This is the whole Casebook corpus: 299 cases — first how it divides, then record by record. The filter works along a case’s five axes — domain, technology, maturity stage, region and source confidence — and every record carries its source and that source’s grade, so a claim here can be checked rather than taken. A caveat sits on 85 records, and on each it is the audit’s own reservation about that source, not a general mark. How it was built and what the audit found are in the Overview; the form at the foot of the page takes a case we have missed.
Corpus division
Domain bands: 3
- 1 Viticulture 177 59.2%
- 2 Wine business, marketing, hospitality 75 25.1%
- 3 Winemaking and laboratory 47 15.7%
Technology bands: 7
- 1 Computer vision and deep learning on images 52 17.4%
- 2 Sensors and IoT 38 12.7%
- 3 Large language models and generative AI 37 12.4%
- 4 Autonomous robotics 31 10.4%
- 5 Optimisation, planning, demand forecasting 27 9.0%
- 6 Chemometrics, spectroscopy and ML in the laboratory 25 8.3%
- 7 The rest 89 29.8%
- In the last band: Predictive disease and weather models — 24 · Remote sensing (satellite, drone, aerial imagery) — 21 · Recommender systems — 20 · Yield forecasting — 16 · Other — 5 · Breeding and genomics — 3. The tail is always last, even when it is larger than its neighbours: it is a remainder, not a value.
Stage bands: 6
- 1 Research 83 27.8%
- 2 Pilot 54 18.1%
- 3 Commercial operation 129 43.1%
- 4 Scaled 18 6.0%
- 5 Closed, acquired or wound down 11 3.7%
- 6 Stage not established 4 1.3%
The bands run in the declared order rather than by size. A maturity ladder is a sequence, and sorting it by how many cases sit on each rung would be a lie about it. An unordered facet runs the other way — largest band first.
«Unstated» takes the sparsest hatch in the set — the same one the empty track takes. Missing data is drawn the same way everywhere, so the band shows at a glance that this is not one more step of the scale.
Region bands: 7
- 1 US and Canada 56 18.7%
- 2 Italy, Spain, Portugal 50 16.7%
- 3 Global platforms, cross-border projects and countries outside the groups 37 12.4%
- 4 France 35 11.7%
- 5 Australia and New Zealand 24 8.0%
- 6 Germany, Austria, Switzerland 23 7.7%
- 7 The rest 74 24.8%
- In the last band: South America and South Africa — 22 · China, Japan, Korea — 18 · Eastern Europe, Balkans, Caucasus, Greece, Russia — 16 · United Kingdom and Scandinavia — 12 · Israel, India, Middle East — 6. The tail is always last, even when it is larger than its neighbours: it is a remainder, not a value.
Confidence bands: 4
- 1 A — peer-reviewed publication, official EU/ministry report or independent press with figures 109 36.5%
- 2 B — trade press or an official company statement with verifiable details 110 36.8%
- 3 C — vendor marketing claim without independent confirmation 61 20.4%
- 4 not stated (peer-reviewed science section) 19 6.3%
The bands run in the declared order rather than by size. A maturity ladder is a sequence, and sorting it by how many cases sit on each rung would be a lie about it. An unordered facet runs the other way — largest band first.
«Unstated» takes the sparsest hatch in the set — the same one the empty track takes. Missing data is drawn the same way everywhere, so the band shows at a glance that this is not one more step of the scale.
The scale under the bar is marked in cases, not in per cent: a tick stands where one group ends and the next begins. The shares in the key are rounded by largest remainder and add up to exactly one hundred per cent. The device carries no colour at all — bands differ by hatch and by number — so it prints identically in colour and in black and white, and it cannot collide with the six domain categoricals the leaderboard's figures encode. A division draws at most 7 bands, because an eighth hatch is no longer distinguishable from its neighbours; whatever is left folds into a single «The rest» band.
Band 1 of 3: Viticulture
177 cases · 59.2% No. Case 1 Tule Technologies (→ CropX)
Estimates vine water stress and irrigation need Closed, acquired or wound down 2018–2023 confidence C C
- Technology
- Evaporation sensors + canopy photo analysis
- What it does
- Estimates vine water stress and irrigation need
- Results
- Acquired by CropX in January 2023 as the company's fourth acquisition.
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Closed, acquired or wound down
- Region
- US and Canada
- Country
- United States and Canada
- Years
- 2018–2023
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The payback quote and the estate names are absent from the cited source; only the CropX acquisition is confirmed.
Not stated in the source operator, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 2 Ceres Imaging
Maps of water stress, nutrition and disease foci Commercial operation 2020 confidence B B
- Operator
- Trinchero Family Estates, Michael David Winery
- Technology
- Aerial imaging in the thermal and multispectral bands
- What it does
- Maps of water stress, nutrition and disease foci
- Results
- Trinchero: grape quality up 25–30% on a problem block after correction of the irrigation problem found
- Domain
- Viticulture
- Technology class
- Remote sensing (satellite, drone, aerial imagery)
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States and Canada
- Years
- 2020
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The year 2020 is not confirmed by the cited source: the page carries no date.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 3 Gamble Family Vineyards
Data-driven management of irrigation and treatments Commercial operation 2022 confidence B B
- Operator
- Gamble Family Vineyards (Oakville)
- Technology
- Drones + a soil sensor network + disease detection
- What it does
- Data-driven management of irrigation and treatments
- Results
- "Tens of thousands of gallons of water per acre" saved
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States
- Years
- 2022
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 4 Bouchaine Vineyards / Cisco
Replaces visual assessment of block condition with telemetry Pilot 2022 confidence B B
- Operator
- Bouchaine Vineyards
- Technology
- IoT network of weather and soil sensors
- What it does
- Replaces visual assessment of block condition with telemetry
- Results
- 87 acres under sensors; the system was extended after the pilot.
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Pilot
- Region
- US and Canada
- Country
- United States and Canada
- Years
- 2022
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The source describes the deployment as an experimental "living laboratory" of 2020, later extended.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 5 Monarch Tractor MK-V
Autonomous inter-row operations and row-by-row data collection Closed, acquired or wound down 2019–2025 confidence A A
- Operator
- Wente Vineyards, Crocker & Starr, Constellation Brands
- Technology
- Electric tractor with optional autonomy + the Scout platform
- What it does
- Autonomous inter-row operations and row-by-row data collection
- Results
- More than $240m raised; 500+ machines; 130,000+ operating hours. The company shut down at the end of 2025 and the technology was sold to a large equipment manufacturer.
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Closed, acquired or wound down
- Region
- US and Canada
- Country
- United States and Canada
- Years
- 2019–2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 6 GUSS Automation (→ John Deere)
One operator runs up to eight machines remotely Closed, acquired or wound down 2018–2025 confidence A A
- Operator
- California vineyards and orchards
- Technology
- Autonomous sprayers with GPS and LiDAR
- What it does
- One operator runs up to eight machines remotely
- Results
- 250+ machines worldwide; 2.6m acres treated; 500,000+ hours of autonomous operation; fully acquired by John Deere in August 2025.
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Closed, acquired or wound down
- Region
- US and Canada
- Country
- United States and Canada
- Years
- 2018–2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 7 Saga Robotics Thorvald (UV-C)
Night-time UV-C treatment against powdery mildew and botrytis without fungicides Commercial operation 2016–2026 confidence A A
- Operator
- Castoro Cellars, Bien Nacido Estate, Bonterra Organic Estates
- Technology
- Autonomous robot with ultraviolet lamps
- What it does
- Night-time UV-C treatment against powdery mildew and botrytis without fungicides
- Results
- 600 organic acres at Castoro Cellars and a Bonterra pilot on 200 acres with six robots; chemical treatments cut by 60–90%
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States and Canada
- Years
- 2016–2026
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 8 Bloomfield Robotics "Scout" (→ Kubota)
Bunch size and count, colour, signs of disease and ripeness Closed, acquired or wound down 2018–2024 confidence B B
- Operator
- A vineyard in New York State (pilot)
- Technology
- Computer vision with per-pixel plant analysis
- What it does
- Bunch size and count, colour, signs of disease and ripeness
- Results
- Acquired by Kubota in 2024; after the acquisition the focus shifted to blueberries and the grape line was wound down.
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Closed, acquired or wound down
- Region
- US and Canada
- Country
- United States and Canada
- Years
- 2018–2024
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 9 PhytoPatholoBot (Cornell)
Near-real-time map of disease type, location and severity Pilot 2025 confidence A A
- Operator
- Cornell AgriTech + commercial vineyards in 6 states
- Technology
- Autonomous ground robot + CV + NASA data
- What it does
- Near-real-time map of disease type, location and severity
- Results
- Detection quality "comparable to experienced scouts"; funding from USDA NIFA and NASA JPL.
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Pilot
- Region
- US and Canada
- Country
- United States and Canada
- Years
- 2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 10 Cornell yield model (Random Forest)
Yield and pruning-weight forecast Research 2023 confidence A A
- Operator
- CLEREL, AVA Lake Erie
- Technology
- Random Forest on pre-flowering vigour data
- What it does
- Yield and pruning-weight forecast
- Results
- Yield forecast error 2–8%; pruning weight 15–20%; 321 sampling points, 2018–2021
- Domain
- Viticulture
- Technology class
- Yield forecasting
- Stage
- Research
- Region
- US and Canada
- Country
- United States and Canada
- Years
- 2023
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 11 VineView
Maps of vigour, leafroll and red blotch Commercial operation 2002–2018 confidence B B
- Operator
- Vineyards in California and France
- Technology
- Aerial spectral imaging + algorithmic diagnosis
- What it does
- Maps of vigour, leafroll and red blotch
- Results
- Operating in California since 2002; merged with SkySquirrel in 2018.
- Domain
- Viticulture
- Technology class
- Remote sensing (satellite, drone, aerial imagery)
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States and Canada
- Years
- 2002–2018
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 12 Reservoir Farms (Sonoma County Winegrowers)
A 14-acre vineyard for field-testing robotics before market launch Pilot 2025 confidence A A
- Operator
- Cropmind, Budbreak Innovations, John Deere
- Technology
- Test ground for agricultural robots
- What it does
- A 14-acre vineyard for field-testing robotics before market launch
- Results
- 14 acres; 3 startups at the outset, target 6 by the end of 2025
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Pilot
- Region
- US and Canada
- Country
- United States and Canada
- Years
- 2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 13 E&J Gallo GIS application
Ripeness assessment and harvest planning in 7–14 day cycles Commercial operation 2017 confidence B B
- Operator
- E&J Gallo (internal tool)
- Technology
- Satellite imagery + GIS + field data
- What it does
- Ripeness assessment and harvest planning in 7–14 day cycles
- Results
- Three years of development from 2014
- Domain
- Viticulture
- Technology class
- Remote sensing (satellite, drone, aerial imagery)
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States and Canada
- Years
- 2017
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat As of 2017 the integration with satellite imagery and GIS was a plan two to three years ahead; what was working was the harvest-planning application itself.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 14 Vinsight
Yield forecast ~4 months before harvest Stage not established ≈2017 confidence C C
- Operator
- Large Californian wineries
- Technology
- Landsat/Terra/Aqua + ML on a decade of harvests
- What it does
- Yield forecast ~4 months before harvest
- Results
- According to the company, mean forecast error is 10% against the industry standard of 30–40%; $250,000 seed. Current status unclear.
- Domain
- Viticulture
- Technology class
- Yield forecasting
- Stage
- Stage not established
- Region
- US and Canada
- Country
- United States and Canada
- Years
- ≈2017 as the source has it: “~2017”
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The accuracy is claimed by the company itself and is not independently confirmed; the company's status as of 2026 is unknown.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 15 Vinergy
Cuts the time a picker spends moving along the row Pilot 2019–2020 confidence B B
- Operator
- Anthony Vineyards
- Technology
- Electric picker-assist carts
- What it does
- Cuts the time a picker spends moving along the row
- Results
- According to the company, 7–10 minutes saved per pass, ~2 hours a day per crew; $820/month rental
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Pilot
- Region
- US and Canada
- Country
- United States and Canada
- Years
- 2019–2020
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The figures are given as the company's claim in general, not as a measured result at Anthony Vineyards.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 16 Semios
A reading every 10 minutes per acre; frost forecasting and pest phenology Scaled 2010–2021 confidence A A
- Operator
- Orchards and vineyards (Canada, United States)
- Technology
- Wireless sensor network + ML
- What it does
- A reading every 10 minutes per acre; frost forecasting and pest phenology
- Results
- $225m raised; 120m acres is a figure for the whole platform.
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Scaled
- Region
- US and Canada
- Country
- United States and Canada
- Years
- 2010–2021
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Caveat 120m acres is the coverage of the whole multi-crop platform after the Agworld purchase, including field crops rather than vineyards.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 17 Biome Makers BeCrop
Soil health assessment from the microbiome Commercial operation 2016–2021 confidence A A
- Operator
- Growers of various crops
- Technology
- Soil DNA sequencing + ML
- What it does
- Soil health assessment from the microbiome
- Results
- $15m Series B led by Prosus Ventures
- Domain
- Viticulture
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States and Canada
- Years
- 2016–2021
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Caveat Of the company's figures, only the $15m round is confirmed by the cited source.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 18 AGERpoint
3D phenotyping: trunk diameter, canopy density, yield forecast Commercial operation 2011–2014 confidence B B
- Operator
- Vineyards and orchards
- Technology
- Ground-based LiDAR + HD cameras
- What it does
- 3D phenotyping: trunk diameter, canopy density, yield forecast
- Results
- 550,000 points per second; up to 300 acres a day
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States and Canada
- Years
- 2011–2014
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 19 Vision Robotics
Autonomous vine pruning at speeds above 3 mph Research — confidence C C
- Operator
- Developer, San Diego
- Technology
- 3D mapping + neural networks
- What it does
- Autonomous vine pruning at speeds above 3 mph
- Results
- There is no deployment data.
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Research
- Region
- US and Canada
- Country
- United States and Canada
- Years
- —
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 20 Tastry (smoke taint)
Screening for smoke taint from the chemical profile Commercial operation 2016–2024 confidence C C
- Operator
- Wineries affected by the 2017 fires
- Technology
- Chemometrics + ML
- What it does
- Screening for smoke taint from the chemical profile
- Results
- The source gives no figures.
- Domain
- Viticulture
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States and Canada
- Years
- 2016–2024
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 21 CladisIQ
Hyperlocal smoke and air-quality data during fires Pilot 2026 confidence C C
- Operator
- Napa Valley wineries
- Technology
- Sensor network + AI
- What it does
- Hyperlocal smoke and air-quality data during fires
- Results
- The source gives no figures.
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Pilot
- Region
- US and Canada
- Country
- United States and Canada
- Years
- 2026
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 22 Carbon Robotics LaserWeeder
Kills weeds without chemicals or soil cultivation Commercial operation 2025 confidence B B
- Operator
- General agriculture (grapes not confirmed)
- Technology
- AI-controlled laser weeding
- What it does
- Kills weeds without chemicals or soil cultivation
- Results
- $70m round; investment from NVIDIA's venture arm
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States and Canada
- Years
- 2025
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 23 Brock CCOVI VineAlert adjacent
Alerts on the risk of winter damage Stage not established — confidence C C
- Operator
- Niagara grape growers (Canada)
- Technology
- Bud cold-hardiness monitoring
- What it does
- Alerts on the risk of winter damage
- Results
- No ML methodology is described in the public materials.
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Stage not established
- Region
- US and Canada
- Country
- United States and Canada
- Years
- —
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- adjacent technology, no AI component
Caveat The maturity stage is not confirmed by the cited source.
Why it is classified this way Differential thermal analysis of bud cold hardiness in programmable chambers. A physical measurement with no predictive model. Not AI: 12 of 299. They stay in the corpus as context rather than as examples, and the classification filter removes them in one click.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 24 Naïo Technologies (Ted, Oz, Jo, Orio)
Replaces herbicides with mechanical cultivation Commercial operation 2011–2025 confidence A A
- Operator
- Vineyards in France and 20+ other countries
- Technology
- Autonomous robots for mechanical weeding
- What it does
- Replaces herbicides with mechanical cultivation
- Results
- 350 machines in service; insolvency proceedings in June 2025, rescued by a €6.4m package from Mirova, Bpifrance and the Occitanie region; headcount cut from 80 to 21; target 100 robots a year by 2028.
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Commercial operation
- Region
- France
- Country
- France
- Years
- 2011–2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 25 VitiBot Bakus
Soil cultivation, spraying, mowing Commercial operation 2020–2024 confidence A A
- Operator
- A winemakers' CUMA (Aude) and private estates
- Technology
- Electric autonomous straddle tractor
- What it does
- Soil cultivation, spraying, mowing
- Results
- €180,000 against €115,000 for a conventional tractor with a driver; payback from 440 engine hours a year; subsidy through the CUMA cooperative under the PCAE scheme; RTK subscription €4,800/year
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Commercial operation
- Region
- France
- Country
- France
- Years
- 2020–2024
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Caveat The subsidy rate under the PCAE scheme is not confirmed by the cited source.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 26 Wall-Ye
Pruning, de-suckering, spraying Closed, acquired or wound down 2016–2019 confidence B B
- Operator
- Estates in southern Burgundy
- Technology
- Shape and colour recognition + robotic pruning
- What it does
- Pruning, de-suckering, spraying
- Results
- ~30 machines sold in its entire history; price €25,000. Vitisphere headline: the robot "fascinates but disappoints".
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Closed, acquired or wound down
- Region
- France
- Country
- France
- Years
- 2016–2019
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 27 Vitirover
Mowing the grass between the rows as a service rather than a machine sale Commercial operation 2021–2026 confidence B B
- Operator
- Bordeaux vineyards, Endesa solar farms
- Technology
- Solar-powered autonomous mower
- What it does
- Mowing the grass between the rows as a service rather than a machine sale
- Results
- 100 robots by the end of 2021, target 200 machines by the end of 2022; ~1 robot per hectare; weight 20 kg; speed 200 m/hour
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Commercial operation
- Region
- France
- Country
- France
- Years
- 2021–2026
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 28 Chouette
Disease detection (downy mildew spots from 0.5 cm), ripening, yield assessment, variable-rate treatment maps Commercial operation 2015–2025 confidence A A
- Operator
- ~100 estates
- Technology
- Computer vision on drones and tractor sensors
- What it does
- Disease detection (downy mildew spots from 0.5 cm), ripening, yield assessment, variable-rate treatment maps
- Results
- €5m Series A in 2023
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Commercial operation
- Region
- France
- Country
- France
- Years
- 2015–2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Caveat The main source is blocked by robots.txt; the €5m and the ~100 estates are confirmed from another source.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 29 Vitidrone
Early disease detection at the level of the individual plant Pilot 2024–2026 confidence B B
- Operator
- Château Le Bon Pasteur, Clos L'Apogée (Pomerol, Saint-Émilion)
- Technology
- Drone + digital twin of geolocated vines
- What it does
- Early disease detection at the level of the individual plant
- Results
- ~2 ha per 40-minute flight; incubation at Inria Start-Up Studio; commercial launch planned for the end of 2026.
- Domain
- Viticulture
- Technology class
- Remote sensing (satellite, drone, aerial imagery)
- Stage
- Pilot
- Region
- France
- Country
- France
- Years
- 2024–2026
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 30 Greenshield / VineMapper
Real-time mapping of downy mildew foci Closed, acquired or wound down 2025 confidence A A
- Operator
- Service subscribers
- Technology
- Tractor-mounted vision in the visible band
- What it does
- Real-time mapping of downy mildew foci
- Results
- €15,000 to buy or €3,000/year to rent + €700–800/year for the AI subscription. Judicial liquidation on 27 March 2025.
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Closed, acquired or wound down
- Region
- France
- Country
- France
- Years
- 2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 31 BioScout SporeScout
Identification of spores of downy mildew, powdery mildew, botrytis and trunk diseases Pilot 2025 confidence B B
- Operator
- French vineyards (market entry in 2025)
- Technology
- Spore trap with AI image analysis
- What it does
- Identification of spores of downy mildew, powdery mildew, botrytis and trunk diseases
- Results
- Air intake 10 l/min; €7,000 in the first year, including the device, software and servicing. The claims of "8 years of R&D" and "8m calibration images" are contradicted by the company's own website: it says "over the past four years" and "hundreds of thousands" of labelled images.
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Pilot
- Region
- France
- Country
- France
- Years
- 2025
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 32 Oenoview (Groupe ICV / Vivelys)
Maps of vigour, water status and plot heterogeneity Scaled — confidence A A
- Operator
- Grands Chais de France, Vignobles de Vendéole, La Vigneronne
- Technology
- SPOT satellite sensing (1.5 m) and Sentinel-2
- What it does
- Maps of vigour, water status and plot heterogeneity
- Results
- The Sentinel-2 satellite gives an update every 5 days.
- Domain
- Viticulture
- Technology class
- Remote sensing (satellite, drone, aerial imagery)
- Stage
- Scaled
- Region
- France
- Country
- France
- Years
- —
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 33 DeciTrait (IFV + chambers of agriculture)
Real-time risk of downy mildew, powdery mildew and black rot Scaled — confidence A A
- Operator
- A national network of estates via MesParcelles
- Technology
- Model-based decision support system
- What it does
- Real-time risk of downy mildew, powdery mildew and black rot
- Results
- Saving of 200-750 g of copper per hectare; reduction of the IFT pesticide-load index by up to 35% on average
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Scaled
- Region
- France
- Country
- France
- Years
- —
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 34 Sencrop
Alerts on disease risk, frost and irrigation timing Scaled 2016–2022 confidence A A
- Operator
- 20,000+ farmers and winegrowers in 20+ countries
- Technology
- Connected micro weather stations + predictive models
- What it does
- Alerts on disease risk, frost and irrigation timing
- Results
- $18m Series B in 2022; ~100 employees
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Scaled
- Region
- France
- Country
- France
- Years
- 2016–2022
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 35 Weenat
Irrigation management and downy mildew alerts Scaled 2014–2025 confidence A A
- Operator
- 30,000+ users in 15 European countries
- Technology
- Agricultural sensors + AI analytics for irrigation and crop protection
- What it does
- Irrigation management and downy mildew alerts
- Results
- 25,000 sensors; €8.5m Series C in 2024; processes more than 1bn data points a day; potential water-use reduction of ~20%.
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Scaled
- Region
- France
- Country
- France
- Years
- 2014–2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 36 Fruition Sciences
Irrigation management by the vine's actual water status Commercial operation с 2007 confidence B B
- Operator
- Ovid Winery
- Technology
- Sap-flow sensors + physiological measurements
- What it does
- Irrigation management by the vine's actual water status
- Results
- Operating since 2007; headquarters in Montpellier and Napa.
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Commercial operation
- Region
- France
- Country
- France
- Years
- с 2007 as the source has it: “2007–”
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The cited source confirms a single client, Ovid Winery.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 37 Sabi Agri
Inter-row work on electric drive Commercial operation 2026 confidence C C
- Operator
- French winegrowers
- Technology
- Electric straddle and tracked machinery, SRBC robot
- What it does
- Inter-row work on electric drive
- Results
- SRBC robot from €13,500; for the range as a whole, 2-3 hours of charging for 8-10 hours of operation and a saving of 10.4 t of CO₂ a year are claimed.
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Commercial operation
- Region
- France
- Country
- France
- Years
- 2026
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The charging time and the CO₂ saving relate to the machine range as a whole; a link to the SRBC robot is not confirmed by the source.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 38 Trapview in France
Remote monitoring of the European grapevine moth instead of manual counting Commercial operation — confidence B B
- Operator
- French vineyards
- Technology
- Computer vision in connected pheromone traps
- What it does
- Remote monitoring of the European grapevine moth instead of manual counting
- Results
- More than 90% accuracy on the European grapevine moth, around fifty species recognised
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Commercial operation
- Region
- France
- Country
- France
- Years
- —
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The global platform figures are not tied to the French deployment.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 39 Moët & Chandon "Photobox"
Automatic scan of grapes on intake: berry size, ripeness, rot; fermentation monitoring Pilot — confidence C C
- Operator
- Le Val du Clos press centre, Champagne
- Technology
- Deep learning + computer vision
- What it does
- Automatic scan of grapes on intake: berry size, ripeness, rot; fermentation monitoring
- Results
- The source describes the work of the research centre; it gives no figures.
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Pilot
- Region
- France
- Country
- France
- Years
- —
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat Deployment across the whole intake and vinification chain is not confirmed by the cited source.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 40 Yanmar YV01
Work on slopes of up to 45° at speeds of up to 4 km/h Commercial operation 2019–2024 confidence A A
- Operator
- Moët & Chandon, Champagne
- Technology
- Autonomous sprayer with RTK-GPS
- What it does
- Work on slopes of up to 45° at speeds of up to 4 km/h
- Results
- Price around £130,000; weeding module on sale since January 2024
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Commercial operation
- Region
- France
- Country
- France
- Years
- 2019–2024
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 41 DIVA project
Automatic detection of flavescence dorée, including asymptomatic cases Research 2025 confidence A A
- Operator
- CIVC (Champagne), CIVB (Bordeaux), BIVB (Burgundy), Bernard Magrez, LVMH
- Technology
- AI + multispectral drone imaging
- What it does
- Automatic detection of flavescence dorée, including asymptomatic cases
- Results
- A three-year doctorate under the CIFRE scheme; a cross-regional industry consortium
- Domain
- Viticulture
- Technology class
- Remote sensing (satellite, drone, aerial imagery)
- Stage
- Research
- Region
- France
- Country
- France
- Years
- 2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 42 Champagne Henriot
Soil and canopy analysis, detection of downy mildew and esca Commercial operation 2020–2025 confidence B B
- Operator
- Champagne Henriot
- Technology
- Camera-equipped tractors + drones
- What it does
- Soil and canopy analysis, detection of downy mildew and esca
- Results
- Organic certification obtained in January 2025; the Alliance Terroirs project since 2020.
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Commercial operation
- Region
- France
- Country
- France
- Years
- 2020–2025
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 43 EXAPTA
Scheduling of machinery, treatments and staff with the weather taken into account Pilot 2016–2017 confidence C C
- Operator
- Bordeaux estates
- Technology
- Optimisation algorithms (BaPCod platform, Inria)
- What it does
- Scheduling of machinery, treatments and staff with the weather taken into account
- Results
- Jointly developed by Ertus Group and Inria/CNRS/Univ. Bordeaux
- Domain
- Viticulture
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Pilot
- Region
- France
- Country
- France
- Years
- 2016–2017
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 44 Pellenc RX-20
Inter-row work Pilot 2024 confidence B B
- Operator
- Early-adopter estates
- Technology
- Autonomous tracked robot
- What it does
- Inter-row work
- Results
- Stage of first field outings
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Pilot
- Region
- France
- Country
- France
- Years
- 2024
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 45 AgriDataValue (Horizon Europe)
Multi-crop smart farming platform with wine pilots Research 2023–2029 confidence A A
- Operator
- Conseil des Vins de Saint-Émilion (France), SIVE (Italy)
- Technology
- IoT + drones + satellite + AI analytics
- What it does
- Multi-crop smart farming platform with wine pilots
- Results
- €7,145,500, 100% EU funding. 23 pilots, 181,000 ha, 4,200 estates and 89,000 beneficiaries — the project's targets by 2029, not figures achieved.
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Research
- Region
- France
- Country
- France
- Years
- 2023–2029
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 46 Generative AI package
Meeting transcription, generation of HACCP training, automatic photo tagging, sales analytics Commercial operation 2024–2025 confidence A A
- Operator
- Cave coopérative de Lugny (Burgundy)
- Technology
- LLM, RPA, analytics
- What it does
- Meeting transcription, generation of HACCP training, automatic photo tagging, sales analytics
- Results
- ~30 internal use cases; one of them (analysis of equipment maintenance diagrams) was explicitly acknowledged as immature and discontinued.
- Domain
- Viticulture
- Technology class
- Large language models and generative AI
- Stage
- Commercial operation
- Region
- France
- Country
- France
- Years
- 2024–2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 47 SATELAI (GMV)
Yield forecast two months before harvest Commercial operation 2022–2024 confidence A A
- Operator
- Pago de Carraovejas (Ribera del Duero)
- Technology
- Sentinel-2 + ML on 6 years of data and 6 weather stations
- What it does
- Yield forecast two months before harvest
- Results
- 92% accuracy in the 2022 campaign and 97% in 2023 after climate parameters were added
- Domain
- Viticulture
- Technology class
- Yield forecasting
- Stage
- Commercial operation
- Region
- Italy, Spain, Portugal
- Country
- Italy, Spain, Portugal
- Years
- 2022–2024
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 48 IntelWINES
Forecast of irrigation demand, reduction of sulphite use Pilot 2019–2021 confidence B B
- Operator
- Pago de Carraovejas + University of Salamanca
- Technology
- Sensor networks + drone thermal imaging at 14 cm/pixel
- What it does
- Forecast of irrigation demand, reduction of sulphite use
- Results
- Pilot 2019-2021, results not published.
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Pilot
- Region
- Italy, Spain, Portugal
- Country
- Italy, Spain, Portugal
- Years
- 2019–2021
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 49 VitiGEOSS (Horizon 2020)
Decision support system for the winegrower Research 2020–2024 confidence A A
- Operator
- Familia Torres (Spain), Mastroberardino (Italy), Symington (Portugal)
- Technology
- Satellite + ground sensors + AI models of disease and phenology
- What it does
- Decision support system for the winegrower
- Results
- €3.03m budget, €2.63m from the EU; 9 partners, 4 countries
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Research
- Region
- Italy, Spain, Portugal
- Country
- Italy, Spain, Portugal
- Years
- 2020–2024
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 50 VineScout (Horizon 2020)
Monitoring of vine vigour and water status Research 2016–2020 confidence A A
- Operator
- Symington Family Estates (Douro) and others
- Technology
- Ground robot with a hyperspectral camera
- What it does
- Monitoring of vine vigour and water status
- Results
- €2.13m budget, €1.74m from the EU; the project's target is 5% of the market across 54,540 ha of European vineyards
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Research
- Region
- Italy, Spain, Portugal
- Country
- Italy, Spain, Portugal
- Years
- 2016–2020
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 51 VINBOT (FP7) — a documented failure
Yield estimation from images Research 2013–2016 confidence A A
- Operator
- Test sites in Portugal: ISA Lisboa, Quinta do Pinto, Quinta da Amieira
- Technology
- Autonomous ground robot + computer vision
- What it does
- Yield estimation from images
- Results
- 112 sessions, 27 plots, 17 grape varieties. The EU's official final report: yield estimates per linear metre showed "weak or no agreement and very high error"; acceptable accuracy appeared only when aggregated over 5-10 m.
- Domain
- Viticulture
- Technology class
- Yield forecasting
- Stage
- Research
- Region
- Italy, Spain, Portugal
- Country
- Italy, Spain, Portugal
- Years
- 2013–2016
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 52 CANOPIES (Horizon 2020)
Human-collaborative harvesting and pruning of table grapes Research 2021–2024 confidence A A
- Operator
- A consortium of 10 partners
- Technology
- Dual-arm mobile manipulator + AI perception
- What it does
- Human-collaborative harvesting and pruning of table grapes
- Results
- €6.9m, 100% EU funding; 10 partners
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Research
- Region
- Italy, Spain, Portugal
- Country
- Italy, Spain, Portugal
- Years
- 2021–2024
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 53 VINUM
Autonomous pruning with selection of the cut point Research 2018–2023 confidence B B
- Operator
- Pilot, Piacenza
- Technology
- Quadruped robot + RGB-D vision
- What it does
- Autonomous pruning with selection of the cut point
- Results
- There is as yet no quantitative comparison with manual pruning.
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Research
- Region
- Italy, Spain, Portugal
- Country
- Italy, Spain, Portugal
- Years
- 2018–2023
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The 2018 start year is not confirmed by the cited source.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 54 Villa Sandi Intelligent Vineyard Ecosystem
Integrated monitoring of plots Commercial operation 2024–2026 confidence B B
- Operator
- Villa Sandi (Veneto, Friuli)
- Technology
- IoT sensors + drones + satellite + AI decision support
- What it does
- Integrated monitoring of plots
- Results
- 200+ ha connected; treatments cut by up to 20%; water by an average of 10%, and by up to 70% on trial plots.
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Commercial operation
- Region
- Italy, Spain, Portugal
- Country
- Italy, Spain, Portugal
- Years
- 2024–2026
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 55 SGUARDO
Monitoring of extreme weather events across the whole DOC territory Research 2024–2025 confidence B B
- Operator
- Consorzio Tutela Prosecco DOC, University of Padua, Veneto region
- Technology
- Satellite reading of the chlorophyll gradient + weather radar
- What it does
- Monitoring of extreme weather events across the whole DOC territory
- Results
- The project has been approved; there are no results yet.
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Research
- Region
- Italy, Spain, Portugal
- Country
- Italy, Spain, Portugal
- Years
- 2024–2025
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The source, from September 2024, reports only the project's approval; field work has not yet begun.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 56 Elaisian
Decision support for irrigation and crop protection Commercial operation — confidence C C
- Operator
- Famiglia Malvetani and other estates
- Technology
- IoT sensors + proprietary ML models
- What it does
- Decision support for irrigation and crop protection
- Results
- For the platform as a whole (not only grapes): 50,000 ha, 4,000 estates, 20 countries, average saving on treatments of 25%
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Commercial operation
- Region
- Italy, Spain, Portugal
- Country
- Italy, Spain, Portugal
- Years
- —
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The figures relate to the whole Elaisian platform across all crops, not only to vineyards.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 57 Evja
Irrigation, nutrition, crop protection Commercial operation 2023 confidence B B
- Operator
- Estates in 9 countries
- Technology
- Patented sensor system + predictive agronomic models
- What it does
- Irrigation, nutrition, crop protection
- Results
- 4 patents; €4.2m pre-Series A in September 2023 from CDP Venture Capital
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Commercial operation
- Region
- Italy, Spain, Portugal
- Country
- Italy, Spain, Portugal
- Years
- 2023
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 58 xFarm Technologies (xTrap, xIdro)
Pest identification and irrigation automation Commercial operation — confidence B B
- Technology
- Image recognition in traps + IoT irrigation
- What it does
- Pest identification and irrigation automation
- Results
- Green Innovation award at Enovitis in Campo 2024
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Commercial operation
- Region
- Italy, Spain, Portugal
- Country
- Italy, Spain, Portugal
- Years
- —
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The cited source names no estate using the solution.
Not stated in the source operator, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 59 iVine
Differentiated treatment Pilot 2023–2024 confidence C C
- Operator
- Fèlsina (Chianti Classico), Mulini di Segalari (Bolgheri)
- Technology
- Digital twin of the canopy from smartphone photos + biometric analysis
- What it does
- Differentiated treatment
- Results
- Funding under Tuscany's PSR programme; the stated target is up to −50% of crop protection products
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Pilot
- Region
- Italy, Spain, Portugal
- Country
- Italy, Spain, Portugal
- Years
- 2023–2024
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat −50% of crop protection products is the project's stated target, not a measured result.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 60 Vigneto Sicuro (Abruzzo Trace Technologies)
Forecast of disease risk and weather events Commercial operation — confidence C C
- Operator
- Italian oenologists
- Technology
- Risk-index algorithm from satellite and weather data without field sensors
- What it does
- Forecast of disease risk and weather events
- Results
- According to the company, 90-99% accuracy and 6,000+ registered oenologists
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Commercial operation
- Region
- Italy, Spain, Portugal
- Country
- Italy, Spain, Portugal
- Years
- —
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat Both figures are given on the company's own claim and are not independently confirmed.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 61 Staffilo
Monitoring of plant condition Commercial operation 2026 confidence C C
- Operator
- Staffilo (organic Prosecco)
- Technology
- Drone monitoring + Demetra platform
- What it does
- Monitoring of plant condition
- Results
- Reduction in water and pesticide use, with no percentages disclosed
- Domain
- Viticulture
- Technology class
- Remote sensing (satellite, drone, aerial imagery)
- Stage
- Commercial operation
- Region
- Italy, Spain, Portugal
- Country
- Italy, Spain, Portugal
- Years
- 2026
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The reduction is a company claim; there is no independent verification.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 62 EyesOnTraps
Pest prevention Pilot — confidence B B
- Operator
- Sogevinus Quintas, Adriano Ramos Pinto, Sogrape Vinhos (Douro)
- Technology
- Computer vision + crowdsensing in pheromone traps
- What it does
- Pest prevention
- Results
- Led by Fraunhofer AICOS together with ADVID and GeoDouro.
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Pilot
- Region
- Italy, Spain, Portugal
- Country
- Italy, Spain, Portugal
- Years
- —
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 63 Wine4cast
National system for forecasting the productivity of Portugal's vineyards Research 2023–2026 confidence A A
- Operator
- FCUP/INESC TEC + regional producers and 6 small enterprises
- Technology
- Optical and photonic sensors, satellite, drone, plant tomography, AI
- What it does
- National system for forecasting the productivity of Portugal's vineyards
- Results
- ~€940,000, PRR funding; 3 years
- Domain
- Viticulture
- Technology class
- Yield forecasting
- Stage
- Research
- Region
- Italy, Spain, Portugal
- Country
- Italy, Spain, Portugal
- Years
- 2023–2026
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 64 AgrarIA
Harvest forecasting Pilot 2021–2023 confidence A A
- Operator
- Familia Torres (Penedès)
- Technology
- Satellite imagery + agroclimatology + AI
- What it does
- Harvest forecasting
- Results
- A consortium of 24 organisations led by GMV, part of Spain's National AI Strategy
- Domain
- Viticulture
- Technology class
- Yield forecasting
- Stage
- Pilot
- Region
- Italy, Spain, Portugal
- Country
- Italy, Spain, Portugal
- Years
- 2021–2023
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 65 ALIMENTE 21
Production management Research 2022–2024 confidence A A
- Operator
- Raventós Codorníu (lead), Aldelís, Prolongo
- Technology
- Deep reinforcement learning, digital twins, edge computing
- What it does
- Production management
- Results
- Budget €5,116,810 (grant €3,097,895), CDTI / Next Generation EU; the only food project approved in the Misiones 2021 competition
- Domain
- Viticulture
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Research
- Region
- Italy, Spain, Portugal
- Country
- Italy, Spain, Portugal
- Years
- 2022–2024
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 66 IRTA + Raventós Codorníu (Raimat) adjacent
Irrigation management Commercial operation 2000–2025 confidence A A
- Operator
- Raimat
- Technology
- Precision irrigation (AI so far only planned)
- What it does
- Irrigation management
- Results
- 20% water saving, 65% improvement in plot uniformity over 25 years of work — without AI; the programme head explicitly calls AI a prospect rather than a current tool.
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Commercial operation
- Region
- Italy, Spain, Portugal
- Country
- Italy, Spain, Portugal
- Years
- 2000–2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- adjacent technology, no AI component
Why it is classified this way Precision irrigation achieved over 25 years without AI — the programme's own head states this explicitly. Not AI: 12 of 299. They stay in the corpus as context rather than as examples, and the classification filter removes them in one click.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 67 VitiMeteo
Simulation of the life cycle of downy mildew and powdery mildew for precise timing of treatments Scaled с 2003 confidence A A
- Operator
- Winegrowers in Baden, Pfalz, Bavaria, Switzerland, Austria and Luxembourg
- Technology
- Epidemiological models on weather data
- What it does
- Simulation of the life cycle of downy mildew and powdery mildew for precise timing of treatments
- Results
- ~42,000 ha covered through ~100 weather stations; in daily use since 2003 — the longest-running working case in the catalogue
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Scaled
- Region
- Germany, Austria, Switzerland
- Country
- Germany, Austria, Switzerland
- Years
- с 2003 as the source has it: “2003–”
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 68 FungiSens
Hyperlocal microclimate instead of weather stations 30–40 km away Pilot 2022–2024 confidence B B
- Operator
- LVWO Weinsberg, Felsengartenkellerei Besigheim
- Technology
- Wireless microsensors inside the canopy + the VitiMeteo model
- What it does
- Hyperlocal microclimate instead of weather stations 30–40 km away
- Results
- Conditions lethal to spores were recorded on 9% of days inside the canopy against 1.4% according to standard weather stations — a clear illustration of the data-resolution problem; ~€500,000.
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Pilot
- Region
- Germany, Austria, Switzerland
- Country
- Germany, Austria, Switzerland
- Years
- 2022–2024
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 69 KI-iREPro
Yield forecasting in place of manual estimation Pilot 2022–2025 confidence A A
- Operator
- Deutsches Weintor eG (Pfalz)
- Technology
- Computer vision from tractor cameras, berry counting
- What it does
- Yield forecasting in place of manual estimation
- Results
- BMEL funding
- Domain
- Viticulture
- Technology class
- Yield forecasting
- Stage
- Pilot
- Region
- Germany, Austria, Switzerland
- Country
- Germany, Austria, Switzerland
- Years
- 2022–2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 70 DigiVine
Selective mechanised harvesting and monitoring of sugar and acidity in the hopper Research 2019–2024 confidence A A
- Operator
- Estates in Rhineland-Palatinate
- Technology
- Hyperspectral imaging + AI; miniature spectrometers inside the harvester
- What it does
- Selective mechanised harvesting and monitoring of sugar and acidity in the hopper
- Results
- Term November 2019 – October 2024, federal ministry funding
- Domain
- Viticulture
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Research
- Region
- Germany, Austria, Switzerland
- Country
- Germany, Austria, Switzerland
- Years
- 2019–2024
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 71 Phenoliner (JKI Geilweilerhof)
High-throughput non-destructive phenotyping of the vine Research 2017 confidence A A
- Operator
- JKI Institute for Grapevine Breeding
- Technology
- Multi-sensor platform: RGB, NIR, VNIR/SWIR
- What it does
- High-throughput non-destructive phenotyping of the vine
- Results
- RTK-GPS accurate to 2 cm; RGB/NIR at 5 Hz, hyperspectral at 100–160 Hz
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Research
- Region
- Germany, Austria, Switzerland
- Country
- Germany, Austria, Switzerland
- Years
- 2017
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 72 DIWAKOPTER
Automatic adjustment of the nitrogen dose for each vine Pilot с 2022 confidence B B
- Operator
- Hessische Staatsweingüter Kloster Eberbach (Rheingau)
- Technology
- Tractor-mounted leaf reflectance sensor + variable-rate dosing
- What it does
- Automatic adjustment of the nitrogen dose for each vine
- Results
- €1.8m, 14 sub-projects; described as the first field trial of the method in German-speaking viticulture.
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Pilot
- Region
- Germany, Austria, Switzerland
- Country
- Germany, Austria, Switzerland
- Years
- с 2022 as the source has it: “2022–”
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 73 Smarter Weinberg
Automation of work on steep slopes Pilot с 2021 confidence B B
- Operator
- Winegrowers of the Mosel, the Ahr and the Mittelrhein
- Technology
- Private 5G network + robot with image recognition + IoT
- What it does
- Automation of work on steep slopes
- Results
- Uplink bandwidth of more than 100 MHz
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Pilot
- Region
- Germany, Austria, Switzerland
- Country
- Germany, Austria, Switzerland
- Years
- с 2021 as the source has it: “2021–”
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The bandwidth is confirmed by a third-party publication — it is not on the project's own page.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 74 Pruning with AI (DLR Mosel)
Training and prompts on gentle pruning cuts to prevent wood diseases Research 2018–2021 confidence B B
- Operator
- DLR Mosel, Bernkastel-Kues
- Technology
- AI + augmented-reality glasses
- What it does
- Training and prompts on gentle pruning cuts to prevent wood diseases
- Results
- A three-year project; no quantitative results disclosed.
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Research
- Region
- Germany, Austria, Switzerland
- Country
- Germany, Austria, Switzerland
- Years
- 2018–2021
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 75 SmartVine / Aqua4D
Precision irrigation under Alpine water scarcity Pilot 2022–2023 confidence B B
- Operator
- Vineyards of the commune of Salgesch (Valais, Switzerland)
- Technology
- Drip irrigation + electromagnetic water treatment + moisture sensors + drone
- What it does
- Precision irrigation under Alpine water scarcity
- Results
- −20% water on the trial plot with comparable yield and quality; the commune's target is more than 40% saving; test site ~2,000 m²
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Pilot
- Region
- Germany, Austria, Switzerland
- Country
- Germany, Austria, Switzerland
- Years
- 2022–2023
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 76 Weingärtner Marbach eG
Irrigation decisions in drought Pilot 2023 confidence B B
- Operator
- Three winegrowing families, the Alter Berg site
- Technology
- LoRaWAN soil moisture sensors
- What it does
- Irrigation decisions in drought
- Results
- 6 sensors in the pilot
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Pilot
- Region
- Germany, Austria, Switzerland
- Country
- Germany, Austria, Switzerland
- Years
- 2023
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 77 Agroscope Wädenswil
Weeding and pre-clinical detection of the disease Research — confidence C C
- Operator
- Swiss vineyards
- Technology
- Autonomous mower + drone for early detection of downy mildew
- What it does
- Weeding and pre-clinical detection of the disease
- Results
- No results; the director is sceptical about drone access to the fruiting zone on steep Swiss slopes.
- Domain
- Viticulture
- Technology class
- Remote sensing (satellite, drone, aerial imagery)
- Stage
- Research
- Region
- Germany, Austria, Switzerland
- Country
- Germany, Austria, Switzerland
- Years
- —
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 78 xarvio FIELD MANAGER for Grapes (BASF)
Treatment recommendations for 100+ grape varieties Commercial operation 2025 confidence B B
- Operator
- Winegrowers in France, Spain and Turkey
- Technology
- Agronomic models (the Hort@ engine) for pests, diseases, irrigation and nutrition
- What it does
- Treatment recommendations for 100+ grape varieties
- Results
- The xarvio platform as a whole: 130,000+ users, 20m+ ha; BASF's R&D spending on digital farming — €919m in 2024
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Commercial operation
- Region
- Germany, Austria, Switzerland
- Country
- Germany, Austria, Switzerland
- Years
- 2025 as the source has it: “ноябрь 2025”
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 79 BOKU Wien
Analysis of vine structure, phenotyping Research 2023 confidence C C
- Operator
- Academic research
- Technology
- Geometric deep learning on 3D point clouds
- What it does
- Analysis of vine structure, phenotyping
- Results
- At the stage of conference papers
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Research
- Region
- Germany, Austria, Switzerland
- Country
- Germany, Austria, Switzerland
- Years
- 2023
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 80 GAIA / National Vineyard Scan (Consilium Technology + Wine Australia)
Detection and mapping of every commercial vineyard block in the country Commercial operation 2018–2019 confidence A A
- Operator
- Australian national programme
- Technology
- ML on Maxar satellite imagery
- What it does
- Detection and mapping of every commercial vineyard block in the country
- Results
- 146,128 ha, 75,961 blocks, 463,718 km of rows; 95% agreement with manual annotation nationwide
- Domain
- Viticulture
- Technology class
- Remote sensing (satellite, drone, aerial imagery)
- Stage
- Commercial operation
- Region
- Australia and New Zealand
- Country
- Australia and New Zealand
- Years
- 2018–2019
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 81 VitiVisor / VitiBox (University of Adelaide AIML)
Counting buds and shoots, detecting inflorescences, recommendations on irrigation and pruning Pilot 2020–2022 confidence A A
- Operator
- Riverland Wine, Wine Australia, PIRSA
- Technology
- Deep learning + a multi-sensor “black box”
- What it does
- Counting buds and shoots, detecting inflorescences, recommendations on irrigation and pruning
- Results
- A $5m project; inflorescence detection accuracy up to 98%; the design is open.
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Pilot
- Region
- Australia and New Zealand
- Country
- Australia and New Zealand
- Years
- 2020–2022
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 82 Cropsy Technologies
Per-vine analytics of disease, pruning quality, buds and yield Scaled 2019–2025 confidence A A
- Operator
- Commercial grape-growing estates
- Technology
- Real-time computer vision on tractors and harvesters
- What it does
- Per-vine analytics of disease, pruning quality, buds and yield
- Results
- 20m vine scans in the last 12 months; ~8,000 vines per hour; 40 active scanners
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Scaled
- Region
- Australia and New Zealand
- Country
- Australia and New Zealand
- Years
- 2019–2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Caveat The names of the client estates are not confirmed by the cited source.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 83 The Yield “Sensing+”
14-day forecasts for treatments and harvest Commercial operation 2018–2020 confidence A A
- Operator
- Treasury Wine Estates, Penfolds Magill Estate
- Technology
- IoT network + predictive microclimate model
- What it does
- 14-day forecasts for treatments and harvest
- Results
- A two-year pilot became a three-year commercial contract in 2020.
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Commercial operation
- Region
- Australia and New Zealand
- Country
- Australia and New Zealand
- Years
- 2018–2020
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 84 Robotics Plus “Prospr” (→ Yamaha)
Row-by-row spraying, one operator to several machines Closed, acquired or wound down 2025 confidence A A
- Operator
- Treasury Wine Estates, Matua vineyard (Marlborough)
- Technology
- Autonomous modular sprayer
- What it does
- Row-by-row spraying, one operator to several machines
- Results
- Fuel consumption reduced by ~70%: 1.5 l/h against 9–10 l/h for a diesel tractor; potential to expand to 150 ha across three vineyards; the company was bought by Yamaha Motor.
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Closed, acquired or wound down
- Region
- Australia and New Zealand
- Country
- Australia and New Zealand
- Years
- 2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 85 Smart Machine “Oxin”
Mowing, mulching and defoliation in a single pass Commercial operation 2023–2024 confidence A A
- Operator
- Pernod Ricard Winemakers, Matapiro vineyard (Hawke's Bay)
- Technology
- Autonomous multi-function platform
- What it does
- Mowing, mulching and defoliation in a single pass
- Results
- 2 machines on site; full block mapping, network setup and operator training took 2 weeks.
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Commercial operation
- Region
- Australia and New Zealand
- Country
- Australia and New Zealand
- Years
- 2023–2024
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 86 UCVision “Occlusion”
“Looking” behind the leaves to count bunches Research 2020–2025 confidence A A
- Operator
- The Universities of Canterbury and Lincoln; a site linked to Cloudy Bay; Waipara Springs
- Technology
- 3D reconstruction with multi-camera robots
- What it does
- “Looking” behind the leaves to count bunches
- Results
- NZ$6m, 5 years, MBIE funding; 3,000 vines on 1.5 ha — the secondary site only
- Domain
- Viticulture
- Technology class
- Yield forecasting
- Stage
- Research
- Region
- Australia and New Zealand
- Country
- Australia and New Zealand
- Years
- 2020–2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 87 12-camera scanning robot
Counting inflorescences and berries for yield forecasting Research — confidence A A
- Operator
- The Universities of Lincoln and Canterbury
- Technology
- 3D photogrammetry, ~10 frames/sec from each side
- What it does
- Counting inflorescences and berries for yield forecasting
- Results
- NZ$6.1m; the aim is to beat the traditional estimation error of 5–10%
- Domain
- Viticulture
- Technology class
- Yield forecasting
- Stage
- Research
- Region
- Australia and New Zealand
- Country
- Australia and New Zealand
- Years
- —
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 88 VinEye
Detection of leafroll virus in a head-to-head comparison with a trained human Pilot 2024 confidence A A
- Operator
- Plant & Food Research, Integrape, Bitwise Agronomy; trials at St Clair Family Estate
- Technology
- ML image classification on a Burro robot
- What it does
- Detection of leafroll virus in a head-to-head comparison with a trained human
- Results
- 60 ha of commercial trials in Hawke's Bay and Marlborough
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Pilot
- Region
- Australia and New Zealand
- Country
- Australia and New Zealand
- Years
- 2024
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 89 MaaraTech
National programme for the robotisation of orchards and vineyards Research 2018–2023 confidence B B
- Operator
- The Universities of Auckland, Waikato, Canterbury and Otago; Plant & Food Research
- Technology
- AI + robotics, plus a sociological study of adoption
- What it does
- National programme for the robotisation of orchards and vineyards
- Results
- 5 years, Endeavour Fund 2018
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Research
- Region
- Australia and New Zealand
- Country
- Australia and New Zealand
- Years
- 2018–2023
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 90 Bragato Research Institute
Formalising vine pruning decisions for automation Research — confidence B B
- Operator
- Bragato / Lincoln University
- Technology
- ML on data about the behaviour of experienced pruners
- What it does
- Formalising vine pruning decisions for automation
- Results
- A Pinot Noir programme co-funded by MBIE through the Endeavour Programme: “we will apply machine learning to predict the relationship between wine chemistry and the perception of quality”.
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Research
- Region
- Australia and New Zealand
- Country
- Australia and New Zealand
- Years
- —
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 91 Onside Intelligence
Faster response to biological incursions Scaled — confidence A A
- Operator
- New Zealand Winegrowers (national biosecurity programme)
- Technology
- Traceability platform and network analytics of movements
- What it does
- Faster response to biological incursions
- Results
- 600+ growers and 700+ wineries on the platform; biosecurity plans will become mandatory for Sustainable Winegrowing NZ members from 2026.
- Domain
- Viticulture
- Technology class
- Other
- Stage
- Scaled
- Region
- Australia and New Zealand
- Country
- Australia and New Zealand
- Years
- —
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 92 AI detection of the brown marmorated stink bug
Field identification of quarantine pests threatening viticulture Pilot 2022 confidence B B
- Operator
- Australian Department of Agriculture, CSIRO, Microsoft
- Technology
- Mobile image recognition
- What it does
- Field identification of quarantine pests threatening viticulture
- Results
- Trials from the 2022 BMSB season
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Pilot
- Region
- Australia and New Zealand
- Country
- Australia and New Zealand
- Years
- 2022
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The source describes the application as a trial, not as a working system.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 93 TWE autonomous fleet (United States)
Multi-vendor autonomous operations Commercial operation 2023 confidence A A
- Operator
- Treasury Wine Estates, vineyards in the United States
- Technology
- Agtonomy, GUSS, Monarch, Robotics Plus, SwarmFarm, VitiBot, Yamaha
- What it does
- Multi-vendor autonomous operations
- Results
- 325+ ha worked autonomously in the 2023 financial year.
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Commercial operation
- Region
- Australia and New Zealand
- Country
- Australia and New Zealand
- Years
- 2023
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 94 Complexica "Larry, the Digital Analyst"
Optimisation of sales territories, visit frequency and logistics Commercial operation 2018 confidence B B
- Operator
- Treasury Wine Estates
- Technology
- Prescriptive analytics
- What it does
- Optimisation of sales territories, visit frequency and logistics
- Results
- TWE: 14,000+ ha of vineyards, 70+ brands, 3,400+ employees
- Domain
- Viticulture
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Commercial operation
- Region
- Australia and New Zealand
- Country
- Australia and New Zealand
- Years
- 2018
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 95 Vendimia 5.0 (Inria Chile + Corfo)
Harvest volume forecasting, phenology tracking, grape intake planning Pilot 2024 confidence A A
- Operator
- Viña Concha y Toro
- Technology
- Explainable AI (XAI)
- What it does
- Harvest volume forecasting, phenology tracking, grape intake planning
- Results
- 1.5bn Chilean pesos of investment (900m from Concha y Toro, 600m from Corfo), a 5-year programme
- Domain
- Viticulture
- Technology class
- Yield forecasting
- Stage
- Pilot
- Region
- South America and South Africa
- Country
- South America and South Africa
- Years
- 2024
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 96 Smart Agro Platform
Irrigation planning Pilot 2022 confidence B B
- Operator
- Viña Concha y Toro
- Technology
- Precision irrigation from micro-weather data and ML
- What it does
- Irrigation planning
- Results
- 18% water saving (~500 m³/ha per year) on a 1,160 ha pilot
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Pilot
- Region
- South America and South Africa
- Country
- South America and South Africa
- Years
- 2022
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat Scaling beyond the pilot is not confirmed by the source.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 97 Centro de Investigación e Innovación (CII)
Management of fermentation, pump-overs and harvest timing Commercial operation 2017–2025 confidence B B
- Operator
- Viña Concha y Toro
- Technology
- Big Data + ML in winemaking; "smart" fermentation tanks
- What it does
- Management of fermentation, pump-overs and harvest timing
- Results
- $6m on R&D over five years; separately $3.158m in 2025, up 14%
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Commercial operation
- Region
- South America and South Africa
- Country
- South America and South Africa
- Years
- 2017–2025
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 98 WiseConn DropControl
Automation of valves and fertigation Commercial operation 2006–2024 confidence C C
- Operator
- Grape-growing regions of Chile, California, Spain, Italy and Australia
- Technology
- Cloud IoT platform with predictive water analytics
- What it does
- Automation of valves and fertigation
- Results
- According to the company, 20,000 field nodes by 2024 (against 1,000 in 2017)
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Commercial operation
- Region
- South America and South Africa
- Country
- South America and South Africa
- Years
- 2006–2024
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The figure comes from the company's own statement and is not independently confirmed; the attribution to years is taken from a graphic on the website.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 99 Kilimo + Microsoft
Reducing water use Commercial operation 2022–2025 confidence B B
- Operator
- Estates of the Maipo Valley (Chile)
- Technology
- AI irrigation planning
- What it does
- Reducing water use
- Results
- −13% water, 450 ha, 1.5m m³ saved over 3 years; agriculture consumes ~68% of the available water in the Maipo basin.
- Domain
- Viticulture
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Commercial operation
- Region
- South America and South Africa
- Country
- South America and South Africa
- Years
- 2022–2025
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 100 Taranis AI2
Detection of pests, diseases and nutrient deficiency within 48–72 hours Commercial operation 2018 confidence B B
- Operator
- 100+ Argentine estates; Mendoza is the existing base, expansion into San Juan was planned
- Technology
- ML on aerial imagery at 0.5 mm resolution
- What it does
- Detection of pests, diseases and nutrient deficiency within 48–72 hours
- Results
- 300,000+ ha flown in the 2017/18 season; $20m Series B.
- Domain
- Viticulture
- Technology class
- Remote sensing (satellite, drone, aerial imagery)
- Stage
- Commercial operation
- Region
- South America and South Africa
- Country
- South America and South Africa
- Years
- 2018
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 101 Embrapa "Crops" + Jahde Tecnologia
A daily morning recommendation: whether a powdery mildew treatment is needed Commercial operation 2021 confidence A A
- Operator
- Aurora cooperative, Vale dos Vinhedos (Brazil)
- Technology
- Model on weather-station data, delivered by SMS
- What it does
- A daily morning recommendation: whether a powdery mildew treatment is needed
- Results
- Validated over three years on the cooperative's estates.
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Commercial operation
- Region
- South America and South Africa
- Country
- South America and South Africa
- Years
- 2021
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 102 Embrapa Uzum Uva
Identification of diseases, pests and physiological disorders Commercial operation 2021 confidence A A
- Operator
- Brazilian growers
- Technology
- Diagnostic application with symptom matching, works offline
- What it does
- Identification of diseases, pests and physiological disorders
- Results
- Free Android application
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Commercial operation
- Region
- South America and South Africa
- Country
- South America and South Africa
- Years
- 2021
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 103 FruitLook
Weekly data on evapotranspiration, growth and nitrogen status at 20×20 m resolution Scaled с 2011 confidence A A
- Operator
- Grape growers and orchardists of the Western Cape (South Africa)
- Technology
- Satellite remote sensing, SEBAL algorithm
- What it does
- Weekly data on evapotranspiration, growth and nitrogen status at 20×20 m resolution
- Results
- 2014/15 season data: 160,000+ ha under weekly monitoring; 8,287 irrigation blocks / 15,608 ha; wine grapes are 23–24% of blocks and of area; budget 3.5m rand per year, free to farmers
- Domain
- Viticulture
- Technology class
- Remote sensing (satellite, drone, aerial imagery)
- Stage
- Scaled
- Region
- South America and South Africa
- Country
- South America and South Africa
- Years
- с 2011 as the source has it: “2011–”
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 104 FruitLook + Random Forest
Yield estimation models by region and grape variety Research 2018–2019 confidence A A
- Operator
- Stellenbosch University, Vinpro, Winetech
- Technology
- Random Forest on FruitLook satellite variables
- What it does
- Yield estimation models by region and grape variety
- Results
- Overall accuracy 85% in the Olifants River region; best results for Chenin Blanc and Colombard; 5 seasons of data
- Domain
- Viticulture
- Technology class
- Yield forecasting
- Stage
- Research
- Region
- South America and South Africa
- Country
- South Africa
- Years
- 2018–2019
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 105 SAGWRI (Stellenbosch)
Preparing the methodological basis for industry models Research 2025 confidence A A
- Operator
- South African wine industry
- Technology
- AI yield and phenology models
- What it does
- Preparing the methodological basis for industry models
- Results
- The project started in 2025; there are no results.
- Domain
- Viticulture
- Technology class
- Yield forecasting
- Stage
- Research
- Region
- South America and South Africa
- Country
- South America and South Africa
- Years
- 2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 106 Vine evapotranspiration mapping at 10 m
Daily water-use maps at 10 m resolution Research 2024 confidence A A
- Operator
- Stellenbosch University
- Technology
- Satellite + vegetation indices
- What it does
- Daily water-use maps at 10 m resolution
- Results
- Started in 2024.
- Domain
- Viticulture
- Technology class
- Remote sensing (satellite, drone, aerial imagery)
- Stage
- Research
- Region
- South America and South Africa
- Country
- South America and South Africa
- Years
- 2024
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 107 Portable IR spectroscopy
Field determination of carbohydrates, nitrogen and amino acids in vine organs Research 2024 confidence A A
- Operator
- Stellenbosch University
- Technology
- Spectroscopy + chemometric calibration
- What it does
- Field determination of carbohydrates, nitrogen and amino acids in vine organs
- Results
- Started in 2024.
- Domain
- Viticulture
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Research
- Region
- South America and South Africa
- Country
- South America and South Africa
- Years
- 2024
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 108 TerraClim
Grape variety selection and climate decisions Commercial operation 2022 confidence A A
- Operator
- South African wine industry
- Technology
- High-resolution terrain and climate data platform
- What it does
- Grape variety selection and climate decisions
- Results
- Co-funding from the South African Department of Science and Innovation
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Commercial operation
- Region
- South America and South Africa
- Country
- South America and South Africa
- Years
- 2022
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 109 "The Dassie" (Robot X)
Automatic collection of spatial data Research 2016 confidence B B
- Operator
- Stellenbosch University + CSIR
- Technology
- LiDAR, HD cameras, soil electromagnetic induction sensors
- What it does
- Automatic collection of spatial data
- Results
- Prototype launched in June 2016, a 12-month trial phase.
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Research
- Region
- South America and South Africa
- Country
- South America and South Africa
- Years
- 2016
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat There is no information on the project continuing after the trials.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 110 Aerobotics
Tree health, yield forecasting Commercial operation 2014–2021 confidence B B
- Operator
- 18 countries; table grapes, not wine grapes
- Technology
- AI analysis of drone and satellite imagery
- What it does
- Tree health, yield forecasting
- Results
- $17m Series B in 2021; 81m+ trees processed — company-wide figures, mostly citrus.
- Domain
- Viticulture
- Technology class
- Remote sensing (satellite, drone, aerial imagery)
- Stage
- Commercial operation
- Region
- South America and South Africa
- Country
- South America and South Africa
- Years
- 2014–2021
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The figures cover the whole company, mostly citrus; the split between table and wine grapes is not confirmed by the source.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 111 Wine of Moldova "AI Vintage"
Harvest timing and fermentation decisions; the AI winemaker persona "Chelaris"; AI labels Pilot 2024–2025 confidence C C
- Operator
- National wine promotion office of Moldova
- Technology
- IoT sensor network + generative AI
- What it does
- Harvest timing and fermentation decisions; the AI winemaker persona "Chelaris"; AI labels
- Results
- According to Wine of Moldova, 14 weather stations on 12 demonstration vineyards, 19 data parameters, 3 wine regions; 2 finished wines released (Rubrum Aeon, Elysium).
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Pilot
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Various countries / global platforms
- Years
- 2024–2025
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The data are self-reported by the national industry agency and not independently confirmed.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 112 DJI Agras in Romania
Precision spraying on slopes Commercial operation 2025 confidence A A
- Operator
- Romanian growers
- Technology
- Autonomous spraying drones
- What it does
- Precision spraying on slopes
- Results
- Chemical use on slopes halved; worldwide, around 400,000 DJI agricultural drones by the end of 2024, +33% over the year and +90% since 2020.
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Commercial operation
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Various countries / global platforms
- Years
- 2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 113 Deep Planet "VineSignal"
Decisions on defoliation, pruning height and soil cultivation; soil organic carbon mapping Commercial operation — confidence B B
- Operator
- Château Pape Clément (France), Koonara Wines (Australia)
- Technology
- Satellite + AI
- What it does
- Decisions on defoliation, pruning height and soil cultivation; soil organic carbon mapping
- Results
- The platform monitors more than 80,000 ha, more than 200 users; claimed savings of €45 per tonne of grapes and a quality gain of €2–20 per bottle. The metrics are platform-wide and are not tied to specific estates.
- Domain
- Viticulture
- Technology class
- Remote sensing (satellite, drone, aerial imagery)
- Stage
- Commercial operation
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Various countries / global platforms
- Years
- —
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 114 Trapview (Slovenia)
Automatic monitoring of pest flight Scaled — confidence B B
- Operator
- 50+ countries, including wine regions
- Technology
- UV trap with computer vision and human verification
- What it does
- Automatic monitoring of pest flight
- Results
- 30m+ verified images, 60+ insect species, the result is checked by a human within 24 hours.
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Scaled
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Various countries / global platforms
- Years
- —
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The figures cover the whole platform and all crops, not only vineyards.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 115 Fasal (India)
Disease forecasting, treatment and irrigation recommendations, remote fertigation control Commercial operation — confidence C C
- Operator
- Indian growers
- Technology
- AI + IoT
- What it does
- Disease forecasting, treatment and irrigation recommendations, remote fertigation control
- Results
- Across the platform: 52bn litres of water saved, a 127,000 kg reduction in chemical use (not only grapes).
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Commercial operation
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Various countries / global platforms
- Years
- —
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat Aggregate platform figures across eight crops, according to the company itself.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 116 Bluewhite Robotics (Israel)
Autonomous operation of a conventional tractor Closed, acquired or wound down 2021–2026 confidence A A
- Operator
- Formerly — orchards and vineyards in the United States through the CNH dealer network
- Technology
- Autonomy kit for existing tractors in ~14 hours
- What it does
- Autonomous operation of a conventional tractor
- Results
- $37m Series B in September 2021; partnership with CNH in June 2024; acquired by Elbit Systems in May 2026 and left agriculture entirely for defence autonomy.
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Closed, acquired or wound down
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Various countries / global platforms
- Years
- 2021–2026
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 117 Ecorobotix
Claimed up to 95% reduction in chemical use Commercial operation 2023–2025 confidence B B
- Operator
- 20+ countries, field crops
- Technology
- AI plant recognition, ultra-precise spraying
- What it does
- Claimed up to 95% reduction in chemical use
- Results
- $150m raised across Series C and D (2024–2025); use on grapes is not confirmed.
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Commercial operation
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Various countries / global platforms
- Years
- 2023–2025
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 118 Solinftec Solix
Field scouting at 1 mph, self-refilling sprayers Commercial operation 2025–2026 confidence B B
- Operator
- United States, Brazil, Colombia, China, Mexico
- Technology
- Solar-powered autonomous scouting robot with 10+ sensors
- What it does
- Field scouting at 1 mph, self-refilling sprayers
- Results
- 300+ robots, 35m acres monitored; $50,000 per machine plus a subscription; grapes not confirmed.
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Commercial operation
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Various countries / global platforms
- Years
- 2025–2026
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 119 Burro
Carrying the harvest between the row and the collection point Commercial operation 2024 confidence A A
- Operator
- 6 countries, including table grapes
- Technology
- Autonomous logistics robot assisting at harvest
- What it does
- Carrying the harvest between the row and the collection point
- Results
- $24m Series B; 300+ robots, 300,000+ autonomous hours, 40+ customers
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Commercial operation
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Various countries / global platforms
- Years
- 2024
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 120 WineGB (data infrastructure) adjacent
Sector record-keeping and planning Commercial operation 2025–2026 confidence A A
- Operator
- The industry of England and Wales
- Technology
- Industry database and interactive map
- What it does
- Sector record-keeping and planning
- Results
- 1,158 vineyards and 260 wineries; 16.5m bottles in 2025; 200% growth in British wine sales over 2018–2024; membership covers ~70% of the area.
- Domain
- Viticulture
- Technology class
- Other
- Stage
- Commercial operation
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Various countries / global platforms
- Years
- 2025–2026
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- adjacent technology, no AI component
Why it is classified this way Industry database and interactive map. Data infrastructure, not analysis. Not AI: 12 of 299. They stay in the corpus as context rather than as examples, and the classification filter removes them in one click.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 190 Grapevine genomic selection with AI
Predicts breeding potential from the genome, shortening multi-year crossing cycles Research 2024 confidence A A
- Operator
- Agricultural Genomics Institute at Shenzhen, Chinese Academy of Agricultural Sciences (Zhou Yongfeng)
- Technology
- Genomic selection ML model + the Grapepan v1.0 pangenome
- What it does
- Predicts breeding potential from the genome, shortening multi-year crossing cycles
- Results
- Prediction accuracy 85%; data on 400+ of ~10,000 grape varieties, 29 traits; breeding efficiency rose fourfold; 6 Chinese patents.
- Domain
- Viticulture
- Technology class
- Breeding and genomics
- Stage
- Research
- Region
- China, Japan, Korea
- Country
- China
- Years
- 2024 as the source has it: “2024, Nature Genetics”
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 191 Genomic selection for pest resistance
Damage assessment from images and genomic prediction for wine and table grape varieties Research 2025 confidence A A
- Operator
- Chinese Academy of Tropical Agriculture, Zhengzhou Fruit Research Institute, Washington State University
- Technology
- Deep convolutional networks (VGG16, ResNet) + genomic selection
- What it does
- Damage assessment from images and genomic prediction for wine and table grape varieties
- Results
- VGG16: classification accuracy 95.3%; DCNN-PDS: R² = 0.88; genomic selection 95.7% on binary traits; 69 loci and 139 candidate genes identified.
- Domain
- Viticulture
- Technology class
- Breeding and genomics
- Stage
- Research
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- China + United States
- Years
- 2025 as the source has it: “2025, Horticulture Research”
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 192 AI phenotyping and climate breeding
Screening hybrids and modelling vineyard suitability 10, 30 and 50 years ahead Pilot 2024 confidence B B
- Operator
- Chinese Academy of Sciences, experimental vineyard in Ningxia (Dai Zhanwu, Xie Jun)
- Technology
- Image recognition for phenotype assessment + climate models
- What it does
- Screening hybrids and modelling vineyard suitability 10, 30 and 50 years ahead
- Results
- ~20,000 new genotypes a year are screened.
- Domain
- Viticulture
- Technology class
- Breeding and genomics
- Stage
- Pilot
- Region
- China, Japan, Korea
- Country
- China
- Years
- 2024
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 193 Smart irrigation and IoT in Ningxia
Data-driven drip irrigation and fermentation management Scaled 2024–2025 confidence B B
- Operator
- Wineries at the eastern foot of Helan Mountain, including Huangkou
- Technology
- IoT soil and weather sensors, irrigation control from an app, digital fermentation control
- What it does
- Data-driven drip irrigation and fermentation management
- Results
- Water use cut from 700-800 to 220-260 m³ per mu a year. Five workers now cover more than 7,000 mu instead of 300 — a more than twentyfold rise in productivity.
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Scaled
- Region
- China, Japan, Korea
- Country
- China
- Years
- 2024–2025
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 194 Xige Estate (西鸽酒庄)
The first Chinese winery to fully digitise vineyard management Commercial operation 2024 confidence B B
- Operator
- Xige Estate
- Technology
- IoT + an in-house big-data platform
- What it does
- The first Chinese winery to fully digitise vineyard management
- Results
- 34m data records, 59 data interfaces, 43,000 enquiries handled; the “one bottle — one code” programme produced more than 20m yuan in online sales, up 70% in a year, across 25 wineries and 104 wines.
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Commercial operation
- Region
- China, Japan, Korea
- Country
- China
- Years
- 2024
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 195 Helan Mountain pest monitoring robot
Early warning of pests and diseases Pilot 2024 confidence B B
- Operator
- Helan Mountain wine industry digitalisation team (Zhang Xuejian)
- Technology
- Ground robot with cameras + cloud analytics
- What it does
- Early warning of pests and diseases
- Results
- Demonstration area ~390 mu; pesticide costs −25%, labour −10%
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Pilot
- Region
- China, Japan, Korea
- Country
- China
- Years
- 2024
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 196 Ningxia “1+N+100” platform
A single infrastructure: IoT monitoring, disease warning, water and fertiliser management Scaled 2024 confidence B B
- Operator
- Government of Ningxia, Wine Industry Development Bureau
- Technology
- Centralised industry data centre + application systems
- What it does
- A single infrastructure: IoT monitoring, disease warning, water and fertiliser management
- Results
- Serves 100 wineries; 14 digital service systems; 8 “digital wineries” created.
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Scaled
- Region
- China, Japan, Korea
- Country
- China
- Years
- 2024
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 197 Ningxia Nongken satellite and drone network
Reducing damage from weather events; the data feeds the regional platform Commercial operation 2026 confidence B B
- Operator
- Ningxia Nongken state agricultural corporation
- Technology
- Satellite sensing + drones
- What it does
- Reducing damage from weather events; the data feeds the regional platform
- Results
- 606,000 mu of vineyards and 136 wineries — figures for the regional platform of the whole eastern foot of Helan Mountain, not for the Ningxia Nongken corporation
- Domain
- Viticulture
- Technology class
- Remote sensing (satellite, drone, aerial imagery)
- Stage
- Commercial operation
- Region
- China, Japan, Korea
- Country
- China
- Years
- 2026
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The figures relate to the regional platform of the eastern foot of Helan Mountain, not to the corporation itself.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 198 “Laishan grape growing and winemaking large model”
Climate adaptation protocols, precise water and fertiliser application, disease forecasting; AI simulation of fermentation and ageing with dynamic parameter optimisation and profile assessment Pilot 2026 confidence B B
- Operator
- Inspur Cloud + the computing centre of China Agricultural University + a company in Yantai
- Technology
- AI foundation model
- What it does
- Climate adaptation protocols, precise water and fertiliser application, disease forecasting; AI simulation of fermentation and ageing with dynamic parameter optimisation and profile assessment
- Results
- No quantitative KPIs disclosed.
- Domain
- Viticulture
- Technology class
- Large language models and generative AI
- Stage
- Pilot
- Region
- China, Japan, Korea
- Country
- China
- Years
- 2026
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 201 Fujitsu + Okunota Winery
Microclimate analysis to reduce treatments Commercial operation с 2011 confidence B B
- Operator
- Okunota Winery
- Technology
- Sensor network (temperature, rainfall, humidity) at 10-minute intervals + an agriculture cloud
- What it does
- Microclimate analysis to reduce treatments
- Results
- Running since June 2011. A four-year analysis identified mould risk thresholds, which made it possible to reduce the frequency of treatments; the wine entered the national Wonder 500 list.
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Commercial operation
- Region
- China, Japan, Korea
- Country
- Japan
- Years
- с 2011 as the source has it: “2011–”
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The experience is described for one winery only; scaling is not confirmed by the source.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 202 Downy mildew outbreak forecasting, Yamanashi
Early intervention for organic viticulture with a minimum of treatments Pilot 2022 confidence B B
- Operator
- Consortium of the Wine Science Centre of the University of Yamanashi (Suzuki Shunji)
- Technology
- Sensors + a cloud forecasting model + mobile alerts
- What it does
- Early intervention for organic viticulture with a minimum of treatments
- Results
- No results yet.
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Pilot
- Region
- China, Japan, Korea
- Country
- Japan
- Years
- 2022
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 203 Smart viticulture demonstration project
Combining Japanese long-arm training and European trellising Pilot 2019 confidence B B
- Operator
- Suntory Wine International, Japan Premium Vineyard, Nippon Steel Solutions
- Technology
- Root zone control + robotic support for operations + AI image analysis of the harvest
- What it does
- Combining Japanese long-arm training and European trellising
- Results
- 4 ha of JPV vineyard (Koshu variety), 3 ha in the first phase; support from Japan's Ministry of Agriculture
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Pilot
- Region
- China, Japan, Korea
- Country
- Japan
- Years
- 2019
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 204 Viticulture training via 5G and smart glasses
Passing an experienced farmer's technique to novices with AI support Pilot ≈2021 confidence B B
- Operator
- Yamanashi Prefecture programme (table grapes)
- Technology
- Smart glasses + a local 5G network + AI
- What it does
- Passing an experienced farmer's technique to novices with AI support
- Results
- The source gives no figures.
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Pilot
- Region
- China, Japan, Korea
- Country
- Japan
- Years
- ≈2021 as the source has it: “~2021”
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 205 GRAPEVINE
Downy mildew outbreak forecasting; pilots in PDO Goumenissa (Greece) and Aragon (Spain) Research 2019–2022 confidence A A
- Operator
- Aristotle University of Thessaloniki, ITAINNOVA, CESGA, University of Zaragoza, Atos
- Technology
- ML on supercomputers
- What it does
- Downy mildew outbreak forecasting; pilots in PDO Goumenissa (Greece) and Aragon (Spain)
- Results
- Budget €2,481,670, 75% EU co-funding
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Research
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Greece + Spain
- Years
- 2019–2022
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 206 BACCHUS
Two cooperating robots with adaptive grippers assess ripeness and selectively harvest grapes Research 2020–2023 confidence A A
- Operator
- Aristotle University of Thessaloniki (coordinator)
- Technology
- Computer vision, hyperspectral imaging, AI decision-making
- What it does
- Two cooperating robots with adaptive grippers assess ripeness and selectively harvest grapes
- Results
- Budget €5,104,943, Horizon 2020; tested on several grape varieties.
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Research
- Region
- Eastern Europe, Balkans, Caucasus, Greece, Russia
- Country
- Greece
- Years
- 2020–2023
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 208 UAV + ML vineyard zoning
Automatic detection of rows and management zone boundaries from multispectral imagery Research 2024 confidence A A
- Operator
- University of Novi Sad, Sremski Karlovci, Fruška Gora
- Technology
- YOLO for vine detection + k-means for zoning, NDVI and NPK
- What it does
- Automatic detection of rows and management zone boundaries from multispectral imagery
- Results
- Vine detection accuracy 90%; studies from 2020 and 2022
- Domain
- Viticulture
- Technology class
- Remote sensing (satellite, drone, aerial imagery)
- Stage
- Research
- Region
- Eastern Europe, Balkans, Caucasus, Greece, Russia
- Country
- Serbia
- Years
- 2024
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 209 “Digital vineyard” (Sevastopol)
A smart vineyard with a “digital agronomist's assistant” Pilot 2021–2023 confidence C C
- Operator
- Sevastopol State University, Crimean Federal University, the Magarach institute, the Koktebel winery
- Technology
- Soil, plant and air sensors, drone pest monitoring, a planned AI disease recogniser
- What it does
- A smart vineyard with a “digital agronomist's assistant”
- Results
- 2.5 ha planned, 2 ha replanted, more than 5,000 vines, 8 grape varieties; sensor installation was scheduled for March 2023. The AI disease recognition module was still at the planning stage at the time of publication.
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Pilot
- Region
- Eastern Europe, Balkans, Caucasus, Greece, Russia
- Country
- Russia, Crimea
- Years
- 2021–2023
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat At the time the source was published the sensors were only planned, and AI disease recognition was at the planning stage.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 210 National vineyard cadastre adjacent
Cadastral mapping of vineyards Commercial operation 2014–2021 confidence B B
- Operator
- National Wine Agency of Georgia
- Technology
- Drone orthophotography + GIS (AI not confirmed)
- What it does
- Cadastral mapping of vineyards
- Results
- Rolled out from 2014 to full coverage of Kakheti; about 20,000 grape growers registered by 2021.
- Domain
- Viticulture
- Technology class
- Remote sensing (satellite, drone, aerial imagery)
- Stage
- Commercial operation
- Region
- Eastern Europe, Balkans, Caucasus, Greece, Russia
- Country
- Georgia
- Years
- 2014–2021
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- adjacent technology, no AI component
Why it is classified this way Orthophotography and GIS without recognition. Data infrastructure. Not AI: 12 of 299. They stay in the corpus as context rather than as examples, and the classification filter removes them in one click.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 214 DOCa Rioja consortium
Regional ML model for yield forecasting by plot Commercial operation 2022–2025 confidence B B
- Operator
- Developed by the appellation council
- What it does
- Regional ML model for yield forecasting by plot
- Results
- Accuracy of up to 96% on plots with detailed field data, over 91% overall in 2024; coverage about 66,000 ha
- Domain
- Viticulture
- Technology class
- Yield forecasting
- Stage
- Commercial operation
- Region
- Italy, Spain, Portugal
- Country
- Spain
- Years
- 2022–2025
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 218 Château Montelena
AI analysis of aerial imagery tracks vine health, water use and uneven ripening in real time Commercial operation 2024 confidence C C
- What it does
- AI analysis of aerial imagery tracks vine health, water use and uneven ripening in real time
- Results
- The source gives no figures.
- Domain
- Viticulture
- Technology class
- Remote sensing (satellite, drone, aerial imagery)
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States, Napa
- Years
- 2024
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Not stated in the source operator, technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 223 Le Carline
AI-Grape project: sensors, drone and satellite imagery, AI pest models for organic treatment with orange oil Pilot 2024–2026 confidence C C
- Operator
- Area Science Park, the 4agri.it platform
- What it does
- AI-Grape project: sensors, drone and satellite imagery, AI pest models for organic treatment with orange oil
- Results
- Stated project targets: −20% pesticides, +15% yield (targets, not results)
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Pilot
- Region
- Italy, Spain, Portugal
- Country
- Italy, Friuli
- Years
- 2024–2026
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat −20% pesticides and +15% yield are the project's stated targets, not measured results.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 224 Vinakoper
The second site of the same AI-Grape project Pilot 2024–2026 confidence C C
- Operator
- Area Science Park, 4agri.it
- What it does
- The second site of the same AI-Grape project
- Results
- “Encouraging results” in the first season, with no figures
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Pilot
- Region
- Eastern Europe, Balkans, Caucasus, Greece, Russia
- Country
- Slovenia
- Years
- 2024–2026
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 225 Zhang C., Northwest A&F University
Bunch detection Research 2022 confidence unstated —
- Technology
- YOLOv5s
- What it does
- Bunch detection
- Results
- Precision, recall, mAP and F1 — all 99.40%
- Dataset
- 8,657 field images
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Research
- Region
- China, Japan, Korea
- Country
- China
- Years
- 2022
- Confidence
- not stated (peer-reviewed science section)
- Classification
- AI
Not stated in the source operator. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 226 Pinheiro I., INESC TEC / UTAD
Bunch detection and damage assessment Research 2023 confidence unstated —
- Technology
- YOLOv7-E6E
- What it does
- Bunch detection and damage assessment
- Results
- mAP 77%, F1 94%, precision 98%; bunch condition mAP 71–72%
- Dataset
- 10,010 images
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Research
- Region
- Italy, Spain, Portugal
- Country
- Portugal
- Years
- 2023
- Confidence
- not stated (peer-reviewed science section)
- Classification
- AI
Not stated in the source operator. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 227 Codes-Alcaraz A.M., Miguel Hernández University
Bunch counting from a drone Research 2025 confidence unstated —
- Technology
- YOLOv7x on UAV RGB
- What it does
- Bunch counting from a drone
- Results
- mAP 0.63; R² = 0.64; RMSE 0.78 bunches per vine against manual counting
- Dataset
- 60 aerial images, a 1.03 ha plot, 2,742 plants
- Domain
- Viticulture
- Technology class
- Yield forecasting
- Stage
- Research
- Region
- Italy, Spain, Portugal
- Country
- Spain
- Years
- 2025
- Confidence
- not stated (peer-reviewed science section)
- Classification
- AI
Not stated in the source operator. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 228 Zhang Z., Northwest A&F + University of Sydney
Downy mildew detection from the leaf Research 2022 confidence unstated —
- Technology
- YOLOv5 with coordinate attention
- What it does
- Downy mildew detection from the leaf
- Results
- Precision 85.6%, recall 83.7%, [email protected] 89.55%, 58.8 frames/s
- Dataset
- 820 leaf samples
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Research
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- China, Australia
- Years
- 2022
- Confidence
- not stated (peer-reviewed science section)
- Classification
- AI
Not stated in the source operator. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 229 De Nart D., CREA
Grape variety identification from the leaf Research 2024 confidence unstated —
- Technology
- Comparison of five CNNs
- What it does
- Grape variety identification from the leaf
- Results
- Cross-validation above 0.9 — and on an independent external set top-1 accuracy falls so far that the authors write outright: “no model gives a satisfactory result”; top-3 is 0.75, top-5 is 0.83. The main result is precisely the failure of transferability.
- Dataset
- 27 grape varieties, 26,382 images, 3 regions, 2 seasons
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Research
- Region
- Italy, Spain, Portugal
- Country
- Italy
- Years
- 2024
- Confidence
- not stated (peer-reviewed science section)
- Classification
- AI
Caveat The exact top-1 values cannot be read in the available online version of the paper; only top-3 (0.75) and top-5 (0.83) are confirmed.
Not stated in the source operator. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 230 Nasiri A. and co-authors
Grape variety identification from the leaf Research 2021 confidence unstated —
- Technology
- Modified VGG16
- What it does
- Grape variety identification from the leaf
- Results
- Mean accuracy above 99% under five-fold cross-validation
- Dataset
- not stated
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Research
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- United States, Iran
- Years
- 2021
- Confidence
- not stated (peer-reviewed science section)
- Classification
- AI
Caveat The first author is affiliated with the University of Tennessee (United States); Iran is the co-authors' affiliation and the subject of the study.
Not stated in the source operator. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 231 Bendel N., Julius Kühn-Institut
Esca detection Research 2020 confidence unstated —
- Technology
- Ground-based hyperspectral + airborne multispectral imaging
- What it does
- Esca detection
- Results
- Symptomatic stage 88–95%; pre-symptomatic only 62–82%; from the air 58–73%
- Dataset
- 129/114/112 vines over three seasons
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Research
- Region
- Germany, Austria, Switzerland
- Country
- Germany
- Years
- 2020
- Confidence
- not stated (peer-reviewed science section)
- Classification
- AI
Caveat The pre-symptomatic figures (62–82%) were not found in the available text of the source — but neither were they refuted.
Not stated in the source operator. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 232 Sawyer E., Fresno State / Cornell / UC ANR
Detection of red blotch and leafroll viruses Research 2023 confidence unstated —
- Technology
- Hyperspectral imaging + Random Forest and 3D-CNN
- What it does
- Detection of red blotch and leafroll viruses
- Results
- Binary classification: on the symptomatic set CNN 87%, RF 82.4%; on the pre-symptomatic set RF 82.8%. Four classes: 77.7% and 76.9%. Both models outperformed an expert's visual assessment.
- Dataset
- ~500 images, 250 vines, 3 vineyards
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Research
- Region
- US and Canada
- Country
- United States
- Years
- 2023
- Confidence
- not stated (peer-reviewed science section)
- Classification
- AI
Caveat 87% and 82.8% were obtained on different sets and are not directly comparable.
Not stated in the source operator. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 233 Montalban-Faet G., University of Valencia
Botrytis detection from a drone Research 2026 confidence unstated —
- Technology
- YOLOv8 + chlorophyll absorption index
- What it does
- Botrytis detection from a drone
- Results
- Precision 92.6%, recall 89.6%, F1 91.1%, mAP@50 93.9% against a baseline RGB model with F1 68.1%
- Dataset
- ~1,575 multispectral images, altitude 40 m
- Domain
- Viticulture
- Technology class
- Remote sensing (satellite, drone, aerial imagery)
- Stage
- Research
- Region
- Italy, Spain, Portugal
- Country
- Spain
- Years
- 2026
- Confidence
- not stated (peer-reviewed science section)
- Classification
- AI
Not stated in the source operator. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 234 Zhu J., Hebei Agricultural University
Black rot detection Research 2021 confidence unstated —
- Technology
- Super-resolution + improved YOLOv3-SPP
- What it does
- Black rot detection
- Results
- 95.79% on the PlantVillage dataset; 86.69% in the real field — the gap between laboratory and field in its pure form
- Dataset
- 1,180 laboratory + 108 field images
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Research
- Region
- China, Japan, Korea
- Country
- China
- Years
- 2021
- Confidence
- not stated (peer-reviewed science section)
- Classification
- AI
Not stated in the source operator. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 235 Pacioni E., University of Extremadura / HES-SO Valais
Identifying the cut point for pruning Research 2025 confidence unstated —
- Technology
- YOLOv8 against Mask R-CNN
- What it does
- Identifying the cut point for pruning
- Results
- mAP50 0.883; inference ~55 ms on a Jetson AGX Orin
- Dataset
- 536 images, 6,968 labelled objects
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Research
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Spain, Switzerland
- Years
- 2025
- Confidence
- not stated (peer-reviewed science section)
- Classification
- AI
Not stated in the source operator. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 236 Kapłan M., University of Life Sciences in Lublin
Locating the winter pruning point Research 2026 confidence unstated —
- Technology
- YOLOv8/YOLO11 + PCAcutSeg-V geometry
- What it does
- Locating the winter pruning point
- Results
- 100% correctness — but only on artificial model vines
- Dataset
- 1,500 RGB images of dormant vines
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Research
- Region
- Eastern Europe, Balkans, Caucasus, Greece, Russia
- Country
- Poland
- Years
- 2026
- Confidence
- not stated (peer-reviewed science section)
- Classification
- AI
Not stated in the source operator. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 237 Guadagna P., Catholic University of Piacenza
Pruning-zone detection and segmentation of vine organs Research 2023 confidence unstated —
- Technology
- Faster R-CNN + Mask R-CNN
- What it does
- Pruning-zone detection and segmentation of vine organs
- Results
- Detection: precision up to 0.96 on Sangiovese, but recall only 0.59. Segmentation: precision 0.97, recall 0.81, F1 0.88.
- Dataset
- 1,215 and 119 images, several vineyards and years
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Research
- Region
- Italy, Spain, Portugal
- Country
- Italy
- Years
- 2023
- Confidence
- not stated (peer-reviewed science section)
- Classification
- AI
Not stated in the source operator. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 238 Andrade C.B., Federal University of Santa Catarina
Regional yield forecasting Research 2023 confidence unstated —
- Technology
- PLSR, Cubist, Random Forest
- What it does
- Regional yield forecasting
- Results
- Best model R² = 0.58, RMSE 2.85 t/ha; weather only R² = 0.52; soil only R² = 0.15
- Dataset
- 534 yield records, 14 harvests, 27 grape varieties
- Domain
- Viticulture
- Technology class
- Yield forecasting
- Stage
- Research
- Region
- South America and South Africa
- Country
- Brazil
- Years
- 2023
- Confidence
- not stated (peer-reviewed science section)
- Classification
- AI
Not stated in the source operator. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 239 Giannico V., University of Bari
Forecasting vine water status Research 2024 confidence unstated —
- Technology
- Ensemble of Lasso, Ridge, Elastic Net and Random Forest on Sentinel-2
- What it does
- Forecasting vine water status
- Results
- R² = 0.72, normalised RMSE 12.4% for stem water potential
- Dataset
- 162 observations, 6 plots, 2 years
- Domain
- Viticulture
- Technology class
- Remote sensing (satellite, drone, aerial imagery)
- Stage
- Research
- Region
- Italy, Spain, Portugal
- Country
- Italy
- Years
- 2024
- Confidence
- not stated (peer-reviewed science section)
- Classification
- AI
Not stated in the source operator. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 240 Armstrong C.E.J., University of Adelaide / CSIRO
Predicting wine sensory characteristics from grape spectra Research 2023 confidence unstated —
- Technology
- XGBoost on A-TEEM data + CIELAB colour
- What it does
- Predicting wine sensory characteristics from grape spectra
- Results
- Only 5 of 22 sensory descriptors reached R² above 0.7 (the best being red fruit aroma, 0.849); 15 of 22 above 0.5. The reference is a panel of 9–11 trained tasters.
- Dataset
- 74 samples, 3 vintages, 8 regions of South Australia
- Domain
- Viticulture
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Research
- Region
- Australia and New Zealand
- Country
- Australia
- Years
- 2023
- Confidence
- not stated (peer-reviewed science section)
- Classification
- AI
Not stated in the source operator. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 241 Elsherbiny O., Jiangsu University
Multi-diagnosis of vine diseases Research 2024 confidence unstated —
- Technology
- CNN-LSTM-DNN hybrid + transfer learning
- What it does
- Multi-diagnosis of vine diseases
- Results
- Precision, recall and F1 all 96.6%, IoU 93.4%
- Dataset
- 295 original + 1,770 augmented field images
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Research
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- China, Egypt
- Years
- 2024
- Confidence
- not stated (peer-reviewed science section)
- Classification
- AI
Not stated in the source operator. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 242 Prasad K.V., Vijayanagara University
Leaf disease classification Research 2024 confidence unstated —
- Technology
- Deep CNN based on VGG16
- What it does
- Leaf disease classification
- Results
- Test accuracy 99.06%
- Dataset
- 9,027 images from Kaggle
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Research
- Region
- Israel, India, Middle East
- Country
- India
- Years
- 2024
- Confidence
- not stated (peer-reviewed science section)
- Classification
- AI
Not stated in the source operator. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 243 Grape bunch detection and segmentation benchmarks
Inflorescence detection and bunch segmentation Research 2022 confidence unstated —
- Technology
- Mask R-CNN, PSPNet, DeepLabV3+
- What it does
- Inflorescence detection and bunch segmentation
- Results
- Mask R-CNN on Chardonnay inflorescences: F1 98%, MAPE 6.92%; PSPNet bunch segmentation IoU 87.42%; DeepLabV3+ IoU 88.44% at 60 ms per image
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Research
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- international
- Years
- 2022
- Confidence
- not stated (peer-reviewed science section)
- Classification
- AI
Not stated in the source operator, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 271 FROST (WeatherQuest + WineGB)
ML and sensor data give a frost-risk forecast tied to the site and the grape variety, combining a budburst model with terrain Pilot 2024–2025 confidence A A
- Operator
- Vineyards of England and Wales, Plumpton College
- What it does
- ML and sensor data give a frost-risk forecast tied to the site and the grape variety, combining a budburst model with terrain
- Results
- £300,000 from Innovate UK and Defra
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Pilot
- Region
- United Kingdom and Scandinavia
- Country
- United Kingdom
- Years
- 2024–2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Caveat The £300,000 figure is confirmed by a different publication — it is not on the page cited.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 272 Autopickr “Vinny”
A machine-vision system on an autonomous robot distinguishes ripe bunches from unripe ones for whole-bunch harvesting at hand-picked quality Research 2024 confidence A A
- Operator
- Coopers Croft Vineyard, supported by WineGB
- What it does
- A machine-vision system on an autonomous robot distinguishes ripe bunches from unripe ones for whole-bunch harvesting at hand-picked quality
- Results
- £475,000 of public funding. Context: 1,033 vineyards and 4,209 ha in the country, up 123% over the decade.
- Domain
- Viticulture
- Technology class
- Autonomous robotics
- Stage
- Research
- Region
- United Kingdom and Scandinavia
- Country
- United Kingdom
- Years
- 2024
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 274 VineMAP (Vinescapes) borderline
Site suitability modelling from elevation, aspect, solar radiation, soils and climate; the methodology is published in a peer-reviewed journal Commercial operation — confidence B B
- Operator
- Vineyard developers
- What it does
- Site suitability modelling from elevation, aspect, solar radiation, soils and climate; the methodology is published in a peer-reviewed journal
- Results
- Customer numbers are not disclosed.
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Commercial operation
- Region
- United Kingdom and Scandinavia
- Country
- United Kingdom
- Years
- —
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- borderline case
Caveat The vendor claims neither AI nor machine learning: its page describes this as geospatial GIS analysis.
Why it is classified this way Predictive site-suitability modelling with a peer-reviewed methodology, but the vendor claims no trained model. Not AI: 12 of 299. They stay in the corpus as context rather than as examples, and the classification filter removes them in one click.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 275 CREWS-UK
Modelling of grape-variety and style suitability and of the shift in frost risk to 2050 Research 2022 confidence A A
- Operator
- Consortium of UEA, LSE Grantham, Vinescapes, WeatherQuest
- What it does
- Modelling of grape-variety and style suitability and of the shift in frost risk to 2050
- Results
- Vineyard area grew from 761 to 3,800 ha over 2004–2021, about +400%; growing-season temperature is forecast at +1.4 °C by 2021–2040.
- Domain
- Viticulture
- Technology class
- Predictive disease and weather models
- Stage
- Research
- Region
- United Kingdom and Scandinavia
- Country
- United Kingdom
- Years
- 2022
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 276 Rathfinny Estate
ML analysis of video from routine mower passes extracts bunch counts and shoot height per vine; a separate vendor assesses canopy growth and ripeness in an app; drone 3D mapping from Imperial College Pilot 2020-е confidence B B
- Operator
- Rathfinny Estate
- What it does
- ML analysis of video from routine mower passes extracts bunch counts and shoot height per vine; a separate vendor assesses canopy growth and ripeness in an app; drone 3D mapping from Imperial College
- Results
- The source gives no figures.
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Pilot
- Region
- United Kingdom and Scandinavia
- Country
- United Kingdom, Sussex
- Years
- 2020-е
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 281 Fieldin
Field-operations management platform: machine telemetry, monitoring of treatments, precision spraying with ARAG; grapes are explicitly listed among the supported crops Commercial operation 2025 confidence B B
- Operator
- Vineyards and orchards
- What it does
- Field-operations management platform: machine telemetry, monitoring of treatments, precision spraying with ARAG; grapes are explicitly listed among the supported crops
- Results
- 750,000+ acres under management across all crops
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Commercial operation
- Region
- Israel, India, Middle East
- Country
- Israel, global
- Years
- 2025
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 282 Negev desert viticulture programme adjacent
Irrigation from a smartphone, breeding of salt-tolerant rootstocks, root-zone sensors Commercial operation с 2019 confidence B B
- Operator
- Nana Estate, Carmey Avdat and 30+ Negev vineyards; grape-variety work by Ariel University
- What it does
- Irrigation from a smartphone, breeding of salt-tolerant rootstocks, root-zone sensors
- Results
- More than 30 working vineyards on 250–280 mm of rainfall a year; Ariel University selected 6 promising grape varieties from 600+ samples of wild grapevine.
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Commercial operation
- Region
- Israel, India, Middle East
- Country
- Israel
- Years
- с 2019 as the source has it: “2019–”
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- adjacent technology, no AI component
Caveat There is no AI component in the source: it describes smartphone irrigation and classical breeding.
Why it is classified this way Smartphone irrigation and classical rootstock breeding. There is no AI component in the source at all. Not AI: 12 of 299. They stay in the corpus as context rather than as examples, and the classification filter removes them in one click.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 283 Netafim NetBeat
A cloud “agro-brain” combines sensing, decision support and automatic control of drip irrigation Commercial operation с 2017 confidence C C
- Operator
- Precision irrigation (no vineyard application named in the source)
- What it does
- A cloud “agro-brain” combines sensing, decision support and automatic control of drip irrigation
- Results
- Water savings of 20–50%, plus 15–20% of monitoring time; 150 agronomists on staff
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Commercial operation
- Region
- Israel, India, Middle East
- Country
- Israel, global
- Years
- с 2017 as the source has it: “2017–”
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The savings relate to drip irrigation across all crops; grapes are not mentioned in the source.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 284 ICAR-NRCG generative AI conference
A national conference on the use of generative AI in agricultural research Research 2023 confidence A A
- Operator
- National Research Centre for Grapes (Pune)
- What it does
- A national conference on the use of generative AI in agricultural research
- Results
- 130 researchers, 30 lectures, 6 plenary talks
- Domain
- Viticulture
- Technology class
- Large language models and generative AI
- Stage
- Research
- Region
- Israel, India, Middle East
- Country
- India
- Years
- 2023
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 285 ICAR-NRCG digital viticulture
IoT sensors, drone and satellite imagery, automated recommendations, stress detection Pilot 2025 confidence A A
- Operator
- India's grape-growing regions
- What it does
- IoT sensors, drone and satellite imagery, automated recommendations, stress detection
- Results
- There are no adoption figures. The barriers named: high up-front costs, a shortage of skills, data-management problems.
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Pilot
- Region
- Israel, India, Middle East
- Country
- India
- Years
- 2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Caveat The source describes India as a whole and names AI in a general list, with no technical detail.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 286 Space AG RaptorView
Satellite monitoring and field data collection Commercial operation 2025–2026 confidence C C
- Operator
- Peruvian grape growers
- What it does
- Satellite monitoring and field data collection
- Results
- There are no independent figures.
- Domain
- Viticulture
- Technology class
- Remote sensing (satellite, drone, aerial imagery)
- Stage
- Commercial operation
- Region
- South America and South Africa
- Country
- Peru
- Years
- 2025–2026
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 287 Austral Falcon
A ground-based device with cameras and GPS counts bunches automatically for yield forecasting Commercial operation 2025–2026 confidence C C
- Operator
- Grape growers
- What it does
- A ground-based device with cameras and GPS counts bunches automatically for yield forecasting
- Results
- The vendor claims 95% forecast accuracy.
- Domain
- Viticulture
- Technology class
- Yield forecasting
- Stage
- Commercial operation
- Region
- South America and South Africa
- Country
- Chile
- Years
- 2025–2026
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The figure is given on the company's claim and is not independently confirmed.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 288 Computer vision in viticulture (Crimean Federal University + Koktebel)
Identifying a disease from an image sent in by a grower; automation of planting-material production Pilot 2023 confidence B B
- Operator
- Crimean Federal University, Koktebel winery, Feodosia
- Technology
- Convolutional networks for recognising diseases from photographs + a robotic grafting line
- What it does
- Identifying a disease from an image sent in by a grower; automation of planting-material production
- Results
- The work was scheduled for completion by the end of 2023; no accuracy or adoption figures have been published.
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Pilot
- Region
- Eastern Europe, Balkans, Caucasus, Greece, Russia
- Country
- Russia, Crimea
- Years
- 2023
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 289 DoctorP (JINR, Dubna)
Diagnosis of diseases and pests from a photograph of a leaf, through a mobile and web app Commercial operation 2017–2023 confidence B B
- Operator
- A publicly available app; 19 crops, including grapes
- Technology
- Convolutional network with transfer learning
- What it does
- Diagnosis of diseases and pests from a photograph of a leaf, through a mobile and web app
- Results
- Overall model accuracy is above 95% on synthetic images — but on real user photographs it falls to roughly 50%. 55–60 classes of diseases and pests; 4,000+ training images; more than 40,000 queries cumulatively (not per year); more than 10,000 users of the Android app. Grapes are explicitly named among the 19 crops.
- Domain
- Viticulture
- Technology class
- Computer vision and deep learning on images
- Stage
- Commercial operation
- Region
- Eastern Europe, Balkans, Caucasus, Greece, Russia
- Country
- Russia
- Years
- 2017–2023
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 291 Abrau-Durso
Claimed use of AI in vineyard monitoring, logistics and demand forecasting Stage not established 2025 confidence C C
- Operator
- Abrau-Durso
- Technology
- Vineyard monitoring, demand forecasting on Sber models
- What it does
- Claimed use of AI in vineyard monitoring, logistics and demand forecasting
- Results
- Neither methodology nor figures have been disclosed.
- Domain
- Viticulture
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Stage not established
- Region
- Eastern Europe, Balkans, Caucasus, Greece, Russia
- Country
- Russia
- Years
- 2025
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The record rests on a single source.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 297 Plant Voice
Detection of water stress and disease before visible symptoms appear Pilot 2026 confidence B B
- Operator
- Wineries through the Wine Tech Challenge accelerator
- Technology
- IoT biosensors implanted in the vine trunk, plus sap-flow analytics
- What it does
- Detection of water stress and disease before visible symptoms appear
- Results
- One of eight finalists in the Wine Tech Challenge 2026
- Domain
- Viticulture
- Technology class
- Sensors and IoT
- Stage
- Pilot
- Region
- Italy, Spain, Portugal
- Country
- Italy
- Years
- 2026
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The source has neither machine learning nor a predictive model — it describes a sensor technology.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
Band 2 of 3: Wine business, marketing, hospitality
75 cases · 25.1% No. Case 153 Vivino
Identifies a wine from a photo and matches it to the user's taste Scaled 2010–2026 confidence A A
- Operator
- Consumers
- Technology
- Label recognition + OCR + recommendation algorithm on crowd data
- What it does
- Identifies a wine from a photo and matches it to the user's taste
- Results
- 74m+ users (2026); 3.26bn+ labels scanned; 19.6m+ wines in the database; 300m ratings. Marketplace commission 15%. Not a single profitable quarter in its history.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Computer vision and deep learning on images
- Stage
- Scaled
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Global
- Years
- 2010–2026
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 154 Cambridge study: crowd ratings versus critics
Comparison of Vivino ratings with professional critics for Bordeaux 2004–2016 Research 2025 confidence A A
- Operator
- Journal of Wine Economics
- Technology
- Correlation analysis
- What it does
- Comparison of Vivino ratings with professional critics for Bordeaux 2004–2016
- Results
- Average correlation of Vivino with critics 40%, against a correlation of 63% between critics themselves; best match Jeff Leve (48%), worst Decanter (16%)
- Domain
- Wine business, marketing, hospitality
- Technology class
- Recommender systems
- Stage
- Research
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Global
- Years
- 2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 155 Liv-ex
Rebuilding the platform around AI Commercial operation 2026 confidence A A
- Operator
- Fine wine exchange
- Technology
- Personalisation, automated pricing, generation of market analytics
- What it does
- Rebuilding the platform around AI
- Results
- According to Liv-ex itself, 500+ members from 42 countries; the stated aim is cellar valuation "from hours to seconds"
- Domain
- Wine business, marketing, hospitality
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Commercial operation
- Region
- United Kingdom and Scandinavia
- Country
- United Kingdom
- Years
- 2026
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Caveat The membership figure is given from Liv-ex's own page; the cited article does not contain it.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 156 Vinovest
Selection of investments to match a risk profile Closed, acquired or wound down 2021–2025 confidence A A
- Operator
- Fine wine investors
- Technology
- "Robo-adviser" on a proprietary algorithm
- What it does
- Selection of investments to match a risk profile
- Results
- More than $100m in assets under management and 250,000 users (Whisky Magazine). The StartEngine purchase closed on 17 March 2026, with 8,750,000 shares issued.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Recommender systems
- Stage
- Closed, acquired or wound down
- Region
- US and Canada
- Country
- United States
- Years
- 2021–2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 157 CultX (Cult Wines)
Fine wine trading platform Commercial operation 2025–2026 confidence A A
- Operator
- Collectors and investors
- Technology
- AI filtering by budget and style, liquidity scoring from transaction data
- What it does
- Fine wine trading platform
- Results
- 6,000 live markets; number of transactions +7.2% in 2025 against 2024, with average prices falling 5.6%
- Domain
- Wine business, marketing, hospitality
- Technology class
- Recommender systems
- Stage
- Commercial operation
- Region
- United Kingdom and Scandinavia
- Country
- United Kingdom
- Years
- 2025–2026
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 158 E&J Gallo + Aera Technology
Automation of direct-shipment routing, inventory rebalancing and overdue-stock management Commercial operation 2024–2025 confidence A A
- Operator
- E&J Gallo
- Technology
- Agentic AI on top of ERP
- What it does
- Automation of direct-shipment routing, inventory rebalancing and overdue-stock management
- Results
- $890,000 saved in the first year, payback about a year; 5 scenarios, ~3 months to deploy each; first go-live in February 2025
- Domain
- Wine business, marketing, hospitality
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States
- Years
- 2024–2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 159 Constellation Brands + Blue Yonder
Replaces planning based on historical sales with ML Commercial operation 2024 confidence C C
- Operator
- Constellation (Robert Mondavi, Kim Crawford, Meiomi, The Prisoner, Ruffino)
- Technology
- Predictive demand forecasting and shelf-space planning
- What it does
- Replaces planning based on historical sales with ML
- Results
- Category sales growth of up to 6%, reduction in out-of-stocks
- Domain
- Wine business, marketing, hospitality
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States
- Years
- 2024
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat "Up to 6%" is a maximum, not an average; the figure is given by the vendor and is not independently confirmed.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 160 Treasury Wine Estates "Innovation Engine"
Clusters feedback into themes for development Commercial operation 2024–2026 confidence B B
- Operator
- Treasury Americas (Cali by Snoop, 19 Crimes, Matua Bagnum)
- Technology
- AI product co-creation platform with online consumer panels
- What it does
- Clusters feedback into themes for development
- Results
- Development cycle cut from 12–18 months to under 6; MVP in 87 days against 180 previously; repeat purchases of Cali by Snoop exceeded the 20% target; top 10 among new launches according to Circana.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States
- Years
- 2024–2026
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 161 Pernod Ricard
Management of a portfolio of 240 brands Scaled 2024–2026 confidence B B
- Operator
- Pernod Ricard
- Technology
- Consumer behaviour forecasting, real-time reallocation of the marketing budget, recommendations for sales representatives
- What it does
- Management of a portfolio of 240 brands
- Results
- Annual marketing budget of about €1.5bn, of which roughly 80% is allocated by these models, across 240 brands.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Scaled
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Global
- Years
- 2024–2026
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 162 Diageo + Strategy (MicroStrategy)
Demand forecasting, pricing, anomaly detection Commercial operation 2025–2026 confidence C C
- Operator
- Diageo
- Technology
- AI agents, auto-generated dashboards, scenario modelling
- What it does
- Demand forecasting, pricing, anomaly detection
- Results
- Calculations 25% faster; data availability 96%. Separately: Diageo operates in about 180 countries.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Commercial operation
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Global
- Years
- 2025–2026
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat "About 180 countries" is a separate fact about Diageo's geography; the source does not link it to the data-availability figure.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 163 Diageo / Vivanda FlavorPrint
Personal recommendations across the portfolio Commercial operation с 2022 confidence B B
- Operator
- Diageo
- Technology
- Algorithmic matching of taste preferences
- What it does
- Personal recommendations across the portfolio
- Results
- Vivanda was acquired by Diageo; localised versions launched in China and India.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Recommender systems
- Stage
- Commercial operation
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Global, including China and India
- Years
- с 2022 as the source has it: “2022–”
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 164 LVMH × Google Cloud
Five-year strategic partnership, a joint "Data and AI Academy" Research с 2021 confidence B B
- Operator
- LVMH, including Moët Hennessy
- Technology
- Demand forecasting, personalised offers
- What it does
- Five-year strategic partnership, a joint "Data and AI Academy"
- Results
- Group-wide LVMH figures: wines and spirits 7% of revenue; the division's sales rose 29% in the first quarter of 2021.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Research
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Global
- Years
- с 2021 as the source has it: “2021–”
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The Google Cloud partnership is an announcement; both figures relate to LVMH as a whole and are not connected to AI.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 165 Preferabli
Personal recommendations and food pairing Commercial operation 2024–2025 confidence B B
- Operator
- Retail and restaurants
- Technology
- "Sensory AI" + generative sommelier chat
- What it does
- Personal recommendations and food pairing
- Results
- 15 patents
- Domain
- Wine business, marketing, hospitality
- Technology class
- Recommender systems
- Stage
- Commercial operation
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- United States, United Kingdom
- Years
- 2024–2025
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The source confirms only the number of patents; the other figures and the client names are absent from it.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 166 Drinks "PAIR"
Recommendations based on visual perception Commercial operation 2025 confidence C C
- Operator
- Alcohol e-commerce
- Technology
- AI analyses label design to predict emotional response
- What it does
- Recommendations based on visual perception
- Results
- According to a statement by the company's co-founder, click-through up by more than 50% against standard recommendations
- Domain
- Wine business, marketing, hospitality
- Technology class
- Recommender systems
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States
- Years
- 2025
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The figure is given from a company statement and is not independently confirmed.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 167 Enolytics
Management of direct sales Commercial operation 2023–2025 confidence C C
- Operator
- Wineries (DTC and wholesale)
- Technology
- Wine club retention analytics, segmentation, forecasting
- What it does
- Management of direct sales
- Results
- According to client testimonials posted on the company's own site: average DTC sales growth among clients of 17.9% against 5.2% for the industry; Black Ankle Vineyards — 85 cases worth $60,000 in 24 hours through targeted segmentation; Far Niente — +6% in corporate gift sales
- Domain
- Wine business, marketing, hospitality
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States
- Years
- 2023–2025
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat All three figures are client testimonials posted by the vendor itself; they have not been independently verified.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 168 Hawesko Group (GK Software + nexum)
Personalisation of the storefront, CRM and campaigns Commercial operation 2026 confidence C C
- Operator
- Hawesko.de, Vinos.de, Tesdorpf.de
- Technology
- Recommendation engine
- What it does
- Personalisation of the storefront, CRM and campaigns
- Results
- Up to 12% higher margin per bottle in an A/B test; catalogue of 8,500 wines
- Domain
- Wine business, marketing, hospitality
- Technology class
- Recommender systems
- Stage
- Commercial operation
- Region
- Germany, Austria, Switzerland
- Country
- Germany
- Years
- 2026
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The figures are given in the vendor's case study; "up to 12%" is the maximum in the A/B test, not the average result.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 169 "Dein Weinfreund"
Wine selection by taste and occasion without registration Commercial operation 2024 confidence C C
- Operator
- Weinfreunde.de
- Technology
- LLM chat adviser
- What it does
- Wine selection by taste and occasion without registration
- Results
- 1,000+ products; running since 1 September 2024.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Commercial operation
- Region
- Germany, Austria, Switzerland
- Country
- Germany
- Years
- 2024
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The figure is given from a company statement and is not independently confirmed.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 170 Vinolin
Recommendations strictly from the partner's own range, around the clock Commercial operation 2025 confidence B B
- Operator
- 15 wineries and cooperatives, including Heilbronn Cooperative Cellar
- Technology
- LLM chatbot inside the winery's shop
- What it does
- Recommendations strictly from the partner's own range, around the clock
- Results
- Grants of €160,000 (Stuttgart ministry) and €40,000 (Campus Founders); a team of 4
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Commercial operation
- Region
- Germany, Austria, Switzerland
- Country
- Germany
- Years
- 2025
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 171 WineSecret
Back end for online shops Commercial operation 2025–2026 confidence B B
- Operator
- Online retailers and distributors
- Technology
- AI chatbot plus sales, inventory and POS analytics
- What it does
- Back end for online shops
- Results
- ~1,000 end users at launch; subscription from 4,166 Hong Kong dollars a month
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Commercial operation
- Region
- China, Japan, Korea
- Country
- Hong Kong
- Years
- 2025–2026
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 172 Firstleaf (Penrose Hill)
Personal matching of the monthly selection Commercial operation 2018–2026 confidence C C
- Operator
- Wine club subscribers
- Technology
- Neural network on subscriber ratings
- What it does
- Personal matching of the monthly selection
- Results
- The company claims that 92% of the wines selected hold competition awards.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Recommender systems
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States
- Years
- 2018–2026
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The 92% is a company claim from a 2018 press release; there is no independent confirmation.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 173 Bright Cellars
Matching wine to a taste profile Commercial operation 2015–2026 confidence C C
- Operator
- Wine club subscribers
- Technology
- "Bright Points" algorithm: 18 attributes against a 7-question survey
- What it does
- Matching wine to a taste profile
- Results
- 600,000+ five-star ratings accumulated.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Recommender systems
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States
- Years
- 2015–2026
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The figure is given from the company itself and is not independently confirmed.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 174 Winc — bankruptcy
Matching wine to a taste profile Closed, acquired or wound down 2011–2022 confidence A A
- Operator
- Wine subscription club
- Technology
- Personalisation through the "Palate Profile" survey
- What it does
- Matching wine to a taste profile
- Results
- Revenue growth of 77.5% during the pandemic (2019–2020) proved unsustainable; DTC revenue in the third quarter of 2022 fell by $2.8m. Chapter 11 in December 2022: debts of $36.75m against assets of $50.3m.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Recommender systems
- Stage
- Closed, acquired or wound down
- Region
- US and Canada
- Country
- United States
- Years
- 2011–2022
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 175 Sippd
Matching by taste profile Commercial operation 2021 confidence C C
- Operator
- Consumers, restaurant partners
- Technology
- "Taste Match" score from 1 to 100 + wine list recognition
- What it does
- Matching by taste profile
- Results
- Catalogue of 10,000+ wines
- Domain
- Wine business, marketing, hospitality
- Technology class
- Recommender systems
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States
- Years
- 2021
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The figure is taken from the company's launch press release and is not independently confirmed.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 176 Chai Wine Vault (Maureen Downey × Everledger) adjacent
Digital passport for a bottle Commercial operation 2016 confidence A A
- Operator
- Traders, retailers, auction houses
- Technology
- Blockchain provenance ledger
- What it does
- Digital passport for a bottle
- Results
- 90+ data parameters plus high-resolution photographs per bottle or case
- Domain
- Wine business, marketing, hospitality
- Technology class
- Other
- Stage
- Commercial operation
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Global
- Years
- 2016
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- adjacent technology, no AI component
Why it is classified this way Blockchain provenance ledger. Adjacent traceability technology. Not AI: 12 of 299. They stay in the corpus as context rather than as examples, and the classification filter removes them in one click.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 177 Prosecco DOC AI Brand Protection
Protection of the appellation against counterfeits Pilot 2024 confidence A A
- Operator
- Consorzio Tutela Prosecco DOC, Microsoft Italia, Istituto Poligrafico
- Technology
- Generative AI on Azure OpenAI
- What it does
- Protection of the appellation against counterfeits
- Results
- Scale of the appellation: total Prosecco DOC output in 2023 was about 616m bottles, 81% for export
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Pilot
- Region
- Italy, Spain, Portugal
- Country
- Italy
- Years
- 2024
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Caveat This is the whole Prosecco DOC output for 2023, not the volume protected by the AI system.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 178 Aforza
Single customer profile, real-time credit scoring, dynamic pricing Commercial operation 2023–2024 confidence C C
- Operator
- Distell / Heineken Beverages
- Technology
- AI platform for field sales and trade marketing
- What it does
- Single customer profile, real-time credit scoring, dynamic pricing
- Results
- According to the vendor's case study, growth in average order value and NPS within 8 months of deployment is claimed without figures; the company has ~4,400 employees and $1.8bn turnover.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Commercial operation
- Region
- South America and South Africa
- Country
- South Africa
- Years
- 2023–2024
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The source is a vendor case study: it contains no quantitative growth figures, only qualitative statements.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 179 WineFi × Lay & Wheeler
Building a fine wine portfolio Commercial operation 2025 confidence B B
- Operator
- Retail investors
- Technology
- Algorithmic selection for an investment syndicate
- What it does
- Building a fine wine portfolio
- Results
- WineFi raised £1.1m through crowdfunding in 2025; minimum entry £3,000, horizon 5 years.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Recommender systems
- Stage
- Commercial operation
- Region
- United Kingdom and Scandinavia
- Country
- United Kingdom
- Years
- 2025
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 180 Total Wine & More
Help with choosing at the point of purchase Commercial operation 2024 confidence C C
- Operator
- Large-format retail
- Technology
- Recommendation models and in-store navigation
- What it does
- Help with choosing at the point of purchase
- Results
- "Thousands of wines" in the store's range
- Domain
- Wine business, marketing, hospitality
- Technology class
- Recommender systems
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States
- Years
- 2024
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 181 Tesco (Clubcard Challenges + Roambee)
Individual challenges and rewards for shoppers Commercial operation 2025 confidence C C
- Operator
- Supermarket, wine category
- Technology
- Loyalty personalisation + AI logistics
- What it does
- Individual challenges and rewards for shoppers
- Results
- 4,250+ Tesco stores in the United Kingdom; 3,000 sites at Roambee
- Domain
- Wine business, marketing, hospitality
- Technology class
- Recommender systems
- Stage
- Commercial operation
- Region
- United Kingdom and Scandinavia
- Country
- United Kingdom
- Years
- 2025
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat 4,250+ stores is the whole Tesco retail network, not the wine category and not the reach of the AI.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 182 Winespace
Standardises subjective vocabulary in six languages into visual aroma profiles Commercial operation 2017–2023 confidence B B
- Operator
- Concours Mondial de Bruxelles, the Euralis cooperative
- Technology
- NLP analysis of tasting notes
- What it does
- Standardises subjective vocabulary in six languages into visual aroma profiles
- Results
- A database of ~12,000 wine assessments
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Commercial operation
- Region
- France
- Country
- France
- Years
- 2017–2023
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 183 Estandon Cooperative
Supports HR, the legal function, quality, procurement and sales Commercial operation 2024–2025 confidence A A
- Operator
- Cooperative, Brignoles
- Technology
- Generative AI
- What it does
- Supports HR, the legal function, quality, procurement and sales
- Results
- No staff cuts were made; the tool is positioned as a “sparring partner”.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Commercial operation
- Region
- France
- Country
- France (Provence)
- Years
- 2024–2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 184 vinSUITE “vinSIGHT”
Retention of the club base Commercial operation — confidence C C
- Operator
- Clients of the DTC/CRM platform
- Technology
- Wine club member churn prediction
- What it does
- Retention of the club base
- Results
- According to the vendor: “Flags at-risk club members with up to 94% confidence and explains which factors affect the risk score”.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States
- Years
- —
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The wording and the 94% figure are taken from the vendor's own site; there is no independent verification.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 185 Sogrape
Sustainability programme Commercial operation 2025 confidence B B
- Operator
- Sogrape Vinhos
- Technology
- AI pilots plus water-treatment technologies
- What it does
- Sustainability programme
- Results
- Portugal: emissions −13.2%, water use −40.5% (from 21.9 to 13 litres per 0.75 bottle)
- Domain
- Wine business, marketing, hospitality
- Technology class
- Other
- Stage
- Commercial operation
- Region
- Italy, Spain, Portugal
- Country
- Portugal, Spain
- Years
- 2025
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The source does not link these figures directly to the AI pilots: it attributes them to water-treatment technologies.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 186 ChatGPT versus the Master Sommelier exam
Answers theory questions at Master Sommelier level Research 2023 confidence B B
- Operator
- Independent test
- Technology
- LLM
- What it does
- Answers theory questions at Master Sommelier level
- Results
- Three theory sections of the exam were passed (unofficial test).
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Research
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Global
- Years
- 2023 as the source has it: “март 2023”
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 187 Weinheimer Group AI Marketing Readiness Report
Winery visibility in the answers of AI assistants Research 2026 confidence B B
- Operator
- Survey of wineries
- Technology
- Optimisation for AI search (GEO)
- What it does
- Winery visibility in the answers of AI assistants
- Results
- 93% of the wineries surveyed are experimenting with AI or gathering information; 7% do not treat it as a priority; 60% name better findability as the main opportunity; 36% name the indistinguishability of hype from reality as the main barrier; 29% are waiting for proof of ROI.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Research
- Region
- US and Canada
- Country
- United States
- Years
- 2026 as the source has it: “апрель 2026”
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 188 weine.ai
Wine selection from a description of the need Commercial operation 2025–2026 confidence C C
- Operator
- importweine.de
- Technology
- Natural-language LLM search
- What it does
- Wine selection from a description of the need
- Results
- According to the company, 109 Mosel and Saar wines, 106 Rieslings; 6 selections a week
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Commercial operation
- Region
- Germany, Austria, Switzerland
- Country
- Germany
- Years
- 2025–2026
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat These are live counters from the company's own catalogue — the values change; there is no independent verification.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 189 Familia Morcos
Label design and an internal knowledge base Pilot 2023–2025 confidence C C
- Operator
- Familia Morcos
- Technology
- Generative AI for labels + an internal LLM on oenology
- What it does
- Label design and an internal knowledge base
- Results
- According to the company's statement, a wine “100% designed by AI” has not yet been released.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Pilot
- Region
- South America and South Africa
- Country
- Argentina
- Years
- 2023–2025
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The wording is the company's own; there is no independent confirmation.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 200 Alibaba “new retail” in wine
A showcase of unmanned wine retail Pilot 2018 confidence C C
- Operator
- Hema, Tmall flagships, Future Bar
- Technology
- RFID shelf recognition, face recognition at the checkout, a hostess robot, a smart QR fridge
- What it does
- A showcase of unmanned wine retail
- Results
- Pilots in Shanghai and Hangzhou; scale not disclosed.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Computer vision and deep learning on images
- Stage
- Pilot
- Region
- China, Japan, Korea
- Country
- China
- Years
- 2018
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 211 Blockchain traceability on Cardano adjacent
Origin-protection pilot Pilot 2022–2023 confidence B B
- Operator
- Cardano Foundation, National Wine Agency, Bolnisi Winemakers Association, Scantrust
- Technology
- Blockchain + QR (not AI)
- What it does
- Origin-protection pilot
- Results
- Up to 100,000 bottles of the 2022 harvest; the Bolnisi association produces ~200,000 bottles a year, with a target of 12m over 10 years.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Other
- Stage
- Pilot
- Region
- Eastern Europe, Balkans, Caucasus, Greece, Russia
- Country
- Georgia
- Years
- 2022–2023
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- adjacent technology, no AI component
Why it is classified this way Blockchain and QR codes. Adjacent traceability technology. Not AI: 12 of 299. They stay in the corpus as context rather than as examples, and the classification filter removes them in one click.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 213 Moët Hennessy “Divine”
An interactive AI sommelier trained on the expertise of all the group's estates, presented as a “painting that comes to life” Commercial operation 2024 confidence B B
- Operator
- OpenAI GPT-4
- What it does
- An interactive AI sommelier trained on the expertise of all the group's estates, presented as a “painting that comes to life”
- Results
- Launched in summer 2024; a rollout to shops and online is planned.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Commercial operation
- Region
- France
- Country
- France, global
- Years
- 2024
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 215 Marchesi Frescobaldi
ML recognises the label and opens AR content about the estate's history, terroir and winemaking Commercial operation 2019–2024 confidence B B
- Operator
- AQuest (Ogilvy partner)
- What it does
- ML recognises the label and opens AR content about the estate's history, terroir and winemaking
- Results
- More than 50 digitised labels
- Domain
- Wine business, marketing, hospitality
- Technology class
- Computer vision and deep learning on images
- Stage
- Commercial operation
- Region
- Italy, Spain, Portugal
- Country
- Italy
- Years
- 2019–2024 as the source has it: “2019, развитие 2024”
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 216 Frescobaldi “Vino Perfetto”
A voice skill answers questions about pairings and occasions in natural language Commercial operation 2022 confidence B B
- Operator
- Amazon Alexa
- What it does
- A voice skill answers questions about pairings and occasions in natural language
- Results
- There are no usage metrics.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Recommender systems
- Stage
- Commercial operation
- Region
- Italy, Spain, Portugal
- Country
- Italy
- Years
- 2022
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 217 Grandes Vinos, El Circo brand
Generative AI turns photographs of staff into circus characters; the consumer assembles a personal video by choosing a grape variety, an element and a style Commercial operation 2024 confidence A A
- Operator
- DeuSens
- What it does
- Generative AI turns photographs of staff into circus characters; the consumer assembles a personal video by choosing a grape variety, an element and a style
- Results
- No sales uplift disclosed.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Commercial operation
- Region
- Italy, Spain, Portugal
- Country
- Spain, Cariñena
- Years
- 2024
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 219 Maison Wessman
Generative graphics create unique labels for the limited cuvée “Imprévu” Commercial operation 2024 confidence C C
- What it does
- Generative graphics create unique labels for the limited cuvée “Imprévu”
- Results
- Limited run
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Commercial operation
- Region
- France
- Country
- France, Bergerac
- Years
- 2024
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Not stated in the source operator, technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 244 SommBench
Multilingual benchmark of wine knowledge: theory, attribute completion, food pairing, in eight languages; 18 models tested Research 2026 confidence A A
- Operator
- Academic consortium
- What it does
- Multilingual benchmark of wine knowledge: theory, attribute completion, food pairing, in eight languages; 18 models tested
- Results
- Theory accuracy up to 0.97 for the best models; attribute completion has a ceiling of 0.63; food pairing — the best MCC is only 0.39; one model approved 86% of the pairings offered, regardless of whether they were correct. The parallel with the OenoBench gap is an editorial comparison made by this catalogue: OenoBench itself is not mentioned in the paper.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Research
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- international (Slovakia, Finland, Denmark)
- Years
- 2026
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 245 Wine Access: randomised LLM trials in email marketing
Three randomised controlled trials: human-written emails against LLM-generated against hybrid, about 9,000 customers per cell Research 2026 confidence A A
- Operator
- Wine Access (DTC retailer)
- What it does
- Three randomised controlled trials: human-written emails against LLM-generated against hybrid, about 9,000 customers per cell
- Results
- LLM and hybrid matched or beat humans on profit in two trials out of three, by up to +9.36%. Copywriters cost $375,000 a year, LLM licences $1,000–1,200, hybrid $63,500–94,950. All the AI variants roughly doubled the probability of purchase against no email at all.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Research
- Region
- US and Canada
- Country
- United States
- Years
- 2026
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 247 Bored Gorilla
The first label generated by Midjourney; every bottle is tied to an NFT with 1/1000 of the image, NFC seals by Authena Commercial operation 2022 confidence A A
- Operator
- Schuler St. Jakobs Kellerei
- What it does
- The first label generated by Midjourney; every bottle is tied to an NFT with 1/1000 of the image, NFC seals by Authena
- Results
- 1,000 magnums, a blend of 60% Merlot and 40% Tempranillo
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Commercial operation
- Region
- Germany, Austria, Switzerland
- Country
- Switzerland
- Years
- 2022
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 248 Sommelier.bot
Chatbot for online shops; enriches products with 30+ taste and terroir attributes, learns from purchase history Commercial operation 2023–2026 confidence B B
- Operator
- 40+ retailers, including REWE, Obrist, City Drinks, OneHope Winery
- What it does
- Chatbot for online shops; enriches products with 30+ taste and terroir attributes, learns from purchase history
- Results
- According to the company, 100,000+ active users, 40+ merchant customers, 5 countries, €299 a month
- Domain
- Wine business, marketing, hospitality
- Technology class
- Recommender systems
- Stage
- Commercial operation
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Europe
- Years
- 2023–2026
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The figures are given as claimed by the company and are not independently confirmed.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 249 Wine Engine “GrapevineAI”
An OpenAI-based chatbot answers questions, simplifies tasting notes, suggests pairings Commercial operation 2025 confidence B B
- Operator
- Subscription service
- What it does
- An OpenAI-based chatbot answers questions, simplifies tasting notes, suggests pairings
- Results
- A catalogue of 400 wines at launch
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States
- Years
- 2025
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 250 Preferabli “Tastefuli” at the Napa Valley Marriott
Personal recommendations on wine, spirits, food and local experiences, built into the concierge's work Commercial operation 2025 confidence B B
- Operator
- Hotel guests
- What it does
- Personal recommendations on wine, spirits, food and local experiences, built into the concierge's work
- Results
- 15 patents held by Preferabli, operating in 100 countries. The first such deployment in Napa.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Recommender systems
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States
- Years
- 2025
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The primary source is unavailable; the figures are confirmed from other publications.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 251 WineSpeak.ai + RedChirp
The AI concierge “Goose” handles bookings, club sign-ups and pairing questions around the clock, coupled with SMS Commercial operation 2025–2026 confidence B B
- Operator
- Goosecross Cellars, Valle della Pace, Jessup Cellars
- What it does
- The AI concierge “Goose” handles bookings, club sign-ups and pairing questions around the clock, coupled with SMS
- Results
- SMS open rate about 98%, conversion 21–30% — roughly ten times higher than email
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States
- Years
- 2025–2026
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 252 Pinpointed
Answers pairing questions and suggests 2–3 items that are actually in stock Commercial operation 2025–2026 confidence C C
- Operator
- Applejack Wine & Spirits, Martins Off Licence, Premier Cru
- What it does
- Answers pairing questions and suggests 2–3 items that are actually in stock
- Results
- The vendor claims +27.2% revenue per customer and a click-through rate of 27.9% against an industry 2–5%. The pair $43.77 → $55.66 is not average order value but “stated budget” against “price of the item clicked”, as the vendor itself defines them.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Commercial operation
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Netherlands, Ireland, United States
- Years
- 2025–2026
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat All the figures are vendor data from a sample of 102 sessions; the methodology is not disclosed.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 253 Santé
Automation of invoice scanning, stock records, order management and multichannel communications Scaled 2026 confidence B B
- Operator
- Hundreds of wine and liquor shops
- What it does
- Automation of invoice scanning, stock records, order management and multichannel communications
- Results
- $7.6m seed round (Bonfire Ventures, Y Combinator); 400% growth in a year; processes more than $500m of annual card turnover.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Scaled
- Region
- US and Canada
- Country
- United States
- Years
- 2026
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 254 Wine-Searcher AI
AI added to the platform as a separate “critic” on a par with the human ones — number 124 in the critics database Commercial operation 2025 confidence B B
- Operator
- Platform users
- What it does
- AI added to the platform as a separate “critic” on a par with the human ones — number 124 in the critics database
- Results
- The source gives no figures.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Commercial operation
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- global
- Years
- 2025
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The source site blocks automated requests: the product's existence is confirmed indirectly, the details have not been verified word for word.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 255 Third Aurora
Recognition and translation of labels, tasting notes and promotional video Pilot 2019 confidence B B
- Operator
- Field trials with 88 wineries in Australia, the United States, Lebanon, Israel
- What it does
- Recognition and translation of labels, tasting notes and promotional video
- Results
- 88 wineries in the trials, the target is more than 100 languages
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Pilot
- Region
- Australia and New Zealand
- Country
- Australia
- Years
- 2019
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 256 Wine Spectator survey: how sommeliers actually use AI
A journalistic study of everyday use: formatting tech sheets, entry into the POS, proofreading, study materials, climate lookups — but not tasting and not recommendations to the guest Research 2025 confidence A A
- Operator
- Restaurant sommeliers
- What it does
- A journalistic study of everyday use: formatting tech sheets, entry into the POS, proofreading, study materials, climate lookups — but not tasting and not recommendations to the guest
- Results
- The source gives no figures.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Research
- Region
- US and Canada
- Country
- United States
- Years
- 2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 257 Joe Roberts (1WineDude): a documented hallucination
Answering a question about Roberts himself, an AI tool invented that he writes for Forbes and Wine Enthusiast, and stated that he has never published a book on wine — though he has published several Research 2023–2025 confidence A A
- Operator
- Wine columnist
- What it does
- Answering a question about Roberts himself, an AI tool invented that he writes for Forbes and Wine Enthusiast, and stated that he has never published a book on wine — though he has published several
- Results
- A specific documented case
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Research
- Region
- US and Canada
- Country
- United States
- Years
- 2023–2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 258 Simon Pavitt on Jane Anson's Inside Bordeaux
An article on the effect of AI on wine journalism, at the end of which the author discloses that ChatGPT wrote 90% of the text of that same article Research 2023 confidence A A
- Operator
- Trade press
- What it does
- An article on the effect of AI on wine journalism, at the end of which the author discloses that ChatGPT wrote 90% of the text of that same article
- Results
- Self-disclosure: 90%
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Research
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- United Kingdom, France
- Years
- 2023
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 259 Randy Caparoso, “Behold the Man”
The argument that AI does not reproduce human perception of wine: “AI does not drink wine — people do”; a trend towards craft as a reaction to automation Research 2025 confidence A A
- Operator
- Wine Industry Advisor
- What it does
- The argument that AI does not reproduce human perception of wine: “AI does not drink wine — people do”; a trend towards craft as a reaction to automation
- Results
- The source gives no figures.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Research
- Region
- US and Canada
- Country
- United States
- Years
- 2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 261 Commerce7 Churn Prediction
Predicts the risk of a wine club member leaving, from anonymised industry data Commercial operation 2025 confidence A A
- Operator
- Wineries on the Commerce7 platform
- What it does
- Predicts the risk of a wine club member leaving, from anonymised industry data
- Results
- Accuracy 74%, trained on “terabytes” of anonymised data
- Domain
- Wine business, marketing, hospitality
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States
- Years
- 2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 262 Commerce7 Fraud Prediction
Real-time scoring of fraudulent orders Commercial operation 2025 confidence A A
- Operator
- The same
- What it does
- Real-time scoring of fraudulent orders
- Results
- Keeps the fraud rate below 0.025%.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States
- Years
- 2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 263 Commerce7 ChatDTC (after the WinePulse acquisition)
Natural-language queries against 70 reports and 14 dashboards Commercial operation 2025 confidence B B
- Operator
- Wineries on the Commerce7 platform (240+ wineries — WinePulse's customer base before the acquisition)
- What it does
- Natural-language queries against 70 reports and 14 dashboards
- Results
- Metrics for the whole Commerce7 platform: 60,000 orders and 16m API calls a day
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States
- Years
- 2025
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat In the source ChatDTC is described in the future tense — as an announcement, not as a working feature.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 264 InnoVint AI Analysis Import
Turns photographs of handwritten notes and laboratory forms into structured data Commercial operation 2025 confidence B B
- Operator
- InnoVint customers
- What it does
- Turns photographs of handwritten notes and laboratory forms into structured data
- Results
- “Hundreds of documents, thousands of analyses”
- Domain
- Wine business, marketing, hospitality
- Technology class
- Computer vision and deep learning on images
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States
- Years
- 2025
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 273 Naked Wines
An ML model forecasts a customer's contribution over a five-year horizon from demographics, interactions and transactions, determining whom to recruit and how much to invest in them Commercial operation 2023 confidence A A
- Operator
- Naked Wines plc
- What it does
- An ML model forecasts a customer's contribution over a five-year horizon from demographics, interactions and transactions, determining whom to recruit and how much to invest in them
- Results
- 35.6m customer reviews in the training set; a forecast five-year payback of 1.7x; 14 years of proprietary behavioural data
- Domain
- Wine business, marketing, hospitality
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Commercial operation
- Region
- United Kingdom and Scandinavia
- Country
- United Kingdom
- Years
- 2023
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 277 Systembolaget
The “Liknande vin” similarity model — a “sommelier in your pocket” — finds wines close in taste and aroma, taking bottle volume into account; generative AI improved product search Commercial operation 2022–2024 confidence A A
- Operator
- Customers of the state monopoly's app
- What it does
- The “Liknande vin” similarity model — a “sommelier in your pocket” — finds wines close in taste and aroma, taking bottle volume into account; generative AI improved product search
- Results
- The full digital range was rolled out in 115 stores by the third quarter of 2024.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Recommender systems
- Stage
- Commercial operation
- Region
- United Kingdom and Scandinavia
- Country
- Sweden
- Years
- 2022–2024 as the source has it: “2022, 2024”
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 278 Winevizer
A virtual sommelier that works to rules and checks against actual cellar stock, not an open chatbot Commercial operation — confidence B B
- Operator
- Restaurants and wine bars
- What it does
- A virtual sommelier that works to rules and checks against actual cellar stock, not an open chatbot
- Results
- Claimed +15–30% bottles sold, +25% average wine bill, choosing time cut from 90 to 25 seconds.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Recommender systems
- Stage
- Commercial operation
- Region
- France
- Country
- France, global
- Years
- —
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The figures are given on the vendor's claim; the source names no venue as a customer.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 279 BinWise, Ingest AI, WineDirect Insights
Predictive analytics for purchasing, write-offs and spoilage risk Commercial operation — confidence C C
- Operator
- Restaurants, hotels, bars
- What it does
- Predictive analytics for purchasing, write-offs and spoilage risk
- Results
- There are no independently confirmed figures.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Commercial operation
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- global
- Years
- —
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 280 LLM as a study tool for the WSET Diploma
Students use general-purpose models to summarise theory and make flashcards; MW-level teachers publicly warn against over-reliance Commercial operation 2024 confidence C C
- Operator
- WSET candidates
- What it does
- Students use general-purpose models to summarise theory and make flashcards; MW-level teachers publicly warn against over-reliance
- Results
- The source gives no figures.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Commercial operation
- Region
- United Kingdom and Scandinavia
- Country
- United Kingdom, global
- Years
- 2024
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat All the source confirms is a Master of Wine's recommendation to use an LLM for flashcards.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 290 VINTELLEKT (Vino.ru + Gureev.Pro)
A Telegram sommelier bot: it picks out type, sugar, strength and aromatics from a free-text request and returns three wines with reasons Commercial operation 2024 confidence B B
- Operator
- The Vino.ru marketplace
- Technology
- Fine-tuned YandexGPT 3 Pro + semantic search over the catalogue
- What it does
- A Telegram sommelier bot: it picks out type, sugar, strength and aromatics from a free-text request and returns three wines with reasons
- Results
- Trained on 1,000+ real customer requests with professional sommeliers involved.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Commercial operation
- Region
- Eastern Europe, Balkans, Caucasus, Greece, Russia
- Country
- Russia
- Years
- 2024 as the source has it: “июнь 2024”
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 293 Fanagoria
Automation of sales and marketing, and generation of product ideas Pilot 2025 confidence C C
- Operator
- Fanagoria
- Technology
- Language models
- What it does
- Automation of sales and marketing, and generation of product ideas
- Results
- The source gives no figures.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Pilot
- Region
- Eastern Europe, Balkans, Caucasus, Greece, Russia
- Country
- Russia
- Years
- 2025
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 294 Descartes Underwriting
Payout on the occurrence of a weather event, with no on-site loss assessment Commercial operation 2018–2025 confidence C C
- Operator
- Grape growers; a world leader in cognac (not named)
- Technology
- Parametric insurance against frost and hail: AI reprocesses historical satellite cloud imagery, plus IoT weather stations
- What it does
- Payout on the occurrence of a weather event, with no on-site loss assessment
- Results
- The company publishes no specific figures. All that is confirmed is a parametric insurance line for viticulture and AI in hail-risk modelling.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Predictive disease and weather models
- Stage
- Commercial operation
- Region
- France
- Country
- France → Australia, France
- Years
- 2018–2025
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The source is the company's own pages; there is no independent confirmation.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 295 Hillebrand Gori (DHL Group)
Combines WMO weather data with a route database, predicting temperature and humidity risk for a specific wine shipment Research 2023–2025 confidence B B
- Operator
- Wine importers and exporters on the myHillebrandGori platform
- Technology
- Demand forecasting and weather-risk modelling
- What it does
- Combines WMO weather data with a route database, predicting temperature and humidity risk for a specific wine shipment
- Results
- A database of 110,000 sea routes, 3,300 alternative routings and 2,500 cities
- Domain
- Wine business, marketing, hospitality
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Research
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Netherlands, global
- Years
- 2023–2025
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The company describes the database as an aggregation of weather and logistics data, and mentions AI features in the future tense.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 296 IVDP “Winalytics” (Data+)
Production-volume forecasting, production-cost prediction, optimisation of distribution routes, market recommendations and traceability from berry to bottle Pilot 2020–2021 confidence B B
- Operator
- Douro and Port Wine Institute — the appellation regulator
- Technology
- Descriptive, predictive and prescriptive analytics
- What it does
- Production-volume forecasting, production-cost prediction, optimisation of distribution routes, market recommendations and traceability from berry to bottle
- Results
- A budget of €300,000, the SAMA IA programme under Portugal 2020, project completed in December 2021.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Pilot
- Region
- Italy, Spain, Portugal
- Country
- Portugal
- Years
- 2020–2021
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 298 Dolia
Order automation and real-time stock monitoring Pilot 2026 confidence B B
- Operator
- Wine companies
- Technology
- AI centralisation of sales processes
- What it does
- Order automation and real-time stock monitoring
- Results
- One of eight finalists in the Wine Tech Challenge 2026
- Domain
- Wine business, marketing, hospitality
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Pilot
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Luxembourg
- Years
- 2026
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 299 VinoBuzz
Wine selection in conversation, with delivery Pilot 2026 confidence C C
- Operator
- A consumer marketplace
- Technology
- A conversational AI sommelier on top of a multi-merchant marketplace
- What it does
- Wine selection in conversation, with delivery
- Results
- Claimed: a $10m valuation, 1,000+ sign-ups in two weeks of beta, 4,000+ SKUs. Market context: fewer than 10% of wine purchases in Hong Kong are made online.
- Domain
- Wine business, marketing, hospitality
- Technology class
- Large language models and generative AI
- Stage
- Pilot
- Region
- China, Japan, Korea
- Country
- Hong Kong
- Years
- 2026 as the source has it: “апрель 2026”
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat All the figures come from the company's press release, reprinted by an in-flight magazine; there is no independent confirmation.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
Band 3 of 3: Winemaking and laboratory
47 cases · 15.7% No. Case 121 Schartner, Pouget et al., University of Geneva
Identifies the estate and the vintage directly from the raw chromatogram, without peak identification Research 2023 confidence A A
- Operator
- Peer-reviewed study
- Technology
- Raw GC chromatograms + linear discriminant analysis
- What it does
- Identifies the estate and the vintage directly from the raw chromatogram, without peak identification
- Results
- Estate identification accuracy 99% (N = 80 wines, 7 Bordeaux estates). For vintage: up to 50% on the most informative regions of the chromatogram and 27% on the full chromatogram — against a random baseline of 8% (12 vintage classes).
- Domain
- Winemaking and laboratory
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Research
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- Switzerland / Bordeaux
- Years
- 2023
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 122 Gonzalez Viejo & Fuentes, University of Melbourne
Classification of 12 categories of wine fault (brett, TCA, guaiacol, acetaldehyde, volatile acidity, mercaptans) Research 2022 confidence A A
- Operator
- Peer-reviewed study
- Technology
- Low-cost 9-sensor electronic nose + NIR + neural network
- What it does
- Classification of 12 categories of wine fault (brett, TCA, guaiacol, acetaldehyde, volatile acidity, mercaptans)
- Results
- Accuracy 90–97% for the electronic nose and 94–97% for NIR; ~396 samples
- Domain
- Winemaking and laboratory
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Research
- Region
- Australia and New Zealand
- Country
- Australia
- Years
- 2022
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 123 Sarlo et al., University of Lyon 1
Predicts country, French region and grape variety from the mineral "fingerprint" Research 2024 confidence A A
- Operator
- Peer-reviewed study
- Technology
- Mineral profile (ICP) + XGBoost
- What it does
- Predicts country, French region and grape variety from the mineral "fingerprint"
- Results
- Country 92%, French region 91%, grape variety 85%; specificity above 99%; a database of 12,966 profiles
- Domain
- Winemaking and laboratory
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Research
- Region
- France
- Country
- France
- Years
- 2024
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 124 Hategan et al., Cluj-Napoca
Classification of grape variety, region and vintage of white wines Research 2025 confidence A A
- Operator
- Peer-reviewed study
- Technology
- NMR spectroscopy + kNN and logistic regression
- What it does
- Classification of grape variety, region and vintage of white wines
- Results
- Above 98% under cross-validation, up to 100% on the test set; N = 65
- Domain
- Winemaking and laboratory
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Research
- Region
- Eastern Europe, Balkans, Caucasus, Greece, Russia
- Country
- Romania
- Years
- 2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 125 Ferrier & Block, UC Davis
Reduces the number of trial blends needed to find the optimum Research 2001 confidence A A
- Operator
- Peer-reviewed study
- Technology
- Neural network modelling the non-linear sensory response to blend proportions
- What it does
- Reduces the number of trial blends needed to find the optimum
- Results
- Under 2% composition error with a 30% reduction in the number of trials; up to 11% error with a 50% reduction
- Domain
- Winemaking and laboratory
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Research
- Region
- US and Canada
- Country
- United States
- Years
- 2001
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 126 Tan, Oberholster, Tagkopoulos et al., UC Davis (USDA-NIFA AI institute)
Prediction of the smoke taint sensory index Research 2024 confidence B B
- Operator
- Peer-reviewed study
- Technology
- Lasso regression, SVR, Random Forest on the volatile compound profile
- What it does
- Prediction of the smoke taint sensory index
- Results
- A comparison of four model families (linear, Lasso, SVM, Random Forest); the code is open.
- Domain
- Winemaking and laboratory
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Research
- Region
- US and Canada
- Country
- United States
- Years
- 2024
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 127 Smoke taint prediction from NIR + ANN
Prediction of the level of volatile phenols and glycoconjugates in berries, must and wine Research 2020 confidence A A
- Operator
- University of Melbourne (AWRI — external reference laboratory)
- Technology
- NIR spectroscopy + neural network
- What it does
- Prediction of the level of volatile phenols and glycoconjugates in berries, must and wine
- Results
- R² = 0.95–0.99 across five models; 540 berry samples
- Domain
- Winemaking and laboratory
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Research
- Region
- Australia and New Zealand
- Country
- Australia
- Years
- 2020
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 128 Pinot noir sensory profile model
Prediction of 19 sensory descriptors and wine colour from NIR, weather and agronomic practices Research 2020 confidence A A
- Operator
- Boutique estate, Macedon Ranges
- Technology
- Neural network regression
- What it does
- Prediction of 19 sensory descriptors and wine colour from NIR, weather and agronomic practices
- Results
- R = 0.92 from NIR; R = 0.98 "weather → sensory"; R = 0.99 "weather → colour"; 9 vintages, a panel of 12 people
- Domain
- Winemaking and laboratory
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Research
- Region
- Australia and New Zealand
- Country
- Australia
- Years
- 2020
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 129 Review "from volatile profile to sensory", UTAD
Aggregates published accuracy of "chemistry → sensory" models Research 2025 confidence A A
- Operator
- Peer-reviewed review
- Technology
- Summary of PLSR, SVR, deep networks, LDA
- What it does
- Aggregates published accuracy of "chemistry → sensory" models
- Results
- SVR up to R = 0.96 (correlation coefficient, not R²); deep networks R² above 0.96; LDA classification above 97%; TCA detection limit for the electronic nose 1.4 ng/l — below the human threshold
- Domain
- Winemaking and laboratory
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Research
- Region
- Italy, Spain, Portugal
- Country
- Portugal
- Years
- 2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 130 Review of isotope analysis + ML
Tracing geographical origin and detecting adulteration Research 2025 confidence A A
- Operator
- Peer-reviewed review
- Technology
- δ13C, δ2H, δ18O, 87Sr/86Sr + neural networks, Random Forest, PLS-DA, SVM
- What it does
- Tracing geographical origin and detecting adulteration
- Results
- 64–91.2% is the range from a single study on spirits, not a summary across many works. The threshold of "no less than 90% at country level" is an adequacy criterion recommended by the review, not an achieved result.
- Domain
- Winemaking and laboratory
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Research
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- China / New Zealand
- Years
- 2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 131 Conesa Celdrán et al., Miguel Hernández University
Discrimination of Rioja grape varieties: graciano, garnacha, tempranillo, mazuelo Research 2022 confidence B B
- Operator
- Peer-reviewed study
- Technology
- 8 gas sensors on an Arduino Nano + PCA and k-means
- What it does
- Discrimination of Rioja grape varieties: graciano, garnacha, tempranillo, mazuelo
- Results
- 100% accuracy in clustering, but N = 21 analyses — the sample is too small for conclusions.
- Domain
- Winemaking and laboratory
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Research
- Region
- Italy, Spain, Portugal
- Country
- Spain
- Years
- 2022
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 132 Vismara et al., LIRMM Montpellier
Multi-criteria optimisation of blend composition under oenological constraints Research 2015 confidence B B
- Operator
- Peer-reviewed study
- Technology
- Constraint programming, branch and bound
- What it does
- Multi-criteria optimisation of blend composition under oenological constraints
- Results
- A demonstration of scaling on real problems
- Domain
- Winemaking and laboratory
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Research
- Region
- France
- Country
- France
- Years
- 2015
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 133 Review "Precision Enology", University of Córdoba
Lists the equipment available on the market Research 2026 confidence A A
- Operator
- Peer-reviewed review
- Technology
- Catalogue of commercial fermentation monitoring systems
- What it does
- Lists the equipment available on the market
- Results
- Names Winegrid Wineplus 1110 (Portugal), Enartis B-evolution (Italy), Precision Fermentation BrewIQ, Anton Paar 5100 (Austria).
- Domain
- Winemaking and laboratory
- Technology class
- Sensors and IoT
- Stage
- Research
- Region
- Italy, Spain, Portugal
- Country
- Spain
- Years
- 2026
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 134 Tastry
Matches a wine's chemical profile to an individual buyer's taste Commercial operation 2016–2024 confidence C C
- Operator
- Small and medium wineries, retail; the consumer service BottleBird
- Technology
- Laboratory chemical analysis + ML matching against a database of consumer preferences
- What it does
- Matches a wine's chemical profile to an individual buyer's taste
- Results
- The company claims accuracy above 92% and growth in gross sales at retailers of up to 20%.
- Domain
- Winemaking and laboratory
- Technology class
- Recommender systems
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States
- Years
- 2016–2024
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 135 Enologix
Prediction of the optimal harvest date, the critic score and blend composition Stage not established 1989–2006 confidence C C
- Operator
- Beaulieu Vineyard, Cakebread Cellars, Ridge Vineyards, Joseph Phelps, Peter Michael, Diamond Creek
- Technology
- Correlation of "chemistry → critic score" from phenolic and terpene profiles
- What it does
- Prediction of the optimal harvest date, the critic score and blend composition
- Results
- The methodology was contested by the academic community, including Roger Boulton of UC Davis. According to the company, the critic score is matched to within 2.5 points in 95% of cases; a database of 50,000+ wines by 2005.
- Domain
- Winemaking and laboratory
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Stage not established
- Region
- US and Canada
- Country
- United States
- Years
- 1989–2006
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The figures are given according to the company.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 136 Bruker NMR Wine-Profiling 4.0
Detection of adulteration and mislabelling Scaled 2020–2021 confidence A A
- Operator
- Certification laboratories
- Technology
- NMR fingerprinting against a large reference database
- What it does
- Detection of adulteration and mislabelling
- Results
- The method was included in the OIV Compendium of International Methods of Analysis of Wines and Musts in 2021.
- Domain
- Winemaking and laboratory
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Scaled
- Region
- Germany, Austria, Switzerland
- Country
- Germany → global
- Years
- 2020–2021
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 137 Oritain
Origin verification via a QR code on the bottle Commercial operation 2022 confidence B B
- Operator
- Pyramid Valley Winery
- Technology
- Trace-element and isotopic "origin fingerprint" + statistical models
- What it does
- Origin verification via a QR code on the bottle
- Results
- Partnership since 2022, starting with the 2020 harvest
- Domain
- Winemaking and laboratory
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Commercial operation
- Region
- Australia and New Zealand
- Country
- New Zealand → global
- Years
- 2022
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 138 M&Wine
A wine "passport" to protect against counterfeits; a blind test against sommeliers was run Pilot 2021–2023 confidence B B
- Operator
- Producers and négociants
- Technology
- Analysis of a wine's multi-mineral signature
- What it does
- A wine "passport" to protect against counterfeits; a blind test against sommeliers was run
- Results
- €400,000 raised in 2023, a French Tech grant of €78,000; 7,000+ bottles analysed by April 2023, the target is 50,000.
- Domain
- Winemaking and laboratory
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Pilot
- Region
- France
- Country
- France (Lyon)
- Years
- 2021–2023
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 139 M.A. Silva "Bionic Eye"
Detection of TCA risk, cracks, contamination and insect tunnels Commercial operation 2026 confidence A A
- Operator
- Cork stopper producer
- Technology
- Computer vision, 12 checks on every cork
- What it does
- Detection of TCA risk, cracks, contamination and insect tunnels
- Results
- Consistency of human inspection ~75% against ~100% for the system; throughput up to 40,000 corks per hour; every check is logged independently.
- Domain
- Winemaking and laboratory
- Technology class
- Computer vision and deep learning on images
- Stage
- Commercial operation
- Region
- Italy, Spain, Portugal
- Country
- Portugal → global
- Years
- 2026
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 140 DTWINE
Real-time simulation and optimisation of fermentation Pilot 2021–2024 confidence A A
- Operator
- IATA-CSIC, IIM-CSIC, Bodega Ramón Bilbao
- Technology
- Digital twin of fermentation on 30-litre sensor-equipped vessels
- What it does
- Real-time simulation and optimisation of fermentation
- Results
- €1m budget, 36 months; the experimental winery covers 4 Spanish wine regions.
- Domain
- Winemaking and laboratory
- Technology class
- Sensors and IoT
- Stage
- Pilot
- Region
- Italy, Spain, Portugal
- Country
- Spain
- Years
- 2021–2024
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 141 WINE-PRO (HIGHFIVE)
Optimisation of the fermentation temperature regime Pilot 2023–2024 confidence A A
- Operator
- Puklavec Family Wines
- Technology
- Automatic refractometers + digital twin
- What it does
- Optimisation of the fermentation temperature regime
- Results
- A 3% reduction in energy consumption
- Domain
- Winemaking and laboratory
- Technology class
- Sensors and IoT
- Stage
- Pilot
- Region
- Eastern Europe, Balkans, Caucasus, Greece, Russia
- Country
- Slovenia
- Years
- 2023–2024
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 142 Tailored-Wine (IIM-CSIC)
Prediction and control of the fermentation outcome Research 2026 confidence B B
- Operator
- Early-stage research
- Technology
- Hybrid mechanistic + AI models of yeast metabolism
- What it does
- Prediction and control of the fermentation outcome
- Results
- No results yet.
- Domain
- Winemaking and laboratory
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Research
- Region
- Italy, Spain, Portugal
- Country
- Spain
- Years
- 2026
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 143 Parsec Srl "Sinergia" (→ Enartis)
Micro-oxygenation, fermentation, selective extraction Closed, acquired or wound down 2025 confidence A A
- Operator
- Wineries in 30+ countries
- Technology
- Process control from sensor data
- What it does
- Micro-oxygenation, fermentation, selective extraction
- Results
- In operation since 1995; bought by Enartis in October 2025.
- Domain
- Winemaking and laboratory
- Technology class
- Sensors and IoT
- Stage
- Closed, acquired or wound down
- Region
- Italy, Spain, Portugal
- Country
- Italy
- Years
- 2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 144 Vivelys Scalya (OENEO) borderline
Control of the cap, the temperature and extraction Commercial operation — confidence C C
- Operator
- Customers of the OENEO group
- Technology
- Sensor-driven automation of fermentation and maceration
- What it does
- Control of the cap, the temperature and extraction
- Results
- An ML component is not confirmed in public materials.
- Domain
- Winemaking and laboratory
- Technology class
- Sensors and IoT
- Stage
- Commercial operation
- Region
- France
- Country
- France
- Years
- —
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- borderline case
Why it is classified this way Sensor-based control of fermentation; the vendor does not claim a learning component. Not AI: 12 of 299. They stay in the corpus as context rather than as examples, and the classification filter removes them in one click.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 145 Pellenc Integral'Vision
Berry by berry at intake Scaled — confidence A A
- Operator
- Wineries using Pellenc equipment
- Technology
- Camera-based optical sorting
- What it does
- Berry by berry at intake
- Results
- Up to 2,000 berries per second at a throughput of up to 12 t/h, run by a single operator
- Domain
- Winemaking and laboratory
- Technology class
- Computer vision and deep learning on images
- Stage
- Scaled
- Region
- France
- Country
- France
- Years
- —
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Caveat The vendor describes the system as a sorting programme and does not call it AI.
Why it is classified this way Computer vision on a sorting line — AI under the broader definition.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 146 Ferrari Trento
Quality control of grapes from 700+ supplier estates Commercial operation 2021 confidence B B
- Operator
- Ferrari Trento (Trentodoc)
- Technology
- Deep learning on the intake line, QR-tagged crates
- What it does
- Quality control of grapes from 700+ supplier estates
- Results
- SMAU Innovation Award 2021; accuracy not disclosed.
- Domain
- Winemaking and laboratory
- Technology class
- Computer vision and deep learning on images
- Stage
- Commercial operation
- Region
- Italy, Spain, Portugal
- Country
- Italy
- Years
- 2021
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 147 Air Mixing by Parsec
Self-regulating control of temperature and density through an app Commercial operation 2023–2025 confidence C C
- Operator
- Nieto Senetiner, Cadus Wines (Luján de Cuyo)
- Technology
- Sensor-based control of pump-over with compressed air
- What it does
- Self-regulating control of temperature and density through an app
- Results
- According to the company, fermentation completes in 7–10 days.
- Domain
- Winemaking and laboratory
- Technology class
- Sensors and IoT
- Stage
- Commercial operation
- Region
- South America and South Africa
- Country
- Argentina
- Years
- 2023–2025
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat There is no independent verification of the figure.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 148 Enogis / Apra
Prediction of the harvest window Commercial operation 2024 confidence C C
- Operator
- Italian wineries
- Technology
- Predictive model on Brix, acidity, phenolics, climate and vintage history
- What it does
- Prediction of the harvest window
- Results
- Launched at SIMEI 2024.
- Domain
- Winemaking and laboratory
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Commercial operation
- Region
- Italy, Spain, Portugal
- Country
- Italy
- Years
- 2024
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The model's capabilities are described by the company itself; there is no independent verification.
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 149 Robotic pouring of sparkling wine with video analysis
Quantifies bubbles, foaming and foam persistence in sparkling wine Research 2014–2016 confidence A A
- Operator
- Peer-reviewed study
- Technology
- Standardised robotic pouring + video analysis
- What it does
- Quantifies bubbles, foaming and foam persistence in sparkling wine
- Results
- The results are "comparable" to standard chemometrics and a sensory panel.
- Domain
- Winemaking and laboratory
- Technology class
- Computer vision and deep learning on images
- Stage
- Research
- Region
- Australia and New Zealand
- Country
- Australia
- Years
- 2014–2016
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 150 PINOT (Weincampus Neustadt)
Digitisation of taste, aroma and appearance from berry to bottle Research с 2021 confidence B B
- Operator
- Weingut Lergenmüller
- Technology
- AI + digital sensory system
- What it does
- Digitisation of taste, aroma and appearance from berry to bottle
- Results
- €2.9m of BMEL funding
- Domain
- Winemaking and laboratory
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Research
- Region
- Germany, Austria, Switzerland
- Country
- Germany
- Years
- с 2021 as the source has it: “2021–”
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 151 SmartGrape
Compact mobile device for assessing grape quality and detecting fungal contamination Research с 2021 confidence B B
- Operator
- Pressing stations
- Technology
- Mid-infrared spectroscopy + machine learning
- What it does
- Compact mobile device for assessing grape quality and detecting fungal contamination
- Results
- €1.2m in BMEL funding
- Domain
- Winemaking and laboratory
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Research
- Region
- Germany, Austria, Switzerland
- Country
- Germany
- Years
- с 2021 as the source has it: “2021–”
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 152 Codorníu (digital twin, Eurecat)
Simulation of production processes Pilot 2026 confidence B B
- Operator
- Codorníu
- Technology
- AI twin of production
- What it does
- Simulation of production processes
- Results
- Shown at Wine Innovation Week 2026; no results.
- Domain
- Winemaking and laboratory
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Pilot
- Region
- Italy, Spain, Portugal
- Country
- Spain
- Years
- 2026
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 199 Changyu digital transformation
Traceability, loyalty management, tablet-controlled fermentation Scaled 2012–2024 confidence B B
- Operator
- Changyu (Yantai)
- Technology
- Blockchain traceability, a customer data platform, an “unmanned” workshop
- What it does
- Traceability, loyalty management, tablet-controlled fermentation
- Results
- More than 200m bottles on the blockchain by 2023; 19 lines, 140 tanks under tablet control; 2.8m loyalty programme members; ~20m yuan a year saved on countering counterfeits; 2023 revenue +11.89%, net profit +24.2%
- Domain
- Winemaking and laboratory
- Technology class
- Sensors and IoT
- Stage
- Scaled
- Region
- China, Japan, Korea
- Country
- China
- Years
- 2012–2024
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 207 NMR authentication programme for Hungarian wines
A reference database of Hungarian wines, including Tokaj wines, for verifying origin, grape variety and vintage Commercial operation 2017 confidence B B
- Operator
- Hungarian Ministry of Agriculture, Bruker, Diagnosticum
- Technology
- NMR + chemometric fingerprinting
- What it does
- A reference database of Hungarian wines, including Tokaj wines, for verifying origin, grape variety and vintage
- Results
- A two-year sample collection window; joins the databases for Spain, Italy, France, Chile, Austria and Germany.
- Domain
- Winemaking and laboratory
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Commercial operation
- Region
- Eastern Europe, Balkans, Caucasus, Greece, Russia
- Country
- Hungary
- Years
- 2017
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 212 Domaine Aubert & Mathieu
ChatGPT determined the blend (organic Syrah and Grenache), the wine's name “The End”, the bottle shape and suggested a price, which the estate adjusted Pilot 2023 confidence A A
- Operator
- OpenAI ChatGPT
- What it does
- ChatGPT determined the blend (organic Syrah and Grenache), the wine's name “The End”, the bottle shape and suggested a price, which the estate adjusted
- Results
- Price: according to Vitisphere “around twenty euros”; a run of 600 bottles. The first wine publicly attributed to AI authorship.
- Domain
- Winemaking and laboratory
- Technology class
- Large language models and generative AI
- Stage
- Pilot
- Region
- France
- Country
- France, Languedoc
- Years
- 2023
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Caveat The price is given according to Vitisphere; the claimed €29.90 could not be confirmed, and the first source is unavailable.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 220 Trinchero Family Estates
Modernisation of the data systems for production, bottling and e-commerce; centralised analytics of resource consumption Commercial operation 2022 confidence A A
- Operator
- Hewlett Packard Enterprise (GreenLake)
- What it does
- Modernisation of the data systems for production, bottling and e-commerce; centralised analytics of resource consumption
- Results
- Application response time cut from 5–10 minutes to seconds.
- Domain
- Winemaking and laboratory
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Commercial operation
- Region
- US and Canada
- Country
- United States, Napa
- Years
- 2022
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 221 Accolade Wines
Prescriptive analytics optimises bottling-line schedules: changeovers, sequencing, stock Commercial operation 2017 confidence C C
- Operator
- Ailytic
- What it does
- Prescriptive analytics optimises bottling-line schedules: changeovers, sequencing, stock
- Results
- The vendor claims up to a 30% reduction in cycle time, but the figure relates to a different client in the sector.
- Domain
- Winemaking and laboratory
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Commercial operation
- Region
- Australia and New Zealand
- Country
- Australia
- Years
- 2017
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat “Up to 30%” is a vendor claim: in the source the figure is tied to a different client of theirs, not to Accolade.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 222 Angove Family Winemakers
The same production planning system Commercial operation 2017 confidence C C
- Operator
- Ailytic
- What it does
- The same production planning system
- Results
- The source gives no figures.
- Domain
- Winemaking and laboratory
- Technology class
- Optimisation, planning, demand forecasting
- Stage
- Commercial operation
- Region
- Australia and New Zealand
- Country
- Australia
- Years
- 2017
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 246 AMIC — interpretable model of wine reviews
Predicts the score from chemical parameters and explains which words of the review move the rating Research 2025 confidence A A
- Operator
- Southern Methodist University (Jing Cao)
- What it does
- Predicts the score from chemical parameters and explains which words of the review move the rating
- Results
- Accuracy 89.26% with full interpretability. The words “stained” and “carpet” turned out to be positive predictors, “quick” and “breezy” negative.
- Domain
- Winemaking and laboratory
- Technology class
- Large language models and generative AI
- Stage
- Research
- Region
- US and Canada
- Country
- United States
- Years
- 2025
- Confidence
- A — peer-reviewed publication, official EU/ministry report or independent press with figures
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 260 Cerrion at Verallia
Video analytics recognises jams, fallen bottles and line anomalies, and raises an alert automatically Commercial operation 2025–2026 confidence B B
- Operator
- The Verallia glassworks at Pescia
- What it does
- Video analytics recognises jams, fallen bottles and line anomalies, and raises an alert automatically
- Results
- Incident response time reduced by up to 50%.
- Domain
- Winemaking and laboratory
- Technology class
- Computer vision and deep learning on images
- Stage
- Commercial operation
- Region
- Italy, Spain, Portugal
- Country
- Italy (the Pescia plant)
- Years
- 2025–2026
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 265 Cork Supply Legacy Cork
Analyses the internal structure of every cork, predicting oxygen permeability Commercial operation 2025 confidence C C
- Operator
- The DS100 inspection line
- What it does
- Analyses the internal structure of every cork, predicting oxygen permeability
- Results
- According to the company, ~14.5m corks a year; about 222 man-hours a year — a saving on internal transport, not an overall saving of labour
- Domain
- Winemaking and laboratory
- Technology class
- Computer vision and deep learning on images
- Stage
- Commercial operation
- Region
- Global platforms, cross-border projects and countries outside the groups
- Country
- United States, Portugal
- Years
- 2025
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Caveat The figures are given as claimed by the company and are not independently confirmed.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 266 M.A. Silva Onebyone borderline
Gas-phase spectroscopy and predictive models developed with the University of Aveiro, for piece-by-piece screening for off-notes Commercial operation 2025 confidence C C
- Operator
- M.A. Silva's lines
- What it does
- Gas-phase spectroscopy and predictive models developed with the University of Aveiro, for piece-by-piece screening for off-notes
- Results
- The source gives no figures.
- Domain
- Winemaking and laboratory
- Technology class
- Chemometrics, spectroscopy and ML in the laboratory
- Stage
- Commercial operation
- Region
- Italy, Spain, Portugal
- Country
- Portugal
- Years
- 2025
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- borderline case
Why it is classified this way Predictive models over gas-phase spectroscopy, developed with a university; the details are not disclosed. Not AI: 12 of 299. They stay in the corpus as context rather than as examples, and the classification filter removes them in one click.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 267 Amorim: optical cork sorting
Classification of corks from thousands of images of the body and the head Commercial operation 2025 confidence C C
- Operator
- Amorim's lines
- What it does
- Classification of corks from thousands of images of the body and the head
- Results
- The source gives no figures.
- Domain
- Winemaking and laboratory
- Technology class
- Computer vision and deep learning on images
- Stage
- Commercial operation
- Region
- Italy, Spain, Portugal
- Country
- Portugal
- Years
- 2025
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Why it is classified this way Classification of corks from images — AI under the broader definition.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 268 Winegrid (WATGRID) borderline
Density, temperature, colour and turbidity sensors in tanks, barrels and presses, with automatic event detection Commercial operation с 2018 confidence B B
- Operator
- Producers, including the Sogrape group
- What it does
- Density, temperature, colour and turbidity sensors in tanks, barrels and presses, with automatic event detection
- Results
- Marketing claims more than 150m bottles produced using the system; an EU grant of €50,000.
- Domain
- Winemaking and laboratory
- Technology class
- Sensors and IoT
- Stage
- Commercial operation
- Region
- Italy, Spain, Portugal
- Country
- Portugal
- Years
- с 2018 as the source has it: “2018–”
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- borderline case
Why it is classified this way Automatic event detection in sensor time series; the depth of the model is not disclosed. Not AI: 12 of 299. They stay in the corpus as context rather than as examples, and the classification filter removes them in one click.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 269 Della Toffola membrane press borderline
Optimisation of the pressing cycle from weight, flow and colorimetry sensors, with separation of must fractions by quality Commercial operation с 2021 confidence C C
- Operator
- Wineries
- What it does
- Optimisation of the pressing cycle from weight, flow and colorimetry sensors, with separation of must fractions by quality
- Results
- The source gives no figures.
- Domain
- Winemaking and laboratory
- Technology class
- Sensors and IoT
- Stage
- Commercial operation
- Region
- Italy, Spain, Portugal
- Country
- Italy
- Years
- с 2021 as the source has it: “2021–”
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- borderline case
Why it is classified this way Optimisation of the pressing cycle from sensors; the “neural network” is mentioned in a single interview and is not confirmed on the product pages. Not AI: 12 of 299. They stay in the corpus as context rather than as examples, and the classification filter removes them in one click.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 270 Oculyze Fermentation Wine
Automatic counting of yeast cells, viability and budding in place of a manual haemocytometer Commercial operation — confidence B B
- Operator
- Wineries
- What it does
- Automatic counting of yeast cells, viability and budding in place of a manual haemocytometer
- Results
- “Ten times faster than manual counting”; range 8.5×10⁵–3.5×10⁷ cells/ml
- Domain
- Winemaking and laboratory
- Technology class
- Computer vision and deep learning on images
- Stage
- Commercial operation
- Region
- Germany, Austria, Switzerland
- Country
- Germany
- Years
- —
- Confidence
- B — trade press or an official company statement with verifiable details
- Classification
- AI
Caveat The vendor describes the technology as image recognition and claims neither AI nor machine learning.
Why it is classified this way Image recognition — AI under the broader definition.
Not stated in the source technology, dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
No. Case 292 Kuban-Vino
Detection of defects during bottling Research 2025 confidence C C
- Operator
- Kuban-Vino
- Technology
- Machine vision on the bottling line + an internal AI assistant for documents
- What it does
- Detection of defects during bottling
- Results
- Launch is claimed for the third quarter of 2025; there are no results.
- Domain
- Winemaking and laboratory
- Technology class
- Computer vision and deep learning on images
- Stage
- Research
- Region
- Eastern Europe, Balkans, Caucasus, Greece, Russia
- Country
- Russia
- Years
- 2025
- Confidence
- C — vendor marketing claim without independent confirmation
- Classification
- AI
Not stated in the source dataset size. The field was not lost in parsing — the cited material does not give it. No technology: 57 of 299. No dataset: 281 of 299. A record is not incomplete for saying only what its source says.
None of the 299 cases meets every condition at once.
The intersection is empty, but each condition on its own gives something. Below is how many cases come back if one of them is lifted.