Casebook

The Casebook is a register of audited deployments of artificial intelligence in viticulture, winemaking and the wine business. It holds 299 records, and every one of them was checked against its original source and graded for how reliable that source is.

The corpus was assembled in August 2026 in three waves: the main regions and domains, then the regions and topics those missed, then the Russian market and the industry’s business edges — insurance, logistics, regulators. 177 records are about the vineyard, 47 about the cellar and the laboratory, 75 about business, marketing and hospitality. The tilt towards the vineyard is not an editorial choice: a satellite, a drone and a weather station produce thousands of observations a season; fermentation, dozens a year. The keynote on AI in wine sets out what follows from that asymmetry: what already works, and what becomes normal within five years.

What counted as a case

Three conditions, all required

Wine

The record concerns wine, wine grapes or the wine business.

01

An algorithm

It has an algorithmic component.

02

A source

It is backed by a source that was actually opened and read.

03

We drew the boundary around AI wide and wrote it out in full, in both directions: the industry uses the word loosely, and an unstated definition would have made the catalogue impossible to check.

The boundary does not fall where the industry puts it. Seven well-known systems sit outside it though described as AI almost everywhere: Amorim NDtech’s gas chromatography, Anton Paar’s inline refractometry, Selinko’s cryptographic tag, eProvenance’s fixed rules, Sentia’s biosensor strips, the IWCA emissions calculator and WinePulse’s BI reporting. None learns and none predicts. The reverse also happens — the word “AI” comes from a journalist rather than the company: neither “artificial intelligence” nor “machine learning” appears once on Oculyze’s or Vinescapes’ own pages for the products in the catalogue, and the first of the two does use AI by our definition.

Under that frame 287 records are AI cases proper. A further 5 are borderline, and 7 describe adjacent technologies with no AI component; they are kept and labelled, because the absence of AI where the industry claims it tells the reader as much as its presence.

How to read a record

Two fields should be read before the rest, and both are filters in the register.

Source confidence

A

A peer-reviewed publication, an official EU or ministry report, or independent press with figures.

109

B

Trade press or a company’s own announcement with checkable detail.

110

C

A vendor marketing claim with no independent confirmation; read those figures as “according to the company”.

61

Not stated

The peer-reviewed research section, where the paper is the source.

19

The confidence level rates the source, not the case: the higher the grade, the better established what the record says.

Maturity stage

Research

In the laboratory or the literature.

83

Pilot

Running on a real site, not yet in production.

54

Commercial operation

Sold and running.

129

Scaled

Beyond the first estate or the first market.

18

Closed

Ended, acquired or wound down.

11

Stage not established

Not a point on that ladder but the absence of one: the record’s own sources do not establish a stage. Until the audit these were filed as commercial operation beside a caveat saying exactly that.

4

The stage says where a case stands on that ladder. Commercial operation and scale together are just under half the catalogue, and the half is less solid than it sounds: 40 of those 147 rest on a grade-C source alone — running commercially on the vendor’s word and nothing else. Most of the rest has not left the laboratory or the pilot, or has already ended. Several of those pilots are in their fifth to eighth year, a finding in itself — getting the model onto an estate takes longer than building it.

What is deliberately not here

Forecasts, market sizing and vendor rankings. The confidence level grades the source and says nothing about the product; the stage says how far a case got and nothing about how good it is. Nor are unverified leads here — a mention with no link, or a link with no grapes behind it, does not become a record, however interesting it sounds.

Absence is recorded too. Ten business edges of the industry were checked on purpose and eight came back empty.

None of the eight held a deployment with a learned model and a checkable result. There are no records for them in the catalogue.

The published data

Everything the register draws from is published as a dated release, under CC BY-SA 4.0: all 299 records in one file, in two formats.

• Casebook v2026-09-29 (JSON, 553 KB) — the records, the 5 facet vocabularies with their counts, and a description of every field in both languages, so the file can be read without this site.
• Casebook v2026-09-29 (CSV, 332 KB) — the same records and the same columns, one row each, for a spreadsheet.

Two numbers identify a record and they are not the same number. id runs from 1 to 299 without gaps, and is what this site links to; catalogId is the source catalogue’s original sequence, which reaches 307 with eight gaps where the audit deleted a record. Join on id.

Releases accumulate and none is overwritten, so a dated file keeps resolving after the corpus is next corrected: cite the file, not this page. The licence asks for attribution and a note of what you changed.

How to cite

The plain line suits a bibliography; the BibTeX entry, a reference manager.

@misc{khudov-casebook-v2026-09-29,
  author  = {Nikita Khudov},
  title   = {Vinum ex Machina Casebook: audited cases of AI in wine},
  version = {v2026-09-29},
  year    = {2026},
  url     = {https://vinumexmachina.com/casebook/casebook-v2026-09-29.json},
  note    = {CC BY-SA 4.0}
}

The same files are deposited on Zenodo with a DOI. Cite this dataset: Khudov, N. (2026). Vinum ex Machina Casebook: an audited register of 299 AI applications in viticulture, winemaking and the wine business (Version v2026-09-29) [Data set]. Zenodo. https://doi.org/10.5281/zenodo.23039821