What is Beverage-AI Radar?

Beverage-AI Radar is a curated, source-cited landscape of the companies, people, research, and case studies applying artificial intelligence and data analytics to the beer, whiskey, and wine industries.

Who is applying AI to beer, whiskey, and wine?

The radar tracks over 100 companies and individuals across beer, whiskey, and wine, from flavor prediction and sensory analytics to demand forecasting, computer vision, and generative AI, each with links to public evidence.

How do I get listed on Beverage-AI Radar?

If you build, analyse, or research AI and data for beer, whiskey, or wine, message Ankur Napa on LinkedIn or open a suggestion on the GitHub repository, and your entry will be added.

Beverage-AI Radar | AI, Data & Analytics in Beer, Whiskey & Wine

Companies & ventures Who is building it, how far along, and the evidence for each claim. People The named practitioners, most of them directly reachable. Research & resources Papers, white papers, open datasets, talks, repositories and blogs. Events Where the field gathers, in person and online, with what is coming up. Jobs Open roles applying AI and data in drinks, linked to the original posting. Rules How this is verified The evidence bar every entry is held to, and what is deliberately absent.

Entries that make no AI claim are kept and labelled as such rather than removed or inflated. A landscape is only useful if you can tell which is which.

Where they are click a country to filter

By vertical

By theme

By maturity

Where they are by their company's base · click to filter

Papers, news, case studies, open-source repositories, and talks on AI and data in beverage. Every item links to its source.

Where this industry gathers, and where Ankur has stood up in front of it. Trade fairs and technical conferences alongside his own talks, in person and online. Events whose organiser has not published dates are listed as undated rather than guessed.

Where they are click a country to filter

Open roles applying AI, data, and analytics in beer, whiskey, and wine. Swept from public job postings; every row links to the original posting.

Where they are click a country to filter

Private. Outreach targets for pitching AI, data and analytics into beer, whiskey and wine. This tab only exists on a local copy: prospects.json is gitignored, so it is never published.

Where they are click a region to filter

Why this exists

I am Ankur Napa, a brewer who learned to code. I spent years on the brewery floor and then years in data, and I kept meeting people quietly doing the same thing: using AI, machine learning, and analytics to make better beer, whiskey, and wine. We were all working in parallel and rarely finding each other.

Beverage-AI Radar is my attempt to fix that. It is one honest, source-cited map of who is applying AI and data across the drinks industry: the companies, the open-source projects, the research, the case studies, and above all the people. Every entry links to real evidence. Nothing here is invented.

The rules this dataset follows

A landscape is only worth reading if you can tell what has been checked and what has not. These are the rules every entry is held to, stated so you can hold them against the data yourself.

Evidence

Two sources minimum, one of them the company's own site. A record with no working domain does not get published. If a claim only exists in a press release and not on the vendor's own pages, the entry says so.
Inventing capability is the one unforgivable error. Marketing language is never upgraded into a machine-learning claim. Where a vendor's launch coverage said "computer vision plus an LLM" and the vendor's own methodology page said deterministic text matching, the entry follows the vendor.
Blocked is not the same as absent. Sites behind Cloudflare or built entirely in JavaScript are recorded as blocked, with the reason, rather than quietly dropped. Otherwise the data would imply those vendors publish nothing.

Companies that make no AI claim stay in

Roughly a third of every sweep turns up equipment makers, consultancies and systems of record with real operations and no AI story. They are kept and labelled no AI claim, filterable in one click. Removing them would flatter the field; inflating them would corrupt it. Both are worse than saying plainly what is there.

This rule exists because the opposite once happened here: the maturity scale had no value for a checked negative, so 114 honestly-recorded non-AI companies were published as shipping AI. The fix was a value that can express "checked, and the answer is no".

Verified against provisional

Entries found by an automated scout are marked verified: false until a human confirms them. Treat that flag as the difference between "someone checked this" and "something found this". First-party claims that are not independently evidenced carry the same flag and say so in the description.

Verticals

Beer, whiskey and wine are the subject. Non-alcoholic drinks and food are adjacent lanes, tracked because the same vendors, models and people cross over, and deliberately excluded from the coverage-gap analysis so they never pull the research away from the core three.

How the data is gathered

Curated seed. Hand-verified companies, the authoritative source.
Scout sweeps. Agents work assigned surfaces, then every find is merged through a domain check and a duplicate check before it can enter.
Resource sweeps. Peer-reviewed work through the OpenAlex API; talks from YouTube search; industry white papers from vendor listings. Each sweep applies the same gate: a record must mention a drink AND a technology, and known false friends are excluded by name.
Weekly, on a schedule. The sweeps re-run every Sunday and commit only when something actually changed.

That last gate matters more than it sounds. A search for "wine machine learning" returns almost nothing but tutorials using one teaching dataset, and "brewery machine learning" returns a coding bootcamp called The App Brewery. Both are excluded by rule, not by hand.

What is deliberately not here

No private or personal data. Contact details that appear on public posters are not republished.
No confidential material. Anything shared in a private conversation stays out, however relevant.
No tracking of you. Reading history and recommendations are computed in your own browser and never leave it.

How it helps

Find the field. See at a glance who is shipping, who is piloting, and who is researching across beer, whiskey, and wine.
Find the people. Analysts, data scientists, BI and market-intelligence leaders, and independent builders, each with a public profile.
Find the evidence. Papers, news, vendor and industry-body case studies, repositories, and talks, all in one filterable place.
Find each other. The whole point is to help people working on the same problems meet, compare notes, and build together.

Are you working in this field?

If you build, analyse, or research AI and data for beer, whiskey, or wine and you are not on this radar yet, that is my gap, not yours. I would love to add you. Reach out and I will include you.

Message me on LinkedIn to get listed Suggest an entry on GitHub

Updates

2026-07-22 Added dedicated detail pages, an About and mission page, and full SEO. Extended the tracking window to 10 years.
2026-07-22 Added a Research & resources tab: 18 papers, 26 news and case studies, 23 repositories, and 16 talks, all source-cited.
2026-07-22 Added a People discovery lane for data, BI, and market-intelligence practitioners.
2026-07-21 Rebuilt the dashboard in a Microsoft Fluent style, made the site public, and added logos, avatars, and full filters.
2026-07-20 First public build: curated company landscape with daily refresh.

Method: hand-verified seed data, refreshed by an automated pipeline. Every fact carries a source URL. Corrections and additions are welcome.

← Back to radar