AI readiness for specialty food and drink producers

Allergens, provenance and shipping windows — three facts, all in a PDF.

M4C3E3C2E2

The moment

Callum, who roasts coffee and sells direct

A buyer asks an assistant for a decaf single origin, ground for espresso, that ships within the country before the weekend.

What the agent needs: Origin, process, grind options, allergen and dietary detail, and the shipping cut-off.

What actually happens: Origin and process live in a tasting-notes graphic, grind is a variant with no readable label, and the shipping cut-off is in a policy page written for humans. The assistant recommends a roaster whose product page answers all five.

Which dimensions decide it, and why

AIScan grades five dimensions. These are the ones that carry the outcome for this kind of business.

Commerce

Variants are the trap. Price per size and per grind must be readable per variant, or an assistant quotes the cheapest option and misrepresents you.

Content

Provenance, allergens and dietary claims are exactly the attributes buyers filter on and exactly the ones producers put into artwork. As text with Product structured data, they become the reason you are picked.

Capabilities

Shipping cut-offs, subscription intervals and stock notifications are actions. A machine-readable description of what your site can do turns a browsing agent into a buying one.

The checks that matter here

Fix guides: /docs/checks/commerce · /docs/checks/content · /docs/checks/capabilities

Most sites like this run on Shopify — read the Shopify playbook for where each fix lives.

Questions we get asked

Do allergens need to be in the HTML?

For anyone filtering on them, yes — and buyers with allergies filter first and read second. Text plus structured data, never only an image of a label.

How should subscriptions be described?

As an offer with a price and an interval, in words. 'From £X every four weeks' is quotable; a slider is not.

We sell wholesale too. Same rules?

Publish what is public — minimum order, lead time, whether you supply trade — and keep the price list behind a form if you must. Silence on all four is what costs you the enquiry.

Is a tasting-notes graphic really a problem?

Only as the sole copy. Repeat the notes as text below it and you keep the design and gain the query.

Does provenance help with AI answers specifically?

It is what makes you distinguishable. Assistants summarising 'best decaf single origin' need attributes to justify a pick, and provenance is the attribute you own.

What should we check first?

Scan one product URL and read checks M4 and C3. Those two cover price, variants and structured product data.

Scan your site and see where you actually stand

Free, no account, about twenty seconds. The report names every failing check by ID, shows the evidence we found, and gives the fix for your platform — plus a hand-off prompt you can paste straight into Claude Code or Cursor.

Run a free scan npx aiscan-cli yoursite.com

Prefer to read first? Browse every check we run or the guide library.

Next in Retail & ecommerceAI readiness for fashion and apparel stores

Also covers: coffee roaster, tea merchant, bakery, chocolatier, cheesemonger, butcher, brewery, distillery, winery, hot sauce, olive oil, spice blends, artisan food, farm produce, meal kits, subscription box, direct to consumer food, allergen information.