Vegan, gluten-free, oat milk: how AI actually reads your café menu

Why the AI often doesn't read your café menu at all

Most cafés have their menu as a pretty PDF or as a photo of the chalkboard on Instagram. To humans that looks great; to a language model it is almost invisible. ChatGPT, Gemini and Perplexity draw their answers from text they can read cleanly: from HTML, from structured data, from clear descriptions. A picture of your board, or a menu that only lives inside an ordering app, simply doesn't show up in this process. The information is there, but not in a form a machine can grasp.

You may already know the result: a guest asks the AI about a vegan breakfast in your town and gets three other cafés named, even though you have the best vegan bowl in the area. Not because your offering is worse, but because the information doesn't exist for the machine. Visibility in AI isn't created by beautiful photos, but by clear, readable text on your website.

The good news: you don't need to become a tech pro. It's enough to prepare your menu properly once and to name the key attributes like vegan, gluten-free and lactose-free so clearly that no model can miss them. Remember: what is obvious to a human from context has to be spelled out explicitly for the machine. That is exactly what this guide is about.

How guests really search for your café today

Search queries have changed. People used to type "café downtown" into Google. Today the same person writes full sentences into ChatGPT: "I'm gluten intolerant, where in Regensburg can I get coffee and cake in a cosy spot?" or "Which café nearby has a vegan cappuccino with oat milk and Wi-Fi?" These questions are specific, full of conditions and tailored precisely to a need.

For you as a café this is a huge opportunity. Because someone who asks such a specific question is a very ready-to-buy guest. They aren't looking for just any café, they're looking for exactly your offer. If the AI then names your place because you've cleanly labelled gluten-free cake and oat milk, you've won a guest that a competitor with an unreadable PDF would never have reached.

The key is knowing your guests' real questions. Listen closely in the shop: what gets asked most often at the counter? "Is the banana bread vegan?", "Do you have lactose-free milk?", "Are there nuts in it?" These are exactly the questions people also ask the AI. If your website answers them in plain text, you become the source.

Vegan, gluten-free, lactose-free: name the terms unambiguously

Language models understand language remarkably well, but they hate ambiguity. If your menu only shows a small green leaf symbol next to a dish, the AI can't be sure whether that means "vegan", "vegetarian" or "organic". Spell it out. Instead of a symbol, the carrot cake needs plain text behind it: "vegan, lactose-free, gluten-free on request". Those few words are worth gold to the machine.

Avoid trade abbreviations that only you understand. "VG", "GF" or "nL" on the menu may be handy in the shop, but to a model they're riddles. Always write "vegan", "gluten-free" and "lactose-free" in full. Add the synonyms guests use too: "plant-based", "dairy-free", "wheat-free", "oat milk instead of cow's milk". The more natural phrasings appear, the more reliably the AI recognises the offer.

Also be honest and precise about limitations. If your gluten-free cake is made in the same kitchen as products containing gluten, write "gluten-free recipe, no separate kitchen". That protects your allergic guests and makes you credible as a source. Models prefer content that is precise and responsibly phrased over full-mouthed, vague promises.

Getting the menu onto your website as real text

The most important technical step is also the easiest: your menu has to be on your website as real text, not just as a PDF download and not as an image. Create a dedicated "Menu" page with headings like "Coffee and drinks", "Cakes and pastries", "Breakfast", and beneath them each product as a readable paragraph with name, short description, price and labels. This HTML text is exactly what language models read and quote.

Describe each product the way you'd explain it to a guest. Instead of just "Carrot Cake 4.50", write: "Carrot Cake, moist carrot cake with walnuts, vegan and lactose-free, gluten-free on request, 4.50 euros." That one line answers three typical guest questions at once and gives the AI immediately usable facts. A photo next to it is nice for humans, but the text does the work for your visibility.

Keep the menu up to date. If you change seasonally, update the text and not just the chalkboard in the shop. An outdated online menu advertising a pumpkin latte in July looks unreliable to guests and models alike. A well-maintained, dated menu text, by contrast, signals: this source is reliable and worth quoting.

Structured data: the invisible turbo

Besides the visible text, there's a layer guests never see but which helps search engines and AI systems enormously: structured data using the Schema.org format. For cafés, the relevant types are above all "Restaurant" or "CafeOrCoffeeShop" and "Menu". With them you record in your page's code, in machine-readable form, that you're a café, where you're located, what opening hours apply and which dishes with which dietary features you offer.

Schema.org even has a dedicated property called "suitableForDiet" with values like "VeganDiet", "GlutenFreeDiet" or "LowLactoseDiet". If your menu entries are marked up with these, the machine gets your offerings served in a perfectly unambiguous form. That's the difference between "the AI guesses" and "the AI knows". Even if you don't code this yourself: ask your web provider to do it explicitly, it's not a big effort.

Complement this data with a clean Google Business Profile. Enter the same attributes there that Google offers for hospitality: "Vegan options", "Gluten-free options", Wi-Fi, outdoor seating. Many AI systems draw on this profile data. When website, schema and profile all tell the same story, you become a reliable source.

Answer the typical guest questions directly

Language models love content built as question and answer, because that's exactly the pattern they work in. Set up a small FAQ section on your website that answers your guests' real questions in plain text. "Do you have oat milk?" – "Yes, we offer barista oat milk at no extra charge for all coffee specialities." Such a block is like a ready-made answer for the AI that it can take almost word for word.

Collect these questions from real everyday life. What are the five things most often asked at your counter? For most cafés they're about plant-based milk, gluten-free cakes, allergens, seating and dogs. Answer each of these honestly and concretely. Avoid marketing phrases like "large selection" and instead write what you really have: "three vegan cakes daily, at least one of them gluten-free".

This approach pays off twice. Your human guests find what they're looking for faster, and the AI gets quotable answers. The important thing is that the answers stand on their own and stay understandable even if a model pulls out just this one paragraph and reads it to a user.

Consistency across all channels

Nothing confuses an AI more than contradictions. If your website says you have oat milk but your Google profile doesn't, and on an old review portal someone writes you only have cow's milk, then the model doesn't know what's true. When in doubt it would rather leave your café out of the vegan recommendation. Your job is to tell the same story everywhere.

Go through all the places where your café appears online systematically, once: your own website, Google Business Profile, Instagram bio, delivery services, business directories. Check whether name, address, opening hours and above all the dietary details are the same everywhere. This consistency is one of the strongest signals you can give the AI, because agreement across many sources reads as reliability to a model.

Reviews play into this too. When guests rave in reviews about your "best gluten-free banana bread" or your "large selection of oat milk coffee", it reinforces the signals from your own website. You can happily encourage satisfied regulars to mention concretely in their review what they enjoyed. Such genuine phrasings are valuable confirmation for AI systems.

Start in three steps: your practical plan

Start small, but start. Step one: get your complete menu onto your website this week as real text, with the labels vegan, gluten-free and lactose-free spelled out behind every applicable product. This single step already puts you ahead of most cafés that still just link a PDF.

Step two: build an FAQ section with the five most common guest questions and honest answers. Step three: ask your web provider to add structured data per Schema.org, and maintain your Google Business Profile with the right dietary attributes. After that, regularly check whether all channels tell the same story.

Finally, the simplest test of all: ask ChatGPT or Gemini yourself about a vegan cake or gluten-free breakfast in your town and see whether and how your café gets named. Repeat this every few weeks. That way you see in black and white whether your work on the machine-readable menu is paying off, and where you still need to sharpen up.

Remember that this work compounds. Every guest who finds you via an AI recommendation because you cleanly labelled oat milk and gluten-free cake is a guest who would have ended up with the competition without this preparation. And because very few cafés make their menu machine-readable so far, you have a real head start right now that you can secure with just a few hours of work.