Technical & Structure · 9 min read · July 15, 2026
How to make your menu readable for AI: helping ChatGPT find your dishes and diet options
When a guest asks ChatGPT "Where can I get gluten-free pasta in the evening in Kitzbühel?", it isn't your Google ranking that decides whether you show up — it's whether the AI can actually read your menu. If your menu only exists as a PDF or an image graphic, it's invisible to the machine. This guide shows you how to prepare dishes, prices, and diet options so AI systems can find you and actively recommend you.
Why your beautiful PDF menu is invisible to AI search
Almost every restaurant today has a menu online. The problem: in most cases it's a PDF to download or a nice-looking image the designer put together. For your guests, that looks great. For ChatGPT, Google Gemini or Perplexity, it's a closed door. To a machine, an image is just a surface of pixels, and a PDF often isn't even crawled in the first place. Your beef roulade with bread dumpling at 24.50 euros simply doesn't exist in the AI's knowledge.
This is the key difference between classic SEO and Generative Engine Optimization (GEO). With Google, a PDF at least has some chance of being indexed. An AI assistant, by contrast, needs clean, structured text it can parse sentence by sentence. When a guest asks 'Which restaurant in the old town has vegan mains under 20 euros?', the AI searches through text blocks — not image files, not design artwork.
The first and most important step is therefore unglamorous but effective: your full menu needs to exist as real, readable text on your website — not as an embedded image, not as a download link, but as HTML text you can select with your mouse. Try it yourself: if you can't highlight and copy your dishes on the website with the cursor, the AI can't read them either.
Writing out diet options instead of hiding them in symbols
In restaurants, it's become standard to mark things with small symbols and footnotes: a leaf for vegetarian, a crossed-out ear of wheat for gluten-free, superscript numbers for allergens. On a printed menu that's elegant and saves space. For an AI it's a disaster — it doesn't recognize the wheat icon and has no way of knowing that a 'G' next to a dish means 'contains gluten.' The diet information simply gets lost.
Instead, spell out what's what in plain text. Instead of 'vegetable curry (v, gf),' write: 'Vegetable curry with coconut milk and basmati rice. Vegan and gluten-free.' That costs you two extra seconds, but this is the exact sentence an AI will quote when someone searches for vegan options. Add honest notes where useful, like 'can be made lactose-free on request' or 'available gluten-free on request.' AI systems pick up on phrasing like this readily.
Think in the actual search terms your guests use. People ask for 'gluten-free,' 'wheat-free,' 'suitable for celiac disease,' 'vegan,' 'no animal products,' 'lactose-free,' 'low carb,' or 'keto.' The more of these natural terms actually appear as words in your text, the more reliably the AI recognizes what you offer. A single icon covers none of these search terms.
Structured data: the language machines actually read
There's a technical standard built specifically for restaurants: Schema.org, specifically the types 'Menu,' 'MenuItem,' and 'Restaurant.' This is invisible code running in the background of your website that labels each dish clearly: this is the name, this is the description, this is the price, these are the diet properties. Google and AI systems favor this structure because they don't have to guess — they get the facts served up cleanly.
The Menu schema even has a dedicated field called 'suitableForDiet.' There you can record, using standardized values like 'VeganDiet,' 'GlutenFreeDiet,' or 'LowLactoseDiet,' which diet a dish fits. If your website builder or agency implements this, you have a real edge over the place down the street that only uploaded a PDF. Ask whoever maintains your site about it specifically.
Consistency matters here: the prices, hours, and dishes in your schema code have to match exactly what the guest sees as visible text. If the two contradict each other, the AI gets suspicious and, when in doubt, discards the information altogether. So maintain both in parallel, and update the code whenever you change the menu.
Write descriptions that answer questions, not just name dishes
Many menus are just terse listings: 'Wiener Schnitzel 18.90.' For a regular, that's enough. For an AI, it's thin. Assistant systems favor dishes they can describe, because that lets them give the user a real answer. So add an explanatory sentence: 'Veal Wiener Schnitzel, pan-fried in clarified butter, served with parsley potatoes and lingonberries. A generous portion for a big appetite.' One line turns a listing into usable knowledge.
Good descriptions answer the questions guests are silently asking: where does the meat come from? Is it spicy? How big is the portion? Is it meant for sharing? When you name the regional origin, the preparation method, and typical sides, you're giving the AI exactly the details it needs to set your restaurant apart from everyone else's. 'Trout from our own pond' beats any anonymous 'fish dish.'
Still, stick to the truth. Don't invent awards or organic certifications you don't actually have. AI systems increasingly cross-check details against other sources, and false claims can hurt your visibility over the long run. Authentic, concrete descriptions come across as more credible than inflated claims like 'the best schnitzel in town.'
Hours, reservations, and the context around the meal
'Where can I eat tonight?' isn't purely a menu question. The AI also wants to know whether you're actually open, whether reservations are needed, and whether you fit the guest's situation. That's why an AI-readable menu also needs current opening hours in plain text, a note on reservations, and context like 'terrace,' 'child-friendly,' 'groups up to 20,' or 'quiet corner for business meals.'
Think about the compound searches typical in your industry: 'restaurant with a terrace and vegan options,' 'venue for a family celebration with a kids' menu,' 'where can I get hot food on Sundays until 10 p.m.' Each of these combines food with context. If your text covers these combinations, you get recommended for exactly these purchase-ready queries. Keep your Sunday and holiday hours especially current.
Don't forget consistency across every channel. Your hours on the website, on your Google Business Profile, on TripAdvisor, and on reservation platforms should all match. Conflicting hours are a common reason AI systems get uncertain and choose to recommend a competitor with unambiguous details instead.
Google Business Profile and reviews as amplifiers
AI assistants don't draw their knowledge about your restaurant only from your website. They lean heavily on your Google Business Profile, review platforms, and mentions in blogs or press. A well-maintained Google profile — with the right category, current photos, an uploaded menu, and answered reviews — is a direct lever for your AI visibility. It's the second menu the machine reads.
Reviews are worth their weight in gold here, because guests use the exact natural-language terms other people search for. When several reviews mention 'great vegan selection' or 'finally, gluten-free pizza,' that anchors your offering in the AI's knowledge far more credibly than anything you could say about yourself. Don't hesitate to ask happy guests for an honest review and to mention what brought them in.
Respond to reviews too, especially the critical ones. When you write back, 'Thanks for the feedback — our menu now includes three fully plant-based mains,' you're producing fresh, current text with exactly the right keywords. These exchanges get read and factor into your profile.
Frequent mistakes that make you invisible to the AI
The classic mistake is a pure image or PDF menu with no text alternative. Nearly as damaging are menus buried in Flash-style sliders, third-party JavaScript widgets, or behind a 'view menu' button that a crawler never clicks. If content only appears after a user interaction, the AI often sees nothing but an empty page. Check whether your dishes actually appear in the page's source code.
A second common mistake is an outdated menu. Nothing frustrates a guest more than an AI recommending a dish you dropped a year ago, or quoting a price that's long since gone up. Outdated details erode trust in the recommendation. Build a fixed habit: every time the menu changes, the website gets updated at the same time, not weeks later.
The third mistake is conflicting information across channels: 24.50 euros on the website, 27 euros on the delivery platform, completely different hours on the reservation site. Contradictions like this don't just confuse guests — they make the AI hesitate between sources, and when in doubt, it will favor a restaurant with clear, consistent data.
Your roadmap: becoming AI-readable in five steps
Start with what matters most: turn your PDF or image menu into real website text. Every dish with a name, an honest description, and a price, all as selectable text. That alone puts you ahead of most competitors who never take this step. Next, add diet details to each dish in plain text, written out instead of shown as a symbol.
Third, get your website set up with Schema.org data for menu and restaurant — ask your agency, or check whether your website builder already supports it. Fourth, bring your Google Business Profile in line with the same information: same menu, same hours, same prices. Fifth, build an update routine so everything stays current whenever the menu changes.
The good news: almost everything that makes your menu AI-readable also helps your human guests. Clear descriptions, honest diet information, and current prices build trust with both audiences. So you're not optimizing for cold machines — you're simply making your food visible wherever people are searching for their next meal today.
Common questions
Isn't it enough to just have my menu on the website as a PDF?
No. AI systems often don't read PDFs at all, and image graphics are invisible to a machine. Your dishes need to exist as real, selectable HTML text on the website. Try it yourself: can you highlight and copy a dish with your cursor? If not, the AI can't read it either, and it won't recommend you for related searches.
How do I make gluten-free and vegan dishes recognizable to ChatGPT?
Spell out diet properties in plain text instead of hiding them in symbols or footnotes — 'vegetable curry, vegan and gluten-free' instead of 'vegetable curry (v, gf).' Use the actual terms your guests search for, like 'gluten-free,' 'wheat-free,' 'vegan,' or 'lactose-free.' Ideally, also add the Schema.org field 'suitableForDiet' in your website's background code.
Do I need technical skills or an agency to do this?
The most important step — getting dishes onto the website as real text with descriptions and diet details — you can usually handle yourself in most website builders. For the structured Schema.org data, a quick request to whoever maintains your site often does it. Maintaining your Google Business Profile and answering reviews takes time, not technical skill, and pays off immediately.
Read on