Industries · How restaurants get recommended by ChatGPT and AI search
"Where should we eat tonight?" is a question people now ask AI.
Diners ask assistants for a cuisine, a vibe and a constraint: "quiet Italian place with gluten-free pasta near downtown". The restaurants that get named have that information on their site as text.
Free, no account. We only read public files.
What people ask AI assistants
- “Best tacos in [city] open late”
- “Restaurant near [venue] good for a group of 10”
- “Gluten-free friendly pizza in [neighbourhood]”
- “Romantic dinner spot in [city] that takes reservations”
Why does AI search matter for restaurants?
Restaurant decisions are fast and specific. Assistants answer by matching constraints (dietary needs, hours, price, reservations) against what they can read.
Many restaurant sites put the menu in a PDF or an image and the hours in a footer graphic. Those are exactly the facts assistants need, in the formats they read least reliably.
Review sites and Google Maps will always matter most. Your own site is where you control the facts they draw on.
What we know so far
Full report and method →We haven't scanned restaurants websites yet, so we won't quote numbers for them. Our first study covered 101 dentist, law-firm and home-service sites. The gaps below are the ones we check for on every small-business site, adapted to how restaurants are usually found.
The restaurant checklist
Things to do this week, roughly in order.
- Publish the menu as an HTML page with prices and dietary labels (vegetarian, gluten-free, nut-free).
- Put opening hours on the site as text and update them for holidays.
- Say whether you take reservations, how, and for what group size.
- Describe the place in a sentence a person would use: quiet, family-friendly, good for groups, outdoor seating.
- Add Restaurant schema with servesCuisine, priceRange, hasMenu and acceptsReservations.
- Check that Google, Yelp and your delivery apps show the same hours and phone number.
What are the common mistakes, and the fixes?
- 01
Menu only as a PDF or photo
Publish an HTML menu with dishes, prices and dietary labels. Keep the PDF as a download if you like.
- 02
No Restaurant schema
Restaurant with servesCuisine, hasMenu, priceRange, acceptsReservations and hours turns your basics into structured facts.
Free tool: Schema / JSON-LD Generator → - 03
Hours that disagree with Google
Holiday and seasonal hours drift. Check your site, Google Business Profile and delivery apps say the same thing.
- 04
Site builder bot settings
Some hosted website builders add bot protection or robots rules you never chose. Check what's actually served.
Free tool: AI Crawler Checker →
Which schema should restaurants use?
Restaurant supports servesCuisine, hasMenu, acceptsReservations, priceRange and openingHoursSpecification, which are exactly the facts people ask about.
Generate it for your restaurant →Recommended type
"@type": "Restaurant"
Start with the free check. If you'd rather hand it over, there's the Snapshot.
The free AI Crawler Checker shows in seconds whether AI crawlers can read your site. The $99 Snapshot audit goes further for one restaurant: 10 customer questions × 4 AI engines, crawler, robots.txt & CDN review, llms.txt and schema written for you, prioritised fix list (PDF). Delivered in 5 working days.
FAQ