Restaurant discovery is increasingly shaped by AI answers. Most independent restaurants are not structured for them.
When someone asks ChatGPT where to eat in your city, or Google AI surfaces dining recommendations in search results, the restaurants that appear have Restaurant schema, allowed AI crawlers, and structured menus. The ones that do not appear have none of these things.
AI search cannot recommend a restaurant it cannot classify.
Restaurant schema requires more than a business name and address. The servesCuisine property tells AI systems what food the restaurant serves. The hasMenu property links to the actual menu. acceptsReservations tells AI whether booking is required. Without these properties, AI systems either skip the restaurant or describe it incorrectly.
The commercial impact is practical rather than abstract. If AI cannot tell what cuisine you serve, whether reservations are required, whether a menu is live, or what hours you keep, it is more likely to recommend a restaurant whose site answers those questions cleanly.
Restaurant pages are especially well suited to structured FAQ and answer-first content because diners repeatedly ask the same pre-visit questions: do you take reservations, do you accommodate dietary restrictions, is there parking, do you have outdoor seating, and what is on the menu. Clear answers improve extraction quality even when they do not create a special search feature.
Where restaurant pages usually fail AI extraction
Restaurant schema, answer readiness, and AI access.
- vservesCuisine ("Italian", "Farm-to-Table", "Japanese")
- vhasMenu (URL to menu page)
- vacceptsReservations (Boolean or URL to reservation system)
- vopeningHoursSpecification (full day/time objects)
- vpriceRange ($, $$, $$$)
- vaddress, telephone, geo, image, aggregateRating
- vFAQPage: reservations, dietary, parking, dress code, private events
{
"@type": "Restaurant",
"name": "Oak & Table",
"servesCuisine": "Farm-to-Table American",
"hasMenu": "https://oakandtable.com/menu",
"acceptsReservations": "https://oakandtable.com/reservations",
"priceRange": "$$$",
"openingHoursSpecification": [
{
"@type": "OpeningHoursSpecification",
"dayOfWeek": ["Tuesday","Wednesday","Thursday"],
"opens": "17: 30", "closes": "22: 00"
},
{
"@type": "OpeningHoursSpecification",
"dayOfWeek": ["Friday","Saturday"],
"opens": "17: 00", "closes": "23: 00"
}
]
}Common restaurant audit findings.
No Restaurant schema
AI classifies the site as a generic webpage, not a restaurant.
hasMenu pointing to a PDF
AI crawlers cannot reliably parse PDF menus.
acceptsReservations missing
AI cannot tell diners how to book or whether reservations are required.
tracked U.S. queries with AI Overviews in BrightEdge's February 2026 dataset
BrightEdge, February 2026
of cited URLs also ranked in Google's organic top 10
BrightEdge, February 2026
of cited URLs came from outside Google's organic top 10
Derived from BrightEdge, February 2026
Common questions for restaurants.
OpenTable and Resy have their own AI optimisation and schema. VERIS optimises your own website schema - both should be in place. The acceptsReservations property links to your reservation platform URL.
The hasMenu property links to your menu page URL, which you update normally. The schema property itself does not need updating when the menu changes, only if the URL changes.
No. VERIS implements the schema that references your existing reservation platform. We do not manage or configure the platforms themselves.
Yes. Strong reviews help trust signals but do not substitute for schema. A restaurant with 4.9 stars and no Restaurant schema still cannot be classified correctly by AI systems.
Delivery apps are separate discovery channels. VERIS optimises direct web search visibility, the channel where customers find restaurants before deciding to visit or book.