AI assistants recommend restaurants they can match to a diner's exact question: cuisine, area, occasion, price, dietary needs and opening hours. Put that information on your own website as text, not PDFs or photos; add Restaurant structured data; keep hours identical everywhere; write short answers for the occasions you want; and get onto the local food guides the assistants cite.
The questions diners ask
Diners ask assistants far more specific questions than they type into a search box: "halal steakhouse open late in Downtown", "quiet place for a business lunch near DIFC", "vegan-friendly brunch with outdoor seating", "Sunday roast with a garden for dogs". The assistant searches, reads a few pages and names two or three places, usually with a reason for each.
Notice what those questions contain: cuisine, area, occasion, dietary need, practical detail. A venue gets named when the pages the assistant reads state those things clearly.
Your menu has to be readable
The single most common problem is a menu that exists only as a PDF or an image. Some assistants can read some PDFs some of the time, but a menu in plain HTML text is far more reliable, faster on a phone and better for search.
- Publish the menu as a normal web page, with sections, dishes and prices.
- Label dietary options in words: vegan, vegetarian, gluten-free, halal, contains nuts.
- Keep it current. A changing menu is fine; an out-of-date one is not.
- Keep a PDF as an extra for people who want to print it, not as the only version.
Restaurant structured data
Restaurant is a type of FoodEstablishment in schema.org (others include BarOrPub, CafeOrCoffeeShop and Bakery). The properties that matter most are cuisine, price range, reservations, menu link, address, map location and opening hours. Here's a complete example for a fictional restaurant:
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Restaurant",
"@id": "https://www.biancatrattoria.example/#restaurant",
"name": "Bianca Trattoria",
"url": "https://www.biancatrattoria.example/",
"servesCuisine": ["Italian", "Pizza"],
"priceRange": "$$",
"acceptsReservations": true,
"hasMenu": "https://www.biancatrattoria.example/menu",
"telephone": "+971 4 000 0001",
"address": {
"@type": "PostalAddress",
"streetAddress": "Cluster D, Jumeirah Lake Towers",
"addressLocality": "Dubai",
"addressCountry": "AE"
},
"geo": { "@type": "GeoCoordinates", "latitude": 25.0700, "longitude": 55.1450 },
"openingHoursSpecification": [
{ "@type": "OpeningHoursSpecification",
"dayOfWeek": ["Monday", "Tuesday", "Wednesday", "Thursday", "Sunday"],
"opens": "12:00", "closes": "23:00" },
{ "@type": "OpeningHoursSpecification",
"dayOfWeek": ["Friday", "Saturday"],
"opens": "12:00", "closes": "01:00" }
]
}
</script>If you want to go further, schema.org's Menu, MenuSection and MenuItem types can describe the menu itself, including suitableForDiet for dietary labels. Only mark up what's visible on the page, and validate it with the Schema Markup Validator.
Hours and details that agree everywhere
Restaurants appear in more places than almost any other business: your site, Google Business Profile, review platforms, delivery apps, booking platforms and food guides. When opening hours or phone numbers differ between them, every system becomes less confident.
Pick one source of truth, update it first, then update every listing on the same day, especially for holidays and seasonal hours.
Write for the occasions you want
Decide which occasions you want to win: birthdays, groups, families, business lunches, date nights, late-night dining. Then give each one a short, specific answer on your site: maximum group size, set menus, private room, high chairs, quiet tables, last orders. A few sentences per occasion is enough. These are exactly the details an assistant quotes when it explains a recommendation.
Local food guides and reviews
For restaurant questions, assistants lean heavily on a small number of sources: local food guides and city magazines, award and "best of" lists, review platforms and your Google Business Profile. Find out which ones are cited for your area by asking the assistants your target questions and noting the sources.
Then work on being included the honest way: invite the right writers, enter the right awards, and make it easy for happy guests to leave reviews. Never buy reviews.
Checklist
- Menu as an HTML page, with dietary labels, kept current
- Restaurant (or BarOrPub, Cafe) structured data with cuisine, price range, hours, menu link and location
- Neighbourhood and nearby landmarks stated in plain words
- Identical hours, phone and address on every listing
- Short answers for the occasions you want to win
- A fast mobile site with a clear booking button
- Presence on the food guides cited for your area