Restaurants: get found on Google and in AI assistants (no al

By Équipe EtOH · Published on 13 Nov 2026

Customers now choose restaurants via Google/Maps and AI assistants. A practical plan: Google Business Profile, site pages, reviews system, schema, and tracking.

Restaurant discovery is no longer just a “Google + Maps” game. A growing share of decisions now happens via AI assistants (ChatGPT, Gemini, Perplexity) that summarize and recommend. The good news: most levers are the same — but you need to execute them cleanly, and track what changes.

Here’s a concrete plan so a restaurant gets found, clicked, and chosen.

1) Lock down the essentials: Google Business Profile (GBP) + NAP consistency

Your Google listing (Google Business Profile) is still the first point of contact. A “half-filled” listing costs reservations.

Top-priority actions (by impact):

1. Categories: 1 primary + 3 secondary.

- Example: Italian restaurant (primary), then Pizzeria, Family restaurant, Wine bar.

1. Hours: keep public holidays and service hours updated.

- Wrong hours = negative review + trust drop.

1. Menu: add a stable menu link (ideally a page on your site, not a PDF whose URL changes).
2. Attributes: patio, wheelchair access, vegetarian options, reservations, delivery… These are search filters.
3. Recent photos: 20–40 photos minimum, then 5 new ones/month.

- Reality check: a restaurant that adds 10 recent photos before peak season often sees an increase in clicks for directions.

NAP consistency (Name / Address / Phone):

• Your name, address, and phone number must be identical everywhere (website, social, directories, booking platforms).
• Avoid variants like “Bistrot Martin – Paris 11” on one site and “Bistrot Martin” elsewhere.

Realistic numbers example:

• If your GBP generates 4 000 views/month and 6% click “Directions” (240), improving photos + hours + menu can reasonably move that to 7–8% (280–320). At 20% conversion into actual visits, that’s +8 to +16 potential tables/month (depending on capacity and service).

2) Build “search intent” pages on your site (not just a menu)

AI assistants and Google prefer citing clear sources: structured pages, explicit answers, up-to-date information. Your website should become the official reference.

Pages to create (or improve) first:

• “Book” page: simple, fast, mobile-first, with a highly visible button.
• “Menu” page: HTML (not only PDF), clear sections (lunch/dinner), prices, allergens.
• “Private dining / groups” page: capacity, price range per person, terms.
• Local SEO pages (1 to 3 max): “Italian restaurant in Lyon 6”, “Pizzeria near Part-Dieu”, etc.

- Don’t publish 40 near-identical pages: fewer, genuinely useful pages work better.

A writing method that works (and AI likes):

• Put a 10–15 line direct answer at the top:

- “We serve homemade Italian cuisine, fresh pasta, a short menu, vegetarian options. Open Tuesday to Sunday. Reservations recommended on weekends.”

• Then details: menu, photos, access, FAQ.

Realistic numbers example:

• A well-built “private dining” page that goes from 30 to 120 visits/month, with 5% inquiries (6) and 30% conversion (2 private events), can generate 2 events/month. At 45€/person and 20 people: 1 800€ revenue per event → 3 600€/month.

3) Turn reviews into a system (quality, volume, replies) — not a lottery

Reviews are a major signal for Google, and a source of proof for AI assistants (which summarize “what people say”).

Realistic goal:

• Aim for a pace rather than a spike: for example 12 to 20 reviews/month depending on your traffic.

Simple 3-step process:

1. Ask at the right moment: at checkout or right after dessert.
2. Channel: QR code on the bill + short link.
3. Light reminder: if you have an email list (reservations), send a message the next day (without pestering).

Replying to reviews (operational model):

• Reply to 100% of negative reviews and 30–50% of positive ones.
• Timing: ideally within 72h.

Examples of helpful replies:

• Positive: thank them + mention a distinctive element (“fresh pasta of the day”) + invite them back.
• Negative: apologize, ask for context, propose a direct exchange (no public debate), and mention corrective action.

Realistic impact:

• Moving from 4.2 to 4.5 stars (at comparable volume) can change the game for queries like “best restaurant…”. Even without universal guarantees, you’ll often see more clicks and more “cool-headed” decisions.

4) Structured data & readability: help Google/AI understand your offering

AI tools “read” clear sites better. Google also relies on structured data.

Concrete actions:

• Schema.org Restaurant on your site: name, address, phone, hours, price range, cuisine, booking URL.
• Menu: if possible, menu markup (or at least a clean HTML page).
• FAQ: 6 to 10 common questions on one page (or at the bottom of key pages):

- “Do you accept groups?”, “Gluten-free options?”, “Can we come with a child?”, “Do you have a heated terrace?”

Realistic example:

• A clear FAQ reduces “basic” calls (hours, access, options) and increases booking conversions because uncertainty is removed before guests even reach out.

5) Getting recommended by AI assistants: signals to strengthen (beyond Google)

AI assistants pull from: your website, Google, review platforms, articles, reputable directories, and overall consistency.

Pragmatic levers:

• High-quality local citations: chamber of commerce, tourist office, local guides, neighborhood media.
• Key platforms (depending on your market): TheFork, TripAdvisor, Michelin, Petit Futé, etc. No need to be everywhere: target 3–5 well-maintained profiles.
• “Proof” content: a “Our cuisine” page with 6–10 signature dishes, your producers, your wine list (even short) — AI loves clear lists.
• News: events, seasonal menu, summer hours. A site that hasn’t moved in 18 months feels “out of date.”

Simple tip: publish 1 mini update/month (150–300 words):

• “Truffle menu (January)”, “Terrace open (May)”, “New express lunch (September)”.

Operational checklist (tick off in 60 minutes, then as a routine)

To do this week

• [ ] GBP: categories, hours (including holidays), attributes, menu link, booking link
• [ ] 20 photos minimum + 5 recent photos (dining room, dishes, storefront, team)
• [ ] “Menu” page in HTML (or at least one stable page)
• [ ] “Book” page with a clear CTA
• [ ] Google review QR code ready to print

To do this month

• [ ] 1 “Groups/private dining” page (capacity, pricing, terms)
• [ ] 1 useful local page (neighborhood/city) without artificial duplication
• [ ] FAQ (6–10 questions) + direct answers
• [ ] Schema.org Restaurant (or via your CMS/plug-in)
• [ ] Review routine: weekly target (e.g., 3–5 reviews/week)

To track each month

• [ ] GBP views, direction clicks, call clicks, website clicks
• [ ] Review volume + average rating + recurring themes
• [ ] 1 update published (seasonal, event, offer)

To quickly diagnose your presence (Google + signal consistency), you can use Visibility Check.

Conclusion

Getting found in 2026 means winning on two fronts at once: Google/Maps (immediate intent) and AI assistants (summarized recommendations). There’s nothing magical about the recipe: a flawless Google listing, a readable up-to-date website, a steady review system, and structured information.

If you only do 5 things: update your GBP, publish a clear HTML menu, build an effective booking page, put a review routine in place, and add an FAQ. This foundation is often enough to turn “passive” visibility into real reservations.