Something has quietly shifted in how people decide where to eat.
A few years ago, a hungry diner opened Google. Now, plenty of them open ChatGPT, Gemini or Perplexity and simply ask: “Where’s a good Italian place near me with outdoor seating?” or “Best brunch in the area for a group of six?”
The AI answers with a short list. Three, maybe four restaurants. Not ten blue links a handful of confident recommendations.
Here’s the question that matters for your business: when someone asks an AI where to eat, does your restaurant come up?
For most restaurants, the honest answer right now is no. And the reasons are fixable.
Spark Digital UK works with restaurants on Google visibility, reviews and website SEO, so this is a shift we’re watching closely. This guide explains how AI recommendations actually work, why most restaurants are invisible to them, and what you can do about it.
Most Restaurants Are Already Invisible to AI
Let’s start with the uncomfortable number.
A May 2026 report from Uberall found that 83% of restaurant locations never appear in AI-generated recommendations even though 86% of them are present on Google.
Read that again. Being on Google is no longer enough. Most restaurants have done the basic work to exist on Google Maps, and are still completely absent when a diner asks an AI assistant for a suggestion.
That gap is the opportunity. Almost nobody at a local level is working on this yet. The restaurants that sort it out now will be the ones AI keeps recommending long after everyone else catches on.
How AI Assistants Actually Choose a Restaurant
This is the part worth understanding properly, because it changes what you do.
A traditional search engine sorts pages by ranking position. AI assistants don’t work that way. They don’t pick the “best” restaurant. They pick the restaurant whose information is the most reliable, consistent and easy to state without getting it wrong.
Think about it from the AI’s point of view. It’s about to make a recommendation to a real person. It wants to avoid being wrong sending someone to a place that’s closed, or misdescribing the cuisine. So it favours restaurants where the cuisine, the neighbourhood, the price range, the hours and the reputation are all clear, consistent and confirmed across multiple trusted sources.
If your details are patchy, contradictory or hard to find, the AI hedges or just recommends someone else it feels more sure about.
That’s the core principle. AI recommends what it can confidently verify.
Where Each AI Gets Its Restaurant Information
Here’s something most guides get wrong: the major assistants don’t all draw from the same place.
Only Google owns a built-in database of local businesses Google Maps and Google Business Profile and Gemini is grounded directly in it. The other assistants don’t own anything comparable, so they rely on the open web and on licensed data sources.
What that means in practice:
- Gemini and Google AI Overviews lean heavily on your Google Business Profile, reviews and website. This is Google’s home turf.
- ChatGPT, Perplexity and Claude draw on what they can read across the web pages that already rank on Google, fast websites their crawlers can actually read, structured information and reviews — plus licensed data. It’s been reported that ChatGPT’s local answers lean on licensed directory data, including Yelp.
There’s an important UK caveat here. A lot of the advice written about this is US-centric, and Yelp is far more dominant in the US than in the UK, where Google and TripAdvisor tend to carry more weight. The specific sources also shift as licensing deals change.
So don’t chase one directory. The durable strategy is the same regardless of which source is in favour this month: be accurate, consistent and well-reviewed across all the major trusted sources that matter in your market. For a Harrow restaurant, that means Google first, then the platforms your customers actually use — TripAdvisor, your booking platform, and your own website.
The Signals That Get Your Restaurant Recommended
Across every assistant, the same handful of signals keep coming up. Strengthen these and you improve your odds everywhere at once.
1. Reviews and ratings the entry ticket
This is the big one. AI assistants strongly favour well-reviewed restaurants, and there appears to be a floor.
Reporting from within the industry suggests observed rating thresholds: ChatGPT tends to recommend restaurants averaging around 4.3 stars or higher, with Perplexity and Gemini a little lower. These aren’t official figures published by the AI companies they’re patterns observers have noticed but the direction is clear and consistent. Below a certain rating, you rarely get recommended.
Volume matters as much as the score. A restaurant with a handful of reviews gets less confidence than one with a broad, recent review base, even at the same star rating. A wide sample signals a reputation that’s been validated by lots of people, which is exactly the reassurance an AI is looking for.
Recency counts too. Steady, recent reviews say “this place is still good,” where reviews that stop two years ago raise a quiet question mark.
The takeaway: consistent review generation isn’t just a Google tactic anymore. It’s now the single biggest lever on whether AI recommends you.
2. Consistent, accurate business information
Your name, address, phone, hours and cuisine need to match everywhere they appear your website, Google Business Profile, TripAdvisor, booking platforms, social pages.
Every inconsistency is a reason for the AI to lose confidence. If one source says you close at 10pm and another says 11pm, the AI can’t state your hours reliably, so it’s less likely to put you forward.
3. A website an AI can actually read
This is a genuinely different problem from ranking on Google.
Some assistants use crawlers to read the open web. If your site is built mostly from images, loads slowly, or hides your menu inside a PDF, there’s very little the crawler can actually read and a page it can’t read is a page it can’t use to recommend you.
The fix overlaps with good on-page SEO: real text rather than images, a fast site, your menu published as readable text on the page, and clear information about cuisine, location and price.
4. Details stated clearly and specifically
AI recommends what it can state without risk. So the more specific and verifiable your information, the better.
“Great food in a nice setting” tells an AI nothing it can use. “Wood-fired Neapolitan pizza in Wembley, with outdoor seating and vegan options, open until 11pm” gives it exactly the kind of concrete, checkable detail it needs to match you to a specific request like “vegan-friendly pizza near Wembley with a terrace.”
5. Being mentioned in places the AI trusts
AI cross-references. When your restaurant is mentioned consistently across reviews, local press, food blogs and directories, that consistent picture builds the confidence an AI needs to recommend you. A single strong website matters less than a consistent story told across many sources.
Find Out What AI Says About You Right Now (10 Minutes)
Before you change anything, see where you stand. This is the most useful ten minutes you’ll spend on this.
Open ChatGPT, Gemini and Perplexity, and ask each one the questions your customers would actually ask:
- “Good [your cuisine] restaurant in [your area]”
- “Where should I eat near [a nearby landmark or station]?”
- “Best [your cuisine] in [your area] for [a group / a date / families]”
- “[Your cuisine] with outdoor seating in [your area]”
Then read the answers carefully:
- Does your restaurant get mentioned at all?
- If so, is what the AI says about you accurate cuisine, location, hours, specialities?
- Which competitors come up instead of, or above, you?
- Is the AI describing anyone with wrong or outdated information?
Whatever comes back is your baseline. If you’re not mentioned, you know you’ve got work to do. If you’re mentioned with wrong details, you’ve found exactly which information needs cleaning up. If competitors dominate, you can see who’s currently winning this channel in your area.
Do this again every couple of months. It’s the simplest way to track whether your work is landing.
What to Actually Do
The good news runs through everything above: getting found by AI isn’t a separate project with its own strange rules. It’s mostly the same work that strengthens your whole Google presence — done consistently.
A practical order of priority:
- Get your Google Business Profile complete and accurate. It grounds Gemini and Google AI Overviews directly, and feeds the consistent picture every other assistant cross-references.
- Build reviews steadily and keep your rating healthy. This is the biggest lever. Aim for a consistent flow of recent reviews and reply to them. Given the rating floors, this is where most of the payoff sits.
- Make your information consistent everywhere. Same name, address, phone, hours and cuisine across every platform. Fix contradictions.
- Make your website readable. Real text, fast loading, menu as text not an image, clear details about what you serve and where.
- Be specific about what makes you findable. Cuisine, neighbourhood, price range, dietary options, standout dishes, hours stated plainly so an AI can match you to specific requests.
- Keep your presence strong on the platforms that matter in your market for the UK, Google and TripAdvisor ahead of US-centric directories.
None of this is exotic. It’s disciplined local marketing, pointed at a new channel.
The Honest Limits
A few things worth being straight about, because plenty of content on this topic isn’t.
You can’t force an AI to recommend you. There’s no submission form, no paid placement, no switch to flip. You improve your odds by strengthening the signals AI trusts. Anyone selling guaranteed AI recommendations is selling something that doesn’t exist.
This is moving fast. Which sources each assistant uses, how AI Overviews behave, the rating patterns all of it is shifting month to month. What’s true today will need revisiting. Treat this as an area to monitor, not a box to tick once.
The fundamentals are the safe bet. Because the specifics keep changing, the smart move is to invest in what stays true across all of them: accurate information, strong reviews, a readable website, consistent presence. Those help you in AI search, in Google, and with actual human customers — whatever the algorithms do next.
Frequently Asked Questions
Can I pay to have ChatGPT recommend my restaurant?
Is my Google Business Profile enough to get found by AI?
Why does my restaurant show up on Google but not in ChatGPT?
Do reviews really affect AI recommendations?
Does this matter in the UK, or is it mostly a US thing?
How do I know if any of this is working?
Want to Know What AI Says About Your Restaurant?
If you’d like to see where your restaurant currently stands in AI search and get your Google presence, reviews and website into the shape that gets you recommended Spark Digital UK can review your setup and show you where the gaps are. Get a free Google & AI visibility audit.
