What Diners Do Before They Walk In
How restaurant discovery changed in 2026, which restaurants AI recommends, and what growing operators do differently
The industry is splitting
Technomic's count at the end of 2025 shows the shape of it. Independent restaurants fell to 412,498 locations, down 2.3% in a year, a net loss of about 9,500. Chain locations grew 1.4%. More than nine in ten operators told the National Restaurant Association they're feeling food, labor, insurance, and inflation costs. Scale absorbs that pressure; smaller operators are the ones closing.
Costs aren't something a small restaurant group can fix alone. How customers find it is.
How diners find restaurants now
Toast asked 1,466 U.S. diners how they actually found their most recent restaurant. Among Gen Z, TikTok alone accounted for 38% of discovery, the largest single source for that age group. Diners under 35 lean on social platforms to decide where to eat at meaningfully higher rates than diners over 35.
The newer channel is AI. DoorDash's 2026 industry survey of 3,001 consumers found that 22% of U.S. diners had used a tool like ChatGPT or Gemini to choose a restaurant. Among diners aged 25 to 34, 61% had. A separate survey by Reputation and Nielsen put AI use for researching restaurants and bars at one in five consumers, and at nearly half of people in their late thirties and early forties.
A diner who asks an assistant for "a good patio for a group of eight near downtown" isn't typing keywords. They're asking a question, and the assistant answers with three to five names. Most of those questions never mention a brand: Malou's 2026 analysis found 79% of restaurant searches are non-branded, which means most groups can't rely on their name to bring diners in.
The restaurants AI recommends, and the 83% it doesn't
Local Falcon's analysis this spring found that 83% of restaurant locations don't appear in AI-generated recommendations at all, and that the restaurants which do appear dominate their category repeatedly.
What gets a restaurant onto the list isn't mysterious. The assistants read public sources: Google, Yelp, Foursquare (widely reported to power most of ChatGPT's local answers), delivery apps, and social. Analyses of what they cite find complete listings, ratings in the mid-fours, recent reviews with replies, and specific detail about which dish, which occasion, and which neighborhood. Bloom Intelligence's summary of the restaurant-specific evidence notes an apparent floor around 4.3 stars for ChatGPT's recommendations, and that most AI restaurant prompts are research-style questions of the kind keyword search was never built to answer.
What growing operators do that shrinking ones don't
This September, SOCi published a comparison of eight restaurant chains that were expanding with eight that were closing locations, scored across search, reviews, social, and AI. The differences were consistent on every channel:
- Reputation. The growing chains averaged a 4.39 Google rating against 3.82. They answered 72% of their Google reviews against 44%, two to three times faster.
- Social. The growing chains had roughly 26 times the local social engagement rate of the shrinking ones. SOCi's finding was that local content drove the gap, not volume; pushing the same corporate post to every location page did nothing.
- AI. The growing chains were recommended by ChatGPT in about 20% of test queries against roughly 3%.
SOCi is explicit that this is a pattern, not proof of cause, and the study is both sponsored and drawn from national chains. But every lever in it, response rate, response speed, local content, listing accuracy, is available to a small group this week.
What the Google profile is worth to a restaurant
For a restaurant, the Google Business Profile is where intent turns into a visit. WebFX's 2026 benchmarks put the share of restaurant profile viewers who request directions at 7% to 10%, roughly double the rate across other industries, and found that a spot in the top three map results draws more than twice the traffic of positions four through ten.
That profile responds to attention. Malou's 2025 study of more than 300 restaurant locations found that restaurants actively maintaining their Google profile earned 2.3 times as many reviews and at least 15% more interactions, meaning calls, directions, and website clicks, within six months. Its larger study of 2,000-plus locations found that groups which centralized their listings, reviews, and social management grew organic traffic by 74% in three months.
Operators know, and can't get to it
Restaurant operators are not unaware of any of this. In Popmenu's 2025 survey of 300 U.S. operators, 87% said they planned to focus harder on their online reputation in 2026. Only a third said they currently work on it daily. The gap between intention and practice is the subject of the rest of this paper.
Conclusion
Technomic's unit counts, three consumer surveys from Toast, DoorDash, and Reputation, AI visibility measurements from Local Falcon and SOCi, profile benchmarks from WebFX and Malou, and Popmenu's operator survey come from organizations with no reason to agree with each other. They describe one picture: diners are finding restaurants through social and AI, the AI shortlist is short, the restaurants that get on it keep an accurate, answered, active public record, and most operators know it matters but aren't getting to it. That convergence is why we hold this conclusion with confidence, and it should give a reader the same.
For a smaller restaurant operator, the gap is a set of jobs nobody currently holds. Nobody has asked ChatGPT, Gemini, Perplexity, and Claude what they say about each restaurant, or checked whether the answer is the old name. Reviews arrive faster than managers reply, and the reply rate that separates growing chains from shrinking ones isn't tracked by anyone. The dish that shows up in five hundred reviews has never been posted. Listings drift with every rebrand and holiday schedule. And nothing measures whether any of it moved covers. The costs of running a restaurant in 2026 are largely outside an operator's control. The public record that decides whether a diner walks in is not.
By the numbers
| Figure | What it says | Source |
|---|---|---|
| −2.3% / +1.4% | Independent restaurant locations / chain locations, 2025 | Technomic |
| 38% | Share of Gen Z restaurant discovery from TikTok | Toast, n=1,466 |
| 22% / 61% | U.S. diners / diners 25–34 who have used AI to choose a restaurant | DoorDash with Dynata, n=3,001 |
| 20% / 45% | Consumers / consumers 35–44 using AI to research restaurants and bars | Reputation / Nielsen, Nov 2025 |
| 79% | Restaurant searches that are non-branded | Malou, 2026 |
| 83% | Restaurant locations absent from AI recommendations | Local Falcon, Mar 2026 |
| ~4.3 | Apparent rating floor for ChatGPT restaurant recommendations | Bloom Intelligence synthesis |
| 4.39 vs 3.82 | Google rating, expanding vs contracting chains | SOCi, Sept 2026 |
| 72% vs 44% | Review response rate, expanding vs contracting chains | SOCi |
| ~26x | Local social engagement rate, expanding vs contracting | SOCi |
| ~20% vs ~3% | ChatGPT recommendation rate, expanding vs contracting | SOCi |
| 7–10% | Restaurant profile viewers who request directions | WebFX, 2026 |
| 2.3x / +15% | Reviews / interactions within 6 months, restaurants actively maintaining their Google profile | Malou, 300+ locations |
| +74% | Organic traffic in 3 months, groups that centralized listings, reviews, and social | Malou, 2,000+ locations |
| 87% / 33% | Operators planning to focus more on reputation in 2026 / working on it daily | Popmenu, n=300 |
Sources
- Technomic, unit counts for 2025, via Nation's Restaurant News (Feb 24, 2026) and Restaurant Business (Apr 29, 2026); National Restaurant Association, State of the Industry.
- Toast, 2026 Guest Discovery Report (n=1,466, Oct 2025).
- DoorDash, 2026 Restaurant Industry Trends Report (with Dynata, n=3,001).
- Reputation / Nielsen consumer survey (Nov 2025), as cited by Joli (2026).
- Malou, Local SEO for Restaurants: 2025 Study and 2025 Digital Benchmark for Multi-Location Restaurants. malou.io
- Local Falcon, The AI Visibility Crisis: Why 83% of Restaurants Don't Exist in ChatGPT (Mar 2026).
- Bloom Intelligence, AI Restaurant Discovery 2026: The Data (July 2026).
- SOCi, Why growing restaurant chains win at local search and AI visibility, MarTech, Sept 8, 2026.
- WebFX, restaurant local search benchmarks 2026, via Restaurant Velocity (July 2026).
- Popmenu, survey of 300 U.S. restaurant operators (2025), via PR Newswire (Apr 30, 2026).