We rely on Google's AI Overviews more and more to decide where to eat, asking the assistant for a recommendation instead of scanning a map ourselves. The real question is whether the restaurants it surfaces are the same ones a Google Maps search would show. They aren't, and for most restaurants a strong Maps presence no longer guarantees being found.
Local Falcon ran the same 10,000 US restaurants through both Google Maps and Google AI Overviews in all 50 states and Washington, D.C., then compared who appeared where. The gap between the two is wide, consistent across the country, and largely indifferent to the review counts and star ratings that restaurants have spent years building. Each restaurant was searched from nine spots around its own location: it counts as found if the AI named it at even one, and "invisible" if it was never named at any.
Key Findings
- Nearly 3 in 4 restaurants (74.9%) are invisible in Google's AI recommendations, never surfacing at a single nearby search when a diner asks the AI where to eat.
- A restaurant is nearly four times more likely to be invisible in AI (74.9%) than on Google Maps (19.6%).
- A restaurant with more than 1,000 Google reviews was still left out of AI recommendations 70.9% of the time.
- Among the restaurants the AI did recommend, 5.4% are rated below 3.5 stars, even though every search explicitly asked for highly rated places.
- Yelp is the outside website Google's AI cites most often when recommending restaurants, ahead of TripAdvisor, Facebook, and Reddit.
Most Restaurants Visible on Google Maps Never Appear in AI
The starkest number in the study is the overlap, or the lack of it.
Of the 8,059 restaurants that appeared somewhere on their local Google Maps grid, 5,670 never appeared once in the AI Overview for the identical search.
The gap is clearest in how often a restaurant is invisible, appearing nowhere in the local search.
| Platform | Invisible |
|---|---|
| Google Maps | 19.6% |
| Google AI Overviews | 74.9% |
The deeper story is not just that AI shows fewer restaurants, but that it funnels almost all of its attention onto the same small group.
The top 10% of restaurants capture 74.5% of all AI visibility, compared with 54% on Google Maps. In practice, that means a handful of names surface again and again across a city while everyone else is shut out. Google Maps spreads its visibility across a wide field of restaurants, giving most of them at least some presence somewhere on the grid. Google's AI does the opposite: it crowns a small winners' circle and treats the rest as if they do not exist.
For a restaurant, the result is binary. It is either one of the few names the AI keeps repeating, or it is invisible, with very little ground in between.
The Cuisines AI Hides Most
The disappearance rate is not uniform. It rises sharply as the search gets more generic, because a broad query pushes the AI toward a tiny set of famous names while a specific cuisine forces it to spread wider.
| Cuisine | Invisible on Maps | Invisible in AI |
|---|---|---|
| Restaurant (generic) | 32.0% | 87.5% |
| Bar | 17.8% | 85.1% |
| Sandwich shop | 15.7% | 85.0% |
| Pizza | 17.4% | 84.9% |
| Chicken | 17.8% | 79.9% |
| Coffee shop | 10.6% | 77.9% |
| Breakfast | 17.6% | 77.8% |
| Cafe | 22.8% | 77.2% |
| Fast food | 8.5% | 77.1% |
| Hamburger | 19.3% | 76.5% |
| Mexican | 25.4% | 76.2% |
| Bakery | 16.7% | 74.0% |
| Chinese | 12.4% | 73.9% |
| American | 20.7% | 72.0% |
| Italian | 21.7% | 68.5% |
| Seafood | 19.5% | 67.7% |
| Steak house | 14.4% | 63.1% |
| Barbecue | 19.6% | 62.4% |
| Thai | 14.3% | 54.8% |
| Chicken wings | 12.3% | 54.2% |
| Indian | 10.6% | 50.0% |
| Brewery | 14.1% | 45.3% |
Someone searching the broad phrase "restaurant near me" is the most exposed: nearly 9 in 10 local restaurants never surface in the AI answer. Niche and destination categories fare better, with breweries, Indian restaurants, and wing shops appearing roughly half the time. No category, however, comes close to its Google Maps coverage.
Great Reviews No Longer Save You
The signals that restaurants are told to chase, a high star rating and plenty of reviews, do little to secure a place in the AI answer.
Review volume barely moves the result. A restaurant with more than 1,000 reviews is invisible in AI 70.9% of the time, almost the same rate as a restaurant with a few hundred.
| Google reviews | Invisible in AI |
|---|---|
| Under 50 | 86.8% |
| 50 to 199 | 76.3% |
| 200 to 499 | 70.9% |
| 500 to 999 | 70.1% |
| 1,000 or more | 70.9% |
Star rating helps up to a point and then stops. Restaurants in the 4.5 to 4.7 band are the most visible, while a near-perfect 4.8 or above is actually hidden more often, in part because the highest-rated places tend to be newer spots with thinner review histories.
| Star rating | Invisible in AI |
|---|---|
| Under 3.5 | 90.7% |
| 3.5 to 3.9 | 86.7% |
| 4.0 to 4.4 | 73.6% |
| 4.5 to 4.7 | 60.0% |
| 4.8 or above | 69.0% |
A popular assumption is that AI must be falling back on national chains at the expense of local independents. The data says the opposite. Independent restaurants appear in AI results more than twice as often as chains.
| Type | Appears on Maps | Appears in AI |
|---|---|---|
| Independent | 79.0% | 29.1% |
| Chain (five or more locations) | 84.3% | 13.9% |
Even the Biggest Chains Disappear in AI
Brand recognition offers no protection. The six chains with more than 100 locations in the study are absent from Google's AI recommendations as often as ordinary restaurants, or more.
McDonald's is the clearest example, and the most-tracked brand in the study: 277 of its locations were included, more than any other chain in the sample. Google Maps surfaced them in 96% of the searches around their own neighborhoods. Google's AI recommended a McDonald's in 0% of them: not one of the 277 appeared at a single grid point.
The others are only marginally better. Starbucks surfaces in under 5% of its searches, Subway in 10%, and Dunkin' in 14%, even though all three appear on Google Maps almost everywhere they operate.
| Brand | Locations tracked | Invisible in AI | Invisible on Maps |
|---|---|---|---|
| McDonald's | 277 | 100% | 4.3% |
| Starbucks | 170 | 95.3% | 11.2% |
| Subway | 190 | 90.0% | 14.7% |
| Dunkin' | 132 | 86.4% | 6.1% |
| Burger King | 130 | 83.1% | 46.2% |
| Taco Bell | 139 | 71.2% | 12.2% |
Five of the six biggest chains are invisible in AI more often than the average restaurant (74.9%), and McDonald's never appears at all. The pattern matches the rest of the study: the AI keeps a small winners' circle and leaves out almost everyone else, including the most established brands in the country.
AI Is Recommending Poorly Rated Restaurants
Every query in this study asked the same thing: recommend a highly rated restaurant with great reviews. The AI didn't always oblige. Among the 2,498 restaurants it recommended, only half clear the 4.5-star bar that most people would call highly rated, and roughly one in seven fall below 4.0 stars.
| Star rating of AI's pick | Share of AI's recommendations |
|---|---|
| 4.5 stars and above | 49.8% |
| 4.0 to 4.4 stars | 36.3% |
| 3.5 to 3.9 stars | 8.4% |
| Below 3.5 stars | 5.4% |
The Local Picture: State by State, City by City
The gap is not a coastal or big-city phenomenon. It holds in every state measured, ranging from about 54% invisibility in New Hampshire to nearly 88% in Hawaii.
| State | Invisible on Maps | Invisible in AI |
|---|---|---|
| Hawaii | 25.0% | 87.8% |
| Pennsylvania | 18.7% | 85.1% |
| Nebraska | 23.4% | 84.8% |
| New York | 18.7% | 83.7% |
| Nevada | 24.7% | 83.4% |
| Maryland | 25.4% | 82.5% |
| Illinois | 26.1% | 82.3% |
| Washington, D.C. | 27.9% | 82.2% |
| Georgia | 22.9% | 81.2% |
| Massachusetts | 26.2% | 80.2% |
| Oregon | 21.3% | 79.9% |
| Colorado | 24.5% | 79.9% |
| Oklahoma | 13.9% | 79.4% |
| Florida | 29.3% | 78.9% |
| Indiana | 15.8% | 78.8% |
| Tennessee | 15.0% | 78.6% |
| Kentucky | 22.9% | 78.4% |
| California | 20.4% | 78.2% |
| Virginia | 24.9% | 78.0% |
| South Carolina | 18.7% | 77.1% |
| Texas | 22.1% | 76.6% |
| North Carolina | 23.2% | 76.0% |
| Washington | 19.4% | 75.6% |
| Arizona | 23.1% | 75.0% |
| Missouri | 22.8% | 74.9% |
| New Mexico | 10.3% | 74.2% |
| Idaho | 15.6% | 73.8% |
| Arkansas | 14.1% | 73.3% |
| Utah | 16.7% | 72.0% |
| Ohio | 25.1% | 72.0% |
| Kansas | 12.4% | 71.6% |
| Michigan | 15.2% | 71.5% |
| Connecticut | 20.7% | 70.9% |
| Rhode Island | 14.0% | 70.7% |
| Alaska | 10.1% | 70.6% |
| Delaware | 15.6% | 70.0% |
| Louisiana | 16.1% | 69.8% |
| Alabama | 12.7% | 69.4% |
| Wisconsin | 12.9% | 69.3% |
| Maine | 18.5% | 69.1% |
| North Dakota | 18.6% | 69.0% |
| Mississippi | 9.6% | 68.9% |
| New Jersey | 19.9% | 67.5% |
| Iowa | 15.5% | 66.5% |
| West Virginia | 13.6% | 65.8% |
| Wyoming | 15.0% | 62.9% |
| Montana | 17.2% | 61.0% |
| South Dakota | 18.3% | 60.6% |
| Minnesota | 15.6% | 59.9% |
| Vermont | 12.5% | 59.5% |
| New Hampshire | 12.4% | 53.9% |
The same divide appears at the city level. Among the 30 cities with the largest samples, New York hides 9 in 10 of its restaurants from AI, and even the most AI-friendly major city, Austin, hides more than two-thirds.
| City | Restaurants | Invisible in AI |
|---|---|---|
| New York, NY | 115 | 90.4% |
| Omaha, NE | 126 | 87.3% |
| Honolulu, HI | 147 | 86.4% |
| Philadelphia, PA | 111 | 85.7% |
| Chicago, IL | 128 | 84.5% |
| Las Vegas, NV | 146 | 84.4% |
| Los Angeles, CA | 111 | 83.8% |
| Orlando, FL | 107 | 83.2% |
| Washington, D.C. | 179 | 82.2% |
| Nashville, TN | 104 | 81.7% |
| Houston, TX | 114 | 81.6% |
| San Antonio, TX | 108 | 81.5% |
| Richmond, VA | 104 | 80.8% |
| St. Louis, MO | 104 | 80.0% |
| Portland, OR | 109 | 79.8% |
| Louisville, KY | 104 | 79.0% |
| Indianapolis, IN | 109 | 78.9% |
| Tampa, FL | 109 | 78.2% |
| San Francisco, CA | 108 | 77.8% |
| Miami, FL | 107 | 77.6% |
| Dallas, TX | 108 | 76.9% |
| Atlanta, GA | 106 | 76.4% |
| Brooklyn, NY | 114 | 76.3% |
| Phoenix, AZ | 104 | 76.0% |
| Boston, MA | 108 | 75.9% |
| Jacksonville, FL | 104 | 75.0% |
| San Diego, CA | 111 | 74.8% |
| Albuquerque, NM | 109 | 73.4% |
| Cincinnati, OH | 106 | 69.2% |
| Austin, TX | 107 | 68.2% |
The Sources Google's AI Uses to Pick Restaurants
When Google's AI recommends a place to eat, it cites the sources behind its answer. Across more than 657,000 individual citations, those sources concentrate around a short list of platforms. Google's own listings account for almost half of every citation, meaning the AI largely recycles the same Maps and Business Profile data that powers the map it then diverges from.
Among outside websites, one name dominates: Yelp is cited more often than TripAdvisor, Facebook, Reddit, and Instagram combined. Community and social platforms now feed the AI heavily, and Reddit is cited more than any individual food publication, while professional restaurant criticism barely registers against user reviews and forum posts.
| Source | Share of AI's citations |
|---|---|
| Google's own listings | 48.6% |
| Yelp | 14.1% |
| TripAdvisor | 4.4% |
| 2.6% | |
| 2.3% | |
| 1.9% | |
| OpenTable | 1.6% |
| YouTube | 0.7% |
| Eater | 0.5% |
| Wanderlog | 0.4% |
| TikTok | 0.4% |
| The Infatuation | 0.2% |
What Helps a Restaurant Show Up in AI
A handful of factors track with whether a restaurant gets recommended by AI. The clearest is web presence. Restaurants with no website at all are surfaced only 16.7% of the time, roughly half the rate of those with one, and a Facebook or Instagram page outperforms a conventional website, the same social platforms the AI leans on for its sources. Claiming a Google Business Profile shows a similar lift, with 26.1% of claimed listings visible against 17.9% of unclaimed ones.
| Web presence | Share visible in AI |
|---|---|
| Facebook or Instagram page only | 39.4% |
| Own website | 26.2% |
| Website builder or landing page | 25.0% |
| No website at all | 16.7% |
Methodology
Local Falcon analyzed 10,000 US restaurants across all 50 states and Washington, D.C. in May 2026, running each through both Google Maps and Google AI Overviews, all run midday when restaurants are open. Each restaurant was checked from nine points on a grid around its own location, using its primary cuisine category. Maps queries were typed the way a customer types them ("pizza near me"); AI queries were phrased the way a customer asks a chatbot ("can you recommend some highly-rated pizza near me with great reviews").
A restaurant counts as visible if it appears at even one of its nine grid points, and invisible if it appears at none. Visibility is the share of those nine points where it shows up, and the invisibility rate is the share of restaurants that never appeared at a single point. The same yes-or-no test was applied to both platforms (did this specific restaurant appear at this point or not), and position in the list was not scored, so a restaurant counts as found whether it ranked first or twentieth. Citation sources were taken from the published references that accompany each AI Overview response.
