Your Next Customer Is Asking ChatGPT

A year ago, if you had asked a hundred American consumers whether they'd used ChatGPT or a similar tool to find a restaurant, a dentist, or a plumber, six would have said yes. Ask the same question today and forty-five say yes. That comparison comes from BrightLocal's annual survey of about a thousand U.S. adults, and it represents the fastest change in how people find local businesses since Google Maps arrived on phones.

The number deserves a minute of an owner's attention for a reason that isn't obvious at first. An AI assistant doesn't hand people a list of twenty options to scroll through. It names a few and stops. Whether your business is one of them depends on a public record that, at most smaller businesses, nobody is looking after.

Try it before you read on

Open ChatGPT, Gemini, or Perplexity on your phone. Type the question your customers ask: "best [what you do] near [your neighborhood]." Then ask it directly: "What do people say about [your business]?" Most owners doing this for the first time find they're missing from at least one assistant, or that it's describing an old name, a location they closed, or a two-star review from last spring as though it were the whole story. Whatever you found is what this paper is about.

What changed, and how fast

The finding doesn't rest on one survey. Yext, using different wording and a different sample, found that roughly 47% of U.S. adults had recently used AI to find a local business. Adobe's global survey of 4,000 consumers found that about a quarter now name AI platforms like ChatGPT as their top research tool, ahead of brand websites and online reviews. McKinsey puts the share of consumers in the biggest spending categories who already use AI search to make purchase decisions at between 40% and 55%.

Yext's survey also turned up something that should shape how you think about this: more than nine in ten people who used AI to find a business checked another source, usually Google or a review site, before acting on it. The assistant makes the shortlist, and your Google profile and reviews close the deal. Both have to be right, and both draw on the same public information.

Google itself has gone the same way. Its AI Overviews, the written summary that now sits at the top of many searches, appear on 68% of local business searches, according to Whitespark's 2026 survey of local search practitioners. That is more often than the familiar three-listing map appears.

The shortlist problem

On a search results page, a business ranked eighth still exists. A customer can scroll to it. An AI answer has no eighth position.

SOCi measured this across more than 350,000 business locations. ChatGPT recommended about 1.2% of them. Put another way, for every hundred businesses that could have been named for a given question, roughly one was. Uberall's testing found the same thing from the customer's side: ask an assistant for a restaurant and you get three to five names, never a longer list.

Two things follow. The businesses that get named tend to keep getting named, because the same sources keep confirming them. And there's no way to buy your way onto the list. When OpenAI started selling ads inside ChatGPT in February 2026, the reported minimum was $200,000, which rules out smaller operators entirely. The only route in is to be the best-documented answer to the question.

The customers who arrive this way are worth having. Semrush's analysis of referral traffic found a visitor who comes from an AI answer is worth about four times as much as a typical search visitor, because they've already done their research and arrive ready to decide.

What the machine is actually reading

Owners tend to assume the fix is their website. The evidence points elsewhere. Omniscient Digital analyzed more than 23,000 citations in AI-generated answers and found that about three quarters of what the assistants relied on came from somewhere other than the business's own site: directories, review platforms, forums, video. Foursquare is widely reported to supply most of the underlying data for ChatGPT's local answers, and Yelp has licensed its reviews to OpenAI.

So "digital footprint," a phrase that sounds like it means a website, actually means the sum of what is published about you across:

• your Google Business Profile: name, category, hours, photos, posts, and every review, answered or not;
• Apple Maps, Bing, and Yelp, which the assistants check against Google;
• review and ordering platforms such as TripAdvisor, DoorDash, and OpenTable;
• Instagram, Facebook, and TikTok, where an assistant can see whether you're active and what customers post about you;
• Reddit and local forums, where the unfiltered version of your reputation lives.

When those sources agree with each other, are complete, and show recent reviews with owner replies, an assistant has what it needs to recommend you with confidence. When they disagree, or one of them says you're closed, or the newest review is an unanswered complaint, it has a reason to name someone else.

Why this lands hardest on smaller operators

A national chain is recommended partly on its brand and partly because hundreds of locations reinforce the same data everywhere. A smaller operator has neither advantage. Each of its locations is judged on its own public record, and that record is standing in for a brand the business doesn't have yet. It is both more fragile and more valuable for a smaller business than for a chain.

In practice, the situation at most smaller operators looks like this. Four AI assistants are answering questions about the business every day, and nobody has read the answers. The information they're reading lives in six or more separate logins, maintained by whoever had time, and it has drifted: a rebrand that reached Google but not Instagram, a catering line that reached Yelp but not Apple. Reviews come in faster than anyone replies, and the unanswered ones are the ones most likely to be quoted. There's no ad budget that fixes any of this, and no one whose job it is.

Conclusion

Five separate research efforts, run by BrightLocal, Yext, Adobe, McKinsey, and Whitespark on different samples with different questions, arrive at the same place: consumers have moved to AI assistants faster than to any channel in memory. Two independent measurements, SOCi's and Uberall's, agree that the assistants recommend very few businesses. And the citation analysis shows they choose from the public record rather than from anything on your site. When that many sources converge from different directions, the conclusion is solid, and we hold it with confidence.

What it means for a smaller operator is a gap with a specific shape. Discovery has moved to a channel nobody in the business is watching, the inputs to that channel are scattered across platforms nobody owns, and there is no money that substitutes for getting them right. Every one of those inputs is under the owner's control, which is the good news, and the reason the gap is worth closing now.

By the numbers

Sources

• BrightLocal, Local Consumer Review Survey 2026. brightlocal.com/research/local-consumer-review-survey
• Yext, 2026 Consumer Search Behaviors Report.
• Adobe, 2026 AI and Digital Trends: Consumer Report (with Oxford Economics; 4,000 consumers, Oct–Nov 2025). business.adobe.com
• McKinsey & Company, The new front door to the internet (Nov 2025).
• Whitespark, Local Search Ranking Factors 2026. whitespark.ca/local-search-ranking-factors
• SOCi, 2026 Local Visibility Index. soci.ai/insights/lvi
• Uberall assistant benchmark, via Bloom Intelligence, AI Restaurant Discovery 2026: The Data (July 2026).
• Semrush, AI Search Traffic Study (2025).
• Omniscient Digital, analysis of AI citations (2026), via Elev8 Operations (July 2026).
• Local Falcon, The AI Visibility Crisis (March 2026).