AI Names 12 of the 97 Local Businesses Google Shows. We Measured It

Two numbers explain most of what is happening in local search right now, and we measured both ourselves. Ask Google Maps for a category in a city and it returns about 97 businesses. Ask ChatGPT, Claude, Gemini and Google's own AI Overviews the same question, the way a customer would phrase it, and between them they name 12 of those businesses. The other 85 exist, rank, collect reviews, and never enter the conversation.

That comes from our own study of 48 US markets: 4,657 businesses, 1,440 AI answers, collected in a single pass with the method written down before any data was gathered. Only 11.94 percent of the businesses Google was willing to show got named by any assistant at all.

It is worth sitting with that gap, because it is not a rounding error or a temporary quirk of an early product. It is the structural result of two completely different answer formats.

Why the gap is this large

A Google results page has room. There is a map pack with three to seven listings, ten or more organic links below it, review site snippets, and often ads. A business does not need to be the best in its category to appear somewhere on that page; it needs to be decent, roughly complete, and not badly outranked. That generosity of space is why a hundred businesses can surface for a single query.

A generative AI answer has almost no room. When someone asks "best [category] near me," the assistant does not return a page of options to compare. It reads across whatever sources it trusts for that category and hands back a short, confident answer, typically naming two or three businesses and nothing else. There is no tenth listing, no second page, no map pack runner-up. The model either names you or it does not.

The engines also disagree with each other more than most people expect. In our data Google AI Overviews named 8.07 percent of the businesses Google itself returns, ChatGPT 4.42 percent, Gemini 2.88 percent, and Claude 1.98 percent. Being named by one assistant tells you very little about the other three.

It is not the same competition as Google ranking

This is the part that surprises the most business owners: ranking well on Google does not transfer proportionally to being named by AI. A business can sit in the map pack, look strong on paper, and still never appear in the AI Overview generated directly above its own listing. The mechanisms genuinely differ. Google's ranking rewards a well-built website with decent technical SEO and inbound links. A generative answer rewards a business with strong, current, third-party corroboration, meaning reviews, directory presence, and citations the model has learned to trust, more than it rewards the website itself.

That is also why the businesses that get named are not always the objectively best ones. They are the ones that look most unambiguously trustworthy to a system reading the web at a glance: complete profiles, recent reviews in volume, consistent listings across the sources that get cited, and clean structured data confirming basic facts. Our piece on GEO vs SEO goes deeper on where the two disciplines overlap and where they genuinely diverge.

Category turned out to matter more than city. Across our eight cities the spread was narrow, but gyms and coffee shops were named at roughly 14 percent while hair salons sat near 9.7 percent. Where you are matters less than what you sell.

What this means for your business, concretely

If you have not checked, you do not know which side of this gap you are on, and your Google ranking will not tell you. A business can be page-one on Google and completely invisible to ChatGPT for the exact same query, with no warning, because nothing about a stable Google rank hints at generative absence.

The only way to know is to ask the assistants the questions your customers actually ask, more than once, since generative answers vary between runs. Our free quick check tool does this in a couple of minutes and shows you whether you are named. If you are not, the fix is not mysterious, it is the same handful of things across every category we cover: a complete Google Business Profile, a steady stream of real reviews, presence on the sources AI actually cites for your category, and clean schema markup. The full ordered version of that work is in how to get recommended by ChatGPT.

Why we replaced the number that used to be here

This article used to open with a different pair of figures, 1.2 percent and 35.9 percent, taken from a third-party summary we could not audit. They were widely repeated and they made a good headline. We could not check them, so we stopped using them and measured the question ourselves instead.

Our number is not a drop-in replacement, and that difference is the point. The old figure claimed a share of all local businesses. Ours is a share of the businesses Google Maps actually returns for the same query, which is a smaller, stated, checkable universe. The full rule set, every market row, the limits, and the raw counts are on the research page, including the parts that went against us.

The number that matters going forward

Twelve out of ninety-seven will not stay fixed. As more people route local decisions through an assistant instead of a search results page, being one of the few names an AI offers stops being a curiosity and starts being where a growing share of new customers actually comes from. The businesses treating this as worth measuring now are the ones that will not be surprised later.

LocalSeen scans ChatGPT, Claude, Gemini, and Google AI Overviews and tells you exactly why a competitor is named when you are not. If you want to find out where your business stands, run your first scan.