If you have asked ChatGPT or Gemini about your own business lately, it probably knew who you were. That is the good news, and it is also the trap. A new monthly benchmark launched in October 2026 by Insites, audited across seven AI assistants in September, found AI could identify 99.1% of the 8,016 local businesses it checked. Being invisible to AI is no longer the real risk for most small businesses.
The same benchmark found something far less comfortable: 40.3% of those businesses were never recommended for a single one of their services, even though AI knew exactly who they were. Knowing you exist and telling a customer to call you turn out to be two very different things — and the gap between them is where most Northeast Indiana businesses are quietly losing to a competitor down the road.
This post walks through what the newest 2026 data actually shows about AI and local search, why your business can be “known” and still invisible at the moment it matters, and the concrete steps that move a business from known to recommended.
Key Takeaways
- AI assistants can identify 99.1% of local businesses, but 40.3% are never recommended for any service — knowing and recommending are separate problems.
- ChatGPT knows a business by name 96.0% of the time but recommends it only 27.8% of the time.
- Google shows an AI Overview on just 4.0% of “near me” searches but 75.4% of specific-need searches, and the rate varies sharply by business category.
- Google began turning some local knowledge panels into chat-style AI Overviews in September 2026, pulling from a business's wider web presence, not just its Business Profile.
- AI gets basic facts wrong often: published phone numbers were correct only 53.2% of the time, and 94.3% of businesses had at least one wrong detail somewhere.
- Reviews and listing accuracy, not technical AI schema tricks, are the strongest signals that move a business from known to recommended.
Known vs. Recommended: Why the Gap Matters
The Insites Local AEO Benchmark, a free monthly measure of how seven AI assistants identify and recommend local businesses, draws a sharp line between two questions. A branded question names the business and checks whether AI can find it and describe it accurately. An unbranded question asks what a real customer would ask — “best plumber in town” — and records which businesses actually get named back.
Those two questions produce very different answers. Across the benchmark's first edition, AI found 99.1% of businesses on at least one assistant, but 40.3% were never recommended for any of their services. Broken down by assistant, ChatGPT knew a business by name 96.0% of the time and recommended it only 27.8% of the time — a gap of nearly 70 points between recognition and recommendation.
Insites frames the business implication plainly: telling a local owner they are “invisible to AI” is easy to disprove in one prompt and makes for a weak pitch. Telling them AI knows exactly who they are but sends the customer to a competitor when it matters is, for most businesses, the true and far more useful story.
When Does Google Even Show an AI Overview for a Local Search?
Before worrying about whether AI recommends you, it helps to know how often an AI Overview even appears for the kind of search your customers run. A September 2026 study by Synup pulled 53,900 Google results pages across 49 business categories and 250 U.S. cities to find out, and the answer depends entirely on how the question is phrased.
On desktop, Synup measured AI Overview frequency across four query forms: “[category] near me” triggered an overview on just 4.0% of pages, “[category] [city]” on 8.6%, “best [category] in [city]” on 27.0%, and a “best” search paired with a specific need, such as “for water heater repair,” on 75.4%. The pattern is consistent: the more specific and need-driven the search, the more likely Google answers with AI rather than a plain results list.
Business category swings the numbers even further. Synup found AI Overviews on 80% of “best” searches for home builders, 69% for banks, 69% for accounting firms, and 62% for financial advisors — but only 3.6% for dentists, 1.2% for restaurants, and 0.4% for cafes. A dental practice and a home-remodeling contractor are playing on entirely different boards. The study also found the three-business local map pack appeared on 91 to 97% of pages regardless of query form, making it the single most stable feature in local search — and between 45% and 61% of what AI Overviews cited pointed back to business websites or Google Business Profiles, which is exactly why those two assets deserve the first audit.
Why Local Knowledge Panels Are Turning Into AI Overviews
The ground keeps shifting under these numbers. A detailed 2026 change log from Silverback Marketing notes that on September 17, 2026, Google began labeling some local knowledge panels as AI Overviews, adding a chat-style “Show more” that describes a business by drawing on its broader web presence — not only the facts in its Google Business Profile.
That single change widens what feeds the answer a customer sees about you. It is part of a bigger pattern documented across 2026: Google merged AI Overviews and AI Mode into one continuous experience in May, pushed past 2.5 billion monthly AI Overview users by June, and on August 28 started auto-expanding some AI Overviews into full-screen AI Mode answers for certain queries, with the old “Show more” button simply gone. The result is a results page where, for a growing share of searches, there is nothing to rank above — there is only a question of whether you are a source the overview cites.
For a local business, the practical read is the same one we laid out in our breakdown of Google's Collected Info rollout: your Business Profile is no longer the only thing Google is reading about you before it answers a customer's question. Your website, your reviews, and whatever else is publicly findable about your business are all part of the same pool.
Where AI Gets Your Business Details Wrong
Being known is not the same as being described accurately, either. The Insites Benchmark checked basic facts against what each assistant reported and found published phone numbers correct only 53.2% of the time. Across the full sample, 94.3% of businesses had their name, website, or phone number wrong in at least one assistant's answer.
Just as telling: only 13.2% of the sources AI cited about a business were the business's own website. Most of what an AI assistant says about you comes from somewhere else — a directory, a review site, a mention on someone else's page — which means errors can originate well outside your control and still end up in the answer a customer hears. This is the same upstream problem we cover in how Google Maps ranks Fort Wayne businesses: the signals compound, whether the surface is a map pack or an AI answer.
| Metric (Insites, Sept. 2026, n=8,016) | Result |
|---|---|
| AI can find the business on at least one assistant | 99.1% |
| Never recommended for any service | 40.3% |
| ChatGPT knows the business by name | 96.0% |
| ChatGPT recommends the business | 27.8% |
| Published phone number reported correctly | 53.2% |
| Had a wrong name, website, or phone number somewhere | 94.3% |
| Citations pointing to the business's own website | 13.2% |
How to Close the Gap Between Known and Recommended
The least useful reaction to this data is chasing AI-specific tricks. Insites' own research, cited in the same benchmark, found that readiness checks like FAQ schema, LocalBusiness markup, or an llms.txt file are cheap and sensible hygiene — but they are not what moves a business from known to recommended. Review volume, complete and accurate listings, and a genuinely useful website do most of the work.
A Practical Closing-the-Gap Checklist
- Ask AI the unbranded question. Open ChatGPT or Gemini and ask what a real customer would ask — “best [your trade] in [your town]” — and see whether you're named. That is the test that matters, not whether the assistant recognizes your name when you type it in.
- Audit your phone number and name everywhere. With correct phone numbers reported only 53.2% of the time in the benchmark, verify the number on your website, your Google Business Profile, and major directories all match exactly.
- Grow review volume deliberately. Since reviews are the strongest signal Insites has found separating recommended businesses from ignored ones, build a simple, consistent process for asking recent customers to leave one.
- Make your own website the clearest source. Since only 13.2% of AI citations point to a business's own site, state your services and service area in plain, specific language on your site itself — the same direct-answer approach we cover in our Gemini local AI citations playbook.
- Check your category's AI Overview rate before over-investing. A dental practice chasing AI Overview visibility is fighting a 3.6% baseline; a home builder is fighting an 80% one. Know which fight you're actually in.
How Northeast Indiana Businesses Should Respond
The category variance in the Synup data matters directly for Auburn, Fort Wayne, and the surrounding DeKalb and Allen County market. A trades business — HVAC, plumbing, lawn care — sits closer to the high end of AI Overview exposure once a customer's search includes a specific need, which is exactly how an emergency call is phrased. A dental or medical practice sits at the low end for “best” searches but still depends heavily on the local map pack, which Synup found present on 91 to 97% of pages no matter how the search was worded.
Either way, the fix is the same foundation: accurate, consistent facts everywhere a customer or an AI assistant might look, and enough recent reviews that the signal is unmistakable. That is the same ground we cover in what 1.8M Google Business Profiles reveal about local SEO — complete, accurate, well-reviewed profiles keep winning, whether the surface reading them is a human or a model.
If you want to see exactly where your business stands — known, recommended, or somewhere getting the facts wrong — that audit is the first thing we run with Northeast Indiana clients before touching anything else.
Find Out If AI Is Actually Recommending You
Button Block audits how ChatGPT, Gemini, and Google AI Overviews describe and recommend Northeast Indiana businesses, then fixes the listing, review, and website gaps that separate known from recommended.
Frequently Asked Questions
- Does AI actually know my business exists?
- Almost certainly yes. The Insites Local AEO Benchmark, built on 8,016 local businesses audited across ChatGPT, Gemini, Copilot, Perplexity, Grok, and Google's AI Overviews and AI Mode in September 2026, found AI could identify 99.1% of the businesses it checked on at least one assistant. Being unknown to AI is rare. Being recommended by it is a different story.
- What's the difference between AI 'knowing' and 'recommending' my business?
- Knowing means an assistant can find and describe your business when asked by name. Recommending means it names your business, unprompted, when someone asks an unbranded question like "best plumber near me." The Insites Benchmark found 40.3% of businesses it checked were never recommended for a single service, even though AI knew who they were.
- Why does ChatGPT know my business but send customers to a competitor?
- The Benchmark found ChatGPT knew a business by name 96.0% of the time but recommended it only 27.8% of the time. Recommendation depends on different signals than recognition — reviews, listing completeness, and how clearly your website states what you do and where, not just whether your name appears somewhere on the web.
- Does Google show an AI Overview for every local search?
- No, and the rate varies enormously by how the search is worded. A Synup study of 53,900 local results pages in September 2026 found AI Overviews appeared on just 4.0% of "[category] near me" searches but 75.4% of searches that named a specific need, such as "best plumber in Fort Wayne for water heater repair." Category matters too: home builders saw AI Overviews on 80% of "best" searches, dentists on only 3.6%.
- Is my phone number or address likely to be wrong in AI answers?
- It's a real risk. The Insites Benchmark found AI assistants got a published phone number right only 53.2% of the time, and 94.3% of audited businesses had their name, website, or phone number wrong in at least one assistant's answer. Checking this across assistants is now a basic hygiene task, similar to checking your Google Business Profile for errors.
- What actually moves a business from known to recommended?
- Insites' own research points to review volume as the strongest signal, followed by complete, accurate listings and a website that clearly states services and location in plain language. Technical hygiene like FAQ schema or an llms.txt file is cheap to do but, by their data, does far less of the work than reviews and accuracy.
- Should I just add AI-specific schema or an llms.txt file?
- Do it, but don't expect it to move the needle on its own. The Insites Benchmark describes these readiness checks as sensible hygiene, not what gets a business recommended. Reviews, listing accuracy, and a genuinely useful website carry far more of the weight — the same fundamentals that drive traditional local SEO.
Sources & Further Reading
- Insites: insites.com/blog/introducing-the-insites-local-aeo-benchmark — Introducing the Insites Local AEO Benchmark: monthly data on how AI recommends local businesses (8,016 businesses audited, September 2026)
- Tech Company News: techcompanynews.com/when-does-google-show-ai-overviews-for-local-searches — When Does Google Show AI Overviews for Local Searches? Findings From Synup's 53,900-Page Study
- Silverback Marketing: silverbackmarketing.com/google-ai-overviews-updates-2026 — AI Overviews in 2026: every major Google change, including the September 17 local knowledge panel shift