Is Your Local Business Recommended by ChatGPT?

Someone looking for a plumber, a lawyer, or a dentist doesn't automatically type into Google anymore. Increasingly, they open ChatGPT and ask directly: “what's the best plumber near me?” The AI answers with one or two names — not ten blue links to compare. Is yours one of them?

A study published in January 2026 measured this question at scale, across nearly 350,000 locations: among businesses that rank well on Google, how many are also recommended when you ask ChatGPT, Gemini, or Perplexity? The numbers are stark — and they explain why so many well-ranked local businesses stay silent to AI.

What the study found

SOCi, a company specialized in marketing for multi-location brands, compared presence in Google's local pack against recommendation rates from the major AI engines, across 2,751 brands and nearly 350,000 locations. The results:

• Google local pack: 35.9% of the locations studied appear there.
• ChatGPT: 1.2% — Perplexity: 7.4% — Gemini: 11% recommendation rate.
• Getting recommended by AI is 3 to 30 times harder than ranking well on Google, according to the study.
• In retail, only 45% of the brands ranking best on Google also show up among the brands AI recommends most — meaning more than half of Google's local-pack winners, at least in that sector, are absent from AI answers.
• And when an AI does cite a business, the information it shows (address, hours, phone) is only accurate 68% of the time on ChatGPT and Perplexity, versus 100% on Gemini, which is grounded in Google Maps.

(Source: Search Engine Land, on SOCi's Local Visibility Index, published January 2026.)

That study measured multi-location brands, not independents — you can't conclude your own business has exactly a 1.2% chance of being cited. But the reverse logic holds: if chains with dedicated marketing teams and standardized listings struggle this much, an independent business with an incomplete Google listing or inconsistent information isn't starting from a better position.

Why Google and AI don't see the same thing

Google ranks pages: being 6th instead of 3rd is still a result. A generative AI chooses one or two names to say out loud: you're either in the answer, or you don't exist. That's not a stricter ranking — it's a different operation entirely.

And AI doesn't read your site the way a human skims a homepage. It relies on structured data: your Google Business Profile, your site's schema.org markup, the consistency of your name, address, and phone number across the web. Without clean data, the machine guesses — and when it guesses, it usually prefers to ignore you rather than cite something wrong.

What actually determines whether AI recommends you

None of this is a mystery. Five things come up, in one form or another, in everything seriously written on the topic:

1. The right category, not a vague one. A listing filed under a generic "Local Business" instead of the precise type — Plumber, Attorney, Dentist, HairSalon, AutoRepair, AccountingService are the actual schema.org types for each of these trades — causes the same problem as campsites listed as "hotels": the machine files you wrong, so it doesn't surface you for the right question.
2. Review volume and freshness. In the SOCi study, locations recommended by ChatGPT averaged a 4.3/5 rating, versus 3.9/5 for the full set of well-ranked Google locations. A handful of three-year-old reviews isn't enough.
3. Consistent name, address, and phone number across your site, your Google listing, directories, and social profiles. That's likely a big part of what drives the 32% error rate the study measured: an AI that finds two different addresses for the same business often chooses to cite neither.
4. A real FAQ, with real customer questions. "Do you work weekends?", "Do you take emergencies?", "What does a consultation cost?" — these are exactly the questions people ask AI, and a structured FAQ on your site is content an AI can quote verbatim.
5. Access, plain and simple. A robots.txt blocking GPTBot, or a misconfigured CDN, and none of the above matters — the AI can't even read your page.

Test your own listing (for free)

Before guessing where you stand, look at your own site the way an AI sees it. That's exactly what our free tool, Trace, does: paste your URL, get an AI-visibility score in 30 seconds, no signup required. You can then run a free, optional agent test: a real AI agent browses your site the way a customer looking for your prices or hours would, and publishes the exact path it takes — including where it gives up.

👉 Test your business with Livada Trace — free, no credit card.

Frequently asked questions

I rank well on Google. Why wouldn't AI recommend me?

Because Google and AI don't measure the same data. A strong local-pack ranking doesn't guarantee a listing an AI can actually use — AI relies on structured data (the right category, schema.org, consistent citations) that many local businesses don't expose correctly.

The study you cite is about large chains. Does this apply to my independent business?

The study measures multi-location networks, not independents — you can't transpose the exact figure to your situation. But nothing suggests an independent does better: a chain with a dedicated marketing team is structurally more likely to have the right data in place than a solo business. The only way to know where you stand is to test your own listing.

Do I need to rebuild my site to be visible to AI?

No. In the vast majority of cases, it's a matter of adding the right structured data and fixing existing inconsistencies — not rebuilding the site or changing its design.

Florin Livada — founder of Livada. 20 years on the ground in tourism, the web, and local SEO, today publisher of the tools that make a local business visible to Google and to AI. livada.io · contact@livada.io