Why ChatGPT Doesn't Recommend Your Business (And How to Fix It)

Here's a test that takes ninety seconds. Open ChatGPT, type "best [your trade] in [your town]," and read what comes back. If your business isn't in that answer, you've found a problem that didn't exist two years ago, and one your competitors probably haven't noticed either. Because 45% of consumers now ask an AI tool for local business recommendations, up from 6% a year earlier. For those people, the answer that comes back isn't the start of the shortlist. It's the whole thing.

How many people are actually asking AI for a local recommendation?

Enough that it stopped being a curiosity. BrightLocal's Local Consumer Review Survey 2026 put the number at 45% of consumers who'd used an AI tool to find a local business in the past year. The year before, it was 6%. I've watched a lot of "this changes everything" claims come and go in twenty years of doing this, and most of them didn't. This one moved faster than anything I've seen.

Source: BrightLocal, Local Consumer Review Survey 2026, a panel of 1,002 US adults surveyed via SurveyMonkey. AI questions were analyzed across the 455 respondents who reported using AI for business recommendations.

Two honest caveats before anyone reorganizes their marketing budget around that table. It's one survey, self-reported, over one year. And that last row matters as much as the first: almost everybody who gets an AI recommendation goes and checks the reviews anyway. AI isn't replacing your reputation. It's deciding which reputations get looked at.

Where do AI assistants get their information about you?

From the same public record you could look up yourself. Your Google Business Profile, your reviews, your website, the directory listings you forgot you had, and whatever local press or community pages mention your name. There's no hidden channel and no submission form.

The useful way to think about it isn't an audition. It's a background check. A model answering "who's the best electrician in Grand Rapids" is trying to name a business it can stand behind, and it gets confidence from corroboration. When your profile says you close at 5, your website says 6, and an old directory listing lists a phone number you gave up in 2019, that's not a small formatting problem. That's three sources disagreeing about a business, and the safest move is to name someone else.

Which is why the businesses that win here often aren't the biggest. They're the ones whose facts are boring and identical everywhere.

What Google actually says about optimizing for AI Overviews

This part is worth reading closely, because there's a lot of money being charged for the opposite. Google's own documentation on AI features and your website says it plainly:

"You don't need to create new machine readable files, AI text files, or markup to appear in these features." And: "There's also no special schema.org structured data that you need to add."

The stated requirement is that a page has to be indexed and eligible to be shown in Google Search with a snippet. Google adds that there are no additional technical requirements. That's the whole gate.

So if a vendor is selling you an "AI schema package" or an llms.txt file as the thing that gets you into AI answers, ask them to point at the documentation that asks for it. They can't, because it doesn't exist. We publish an llms.txt on this site because it costs nothing and might matter later. It is not why anyone gets cited today, and I'm not going to pretend otherwise to sell you something.

None of that makes structured data useless. Schema still earns rich results, still helps other engines parse what you do, and still keeps your facts machine-readable. It just isn't a secret key, and the people telling you it is are usually the ones charging for the key.

The five things that actually decide whether you get named

Strip away the noise and this is the list. It looks suspiciously like good local SEO, because that's what it is.

If you want the long version of number two, we keep a Google Business Profile checklist that walks the whole thing field by field. Number three is its own discipline, and the review capture system we use for contractors applies to any local business that finishes jobs and talks to customers.

How to check whether AI recommends you right now

You don't need a tool or a subscription for this. Ten minutes and a notepad.

Step 1: Write down the question a customer would actually type

Not industry vocabulary. Real phrasing. "Who's the best plumber in Traverse City." "I need an emergency vet near Kalamazoo." "Who should I call for a bathroom remodel in Grand Rapids." Write three or four before you open anything, so you're testing your customers' words instead of your own.

Step 2: Ask three engines the same question

ChatGPT, Google's AI Mode, and Perplexity. Use a logged-out or private window where you can, because your own browsing history can tilt an answer toward your business and give you a falsely comfortable result.

Step 3: Record who got named instead of you

Write down every business in every answer, in order. This list is worth the exercise on its own. It's your actual competitive set in the eyes of the machines your customers are asking, and it's frequently not the set you'd have named.

Step 4: Ask the follow-up question

Ask it why. "Why did you recommend those companies?" "Where did that information come from?" The sources it names are the record it trusts. That tells you which profiles, pages, and listings are doing the actual work, and it's usually a shorter list than you'd guess.

Step 5: Fix the gap the answers point at

Put the named businesses next to yours and compare the five factors above. Nine times out of ten the gap is obvious once you look: they have 180 reviews and you have 22, or their profile lists eight services and yours lists one. Start there. Run the same test again in a month and see whether the answer moved.

What doesn't work

• Buying an "AI optimization" package that's schema plus a blog post. You've read Google's position above. If the deliverable is markup and the pitch is AI visibility, the pitch doesn't match the documentation.
• Stuffing "best plumber in [city]" into your footer. Keyword stuffing didn't fool search engines in 2012 and it doesn't fool a language model now. It mostly makes your page read like spam to the humans who land on it.
• Treating llms.txt as a lever. It's a proposed standard. No major engine has committed to reading it as a ranking or citation input. Publish one if you like, but don't let anyone bill you for it as the fix.
• Buying reviews. Beyond being against every platform's terms and the FTC's rules on fake reviews, it's self-defeating here. Consistency is what earns the recommendation, and purchased reviews are the least consistent thing about a business.

The bottom line

The uncomfortable part of AI search is that it's a mirror. It reads what's publicly true about your business and repeats it back. If the record is thin, the answer is thin, and no amount of clever markup papers over that.

The comfortable part is that the fix is unglamorous and mostly free. Complete the profile. Collect the reviews. Make your hours match in all five places they appear. Say what you do in words a person would use. That's the work, and it happens to be the same work that's moved local search rankings for the last decade.

Run the ten-minute test this week. Whatever you find, you'll know more about your own visibility than most of your competitors know about theirs. If you'd rather have someone else run it and hand you the findings, that's what our free audit is for. We're based in Bellaire, and Michigan service businesses are who we build for.