Your Google review count decides if AI recommends you

The one number AI is really counting

The businesses an AI assistant names carry about ten times the Google reviews of the ones it leaves out. That gap comes from a fresh study of 10,000 US businesses, and it points straight at the plainest fix on your list. So if a customer asks ChatGPT or Gemini for a good business in your town and your name never comes up, your review count is most likely the reason. This post covers four things: what the study found, why review volume reads as a safety check to an assistant, a five-minute test you can run before the weekend, and the simple habit that closes the gap.

A firm called Insites ran the numbers on 10,000 US businesses across ChatGPT and Perplexity. They lined up more than 200 different things a business might do online, from reviews to page speed to website structure, against whether AI actually recommended it. That kind of match-up is called a correlation, which just means two things tend to show up together. Nothing fancy. Reviews were the strongest signal by a wide margin, meaning the trait that lined up most often with AI picking a business.

The businesses AI recommended had about 133 Google reviews on average, and the ones it left out had about 11. No other single factor came close, including page speed and website code. Reviews, by the raw count, did most of the work. Insites isn't the only group seeing this. Coverage of a separate 2026 index by Search Engine Land reported that AI assistants name a much narrower set of businesses than Google's local results do.

You can read that as good news, and you should. The thing separating the named businesses from the skipped ones is cheap and low-tech. It's the least glamorous habit in local marketing: asking happy customers to leave a review.

Why a machine leans on your review count

Think about what an assistant is doing when someone types who is a good plumber near me. It has to pick a few trustworthy names out of many possible options, and above everything else, it wants to look right to the person asking. So it runs a fast confidence check on each business, quietly, in the second before it answers. A shop with 130 reviews reads as a solid bet, a known quantity backed by hundreds of real people. When the count is only 11, the assistant isn't sure, so it plays it safe and sends the customer somewhere else.

Here's how that plays out in real life. A woman new to town asks Gemini for a good vet near her. Three clinics have full profiles and hundreds of reviews between them. Yours has 14. The assistant has no reason to gamble on the unknown option, so it hands her the three busy names and moves on. She books one of them that afternoon. You never knew the question was asked.

Reviews work as a stand-in for trust. A business with a big, current pile of reviews looks like a real, busy operation, the kind plenty of people have used and liked and come back to. The assistant can't visit your shop or call your old customers, so it reads your review count as the next best proof. Volume matters more here than a flawless star average, because a big number is harder to fake than a handful of five-star posts.

The same study found two other things worth knowing. Businesses that show up in Google's Local Pack (the little map with three businesses that appears in local search) were recommended 3.5 times more often. And when AI described a business, it pulled from that business's own website 72% of the time, more than any other source. Both findings point back to the same foundation: a full, active Google presence paired with a clear, current website gives the assistant something solid to lean on when it decides who to name.

One more finding, just for peace of mind. The technical website tweaks that agencies love to sell, like page-speed work and special code tags, showed almost no link to whether AI recommended a business. So if that expensive technical work has been nagging at you, you can set it aside and put your time into reviews instead.

The five-minute check you can run today

You don't need a paid report to see where you stand. Grab your phone. Open ChatGPT or Gemini. Here's the whole test:

1. Ask the question a customer would ask. Type something like who is a good dentist near me or best HVAC company in town.
2. Write down the names the assistant gives back. Note whether yours is one of them.
3. Look up each named business on Google and check its review count.
4. Compare those numbers to yours.

In the audits we run, most owners find the same thing. The names the assistant picked have far more reviews than they do, sometimes ten times more. If the businesses it named sit at 90, 120, or 200 reviews and you're parked at 20, you've found your gap. That's the number to close.

Don't panic if the gap is wide. A business at 20 reviews can climb fast. The businesses AI names today were sitting right where you are at some point, and they got there by asking, one customer at a time. Here is where to start.

The habit that moves the number

Closing the gap is simple to describe, though it takes some discipline to keep up. It comes down to asking every happy customer for a review, day after day, without dropping the habit once it turns boring. Most businesses collect a handful of reviews by accident, feel fine about the number they have, and then quietly stop asking. The ones AI names built a steady routine instead, and kept it running long after it stopped feeling new.

A routine that works looks like this. Pick the moment a customer is happiest, right after a good visit or a finished job. Ask in person when you can, because a spoken ask lands better than a sign on the wall. Then send a short text or email with a direct link to your Google review page, so leaving one takes about ten seconds. Get every person on your team doing the same. A shop that asks ten customers a week can add 20 or 30 real reviews in a quarter. Some weeks will fall short of that, and that is fine.

Not every customer will follow through, and that's fine. If two out of every ten you ask actually post, a steady ask still stacks up, week after week, month after month. The point is to make the ask normal, so it happens on every job instead of the rare great one.

Reply to the reviews you get. The good ones and the hard ones both. It shows customers, and any watching software, that a real person runs the place. Most people read reviews before they choose a local business, and many trust them nearly as much as a tip from a friend (BrightLocal). A steady trickle of fresh reviews works better than one big week, so aim for a habit you can keep every week.

AI recommendations are turning into a real way customers find local businesses, right next to Google and word of mouth. The owners quietly building up their review counts right now, one satisfied customer at a time, will be the ones assistants name a year from now. The ones parked at 15 or 20 reviews will keep watching the shop across town get picked. The work is boring and it's cheap, which is the best kind of work to have in front of you.