Lakefly Productions Logo | White | Transparent | Wide

How AI Assistants Pick Which Local Businesses to Recommend

By , PresidentPublished Updated

Remember waiting to be picked?

Two captains. A row of kids on the blacktop before kickball. One by one, names get called, and you find out exactly where you stand.

Everyone has a spot in that line. Some of us were picked first. Some of us landed somewhere in the middle and told ourselves it didn't matter. Some of us were the last name called, and we still remember it.

That feeling is the best way I've found to explain what's happening to local search right now. There are two rules in stockcar racing that sit behind that feeling though. Ricky Bobby lives by one in Talladega Nights: "If you ain't first, you're last." The other, "second place is just the first loser," is a line most often credited to Dale Earnhardt. Both are funny on a racetrack. They're less funny when an AI assistant is the one choosing and your business never gets called.

For twenty years, search worked like a class lineup: everybody got picked, and everybody got a number. Now more people are asking ChatGPT, Gemini and Google's AI results who to call. BrightLocal's March 2026 survey, as summarized by Getfluence, found that 45% of US consumers used an AI tool to find a local business in the past year, up from 6% a year earlier. And a single captain does the picking.

This post explains how that captain decides, how it's the same game as SEO and how it isn't, and what you can do to get picked sooner.

Traditional SEO is the whole-class lineup

Ask ChatGPT, Gemini or Google's AI results for the best provider near you and you don't get a lineup or a finishing order. You get an answer, and the answer names a few businesses. Sometimes one. It's victory lane: the camera goes to the winner, maybe a couple of others, and the rest of the field never makes the broadcast.

Here's how short that roster is. SOCi's 2026 Local Visibility Index found that ChatGPT recommended about 1.2% of the business locations it tracked. Google's local three-pack showed about 35.9%. Very different odds of being picked.

There are really two captains, and they pick a little differently:

Answer engines (AEO)

Google's AI Overviews and AI Mode sit on top of Google's own index. Google says they use "query fan-out," running several related searches to find supporting pages.

Generative engines (GEO)

ChatGPT, Gemini and Claude combine what the model already knows with a live web search when you ask a question that needs one, then write an answer and cite what they found.

One more twist. Being cited is not the same as being picked. In a September 2026 study of 3,118 AI answers, when an engine cited a company's own website, it named that company in the answer only about 27% of the time. In kickball terms, showing up to the game gets you on the field. Hearing your name called is something else.

What the captain looks at before picking

Nobody publishes the captain's scorecard, and it differs by assistant. But the research and Google's own guidance point at the same handful of things. Think of them as the reasons a captain picks one kid over another.

  • You're clear about what you do.

The captain picks the kid who obviously plays the position. Say what you do and where you do it, in plain text on your pages. Google's guidance is to keep important information as text, not buried in images.

  • Everyone agrees who you are.

Your name, address, phone and hours should match across your website, your Google Business Profile and your listings. SOCi found business profile accuracy of about 68% on ChatGPT and Perplexity, versus 100% on Gemini, which it says draws directly from Google Maps.

  • Your teammates vouch for you.

Reviews work more like a threshold than a ranking. SOCi reports that locations ChatGPT recommends average about 4.3 stars, and that ones near 3.4 stars with very low response rates are effectively left out.

  • Other captains talk about you.

Mentions in local press, industry sites and "best of" lists carry weight. In one summary of several studies, editorial mentions made up about 27% of ChatGPT's sources, and the Continuum study found "best X" lists in about 12% of citations.

  • You answer the question being asked.

Pages that open with a direct answer, then back it up, are easier to quote than pages that circle the point. An FAQ written in the way customers actually phrase things is one of the simplest ways to do this.

  • You're easy to reach.

Your pages have to be indexed and eligible to show a snippet, and not blocked by robots.txt or your host. Google says that's the baseline for appearing in its AI features, and it doesn't guarantee it.

Notice what's on that list: the same things a good local SEO program already works on. That's not a coincidence.

Where the racing metaphor holds, and where it breaks

The idea that if you’re second, you’re just first loser runs straight from the racetrack to Ricky Bobby to AI search. when one captain names a few businesses, being fourth in line can mean not being named at all.

Here's the whole comparison side by side.

Traditional SEOAEO & GEO
Who picksA search engine ranks every pageAn answer engine writes one response
What the customer seesAn organized list from best to worst of resultsA short answer naming a few businesses at most
Spots availableDozens, plus three in the map packA handful at most, often just one
Seeing where you standEasy: rank trackers show your positionHard: it varies by question, assistant, and is different each time
What moves you upRelevance, authority, technical health, linksThe same foundations, plus clarity, consistent data, reviews, and mentions
StabilityFairly steady from day to dayIn the Continuum study, only about 30% of the local businesses named showed up in all three repeat runs of the same question

The good news is that the two games share a playbook. Google says its AI features need no special optimization beyond standard SEO, and most of the picking criteria above are things a strong SEO program already covers.

The metaphor breaks in three places, and they matter:

The captain changes.

A kickball captain picks the same kid first every time. AI assistants don't. Different assistants pick differently, and the same assistant can give you a different list tomorrow. Continuum found that about 44% of named local businesses appeared in just one of three runs.

Your Google rank doesn't carry over.

SOCi's index found that fewer than half of the brands leading traditional local search also appear in AI recommendations. Being picked first on one day says little about the next.

Second place isn't always the end.

People check the answer. BrightLocal's March 2026 survey, as summarized by Getfluence, found that 88% of AI users check the cited sources. If you're named and your site and reviews hold up, you get the call. If you're named and your information is wrong, you lose it.

How to get picked earlier

You can't bribe the captain, but you can be the obvious pick. Start here, in roughly this order:

Finish your Google Business Profile

Make your facts match everywhere

Put the answer on the page.

Collect reviews steadily and reply to them

Get mentioned by people other than you

Clear the path for crawlers

The honest part: nobody can promise you the first pick

No one can guarantee you'll be recommended. Not a tool, not an agency promising "#1 on ChatGPT," and not us. The captains differ, they change their minds, and none of them publish the rules. It's the same answer I give about traditional SEO: nobody controls the algorithm, so the honest commitment is sound work and transparent reporting.

A caution on the numbers in this post, too. They come from snapshots, mostly of US searches, and some come from companies that sell AI-visibility services. Treat them as direction, not gospel.

Your next move

Being picked last is a rotten feeling. The difference here is that you can do something about it before the next game. Run the ten-minute check, then work through the six steps in order.

If you need a second set of eyes, let’s talk.

AEO/GEO FAQs

SEO gets your pages ranked in traditional search results, while AEO and GEO get your business named in AI answers. AEO (answer engine optimization) targets Google's AI Overviews and AI Mode. GEO (generative engine optimization) targets assistants like ChatGPT, Gemini and Claude. They overlap heavily: the same clear content, consistent business details, reviews and credible mentions help all three.

Mostly no. Google says AI Overviews and AI Mode need no special optimization beyond standard SEO, and many signals the assistants rely on, like accurate listings, reviews and clear service pages, are already part of good local SEO. What changes is how you measure: you track whether you're named in answers, not only where you rank.

Because the answers aren't a fixed ranking. In a September 2026 study that asked the same local-service questions three times, only about 30% of the businesses named appeared in all three runs. That's why one screenshot proves little, and a log kept over several days tells you far more.

A few bad reviews won't, but a weak overall pattern can. SOCi reports that locations ChatGPT recommends average about 4.3 stars, and that locations near 3.4 stars with very low response rates are effectively left out. A steady flow of new reviews, with replies, is the fix.

There's no reliable timeline. Traditional SEO usually shows early movement within about three months and stronger gains over six to twelve, and AI visibility rests on the same foundations. Anyone quoting an exact date for being recommended by an assistant is guessing.