They Googled You. Then They Asked ChatGPT.
Finding you is no longer the last step before calling you. A growing share of customers now take the names Google gave them and ask ChatGPT about them: which of these is best, is this one trustworthy, who would you pick. The answer works like a referral from a friend who has read everything ever written about you. And you pass that check, or lose the call, without ever knowing it ran.
Here’s what the double-check looks like, and how a business survives it.
Three tabs open, and a new tab
Tuesday, 9 PM. Kitchen table, laptop open, house finally quiet. The fender-bender from last month has turned into physical therapy and an insurance company that stopped returning calls. Time to get an attorney.
Google produced three names, and all three tabs look fine. Confident websites. Decent ratings. Similar promises. Which is the problem, because there’s no obvious way to choose between three fine-looking law firms at 9 PM.
A year ago this person texts a friend and sleeps on it. Tonight they open a new tab and type: “Which of these three injury attorneys is most trustworthy?” Ten seconds later there’s a considered answer. One firm named first. Reasons attached: consistent recent reviews, a clear focus on injury cases, responses that mention fast communication. It reads like advice from someone who did the homework.
It gets treated like advice too. That’s the part most owners haven’t priced in.
The double-check is the new referral call
This behavior went mainstream fast. In BrightLocal’s 2026 Local Consumer Review Survey, 45% of consumers said they’d used ChatGPT or another generative AI tool for local business recommendations, up from 6% one year earlier. ChatGPT alone was used by 31%. AI tools are now the third most-used way people find local businesses, behind only Google and Facebook, ahead of Yelp. Among 30-to-44-year-olds, 64% have asked an AI for a business recommendation.
The old version of this moment was the referral call. You borrowed confidence from someone who knew more than you did. The AI now plays that friend, because it’s instant, it sounds sure of itself, and it appears to have read everything.
And the double-check concentrates exactly where the stakes are highest. Nobody cross-examines a coffee shop. They cross-examine the attorney, the surgeon’s practice, the contractor about to open their roof. The bigger the decision, the more soothing a second opinion feels, and the more weight its answer carries.
What the AI reads before it vouches for you
The short version: nothing private, nothing paid. The AI builds its answer from your reviews and what they say, how plainly your website answers real questions, the structured facts published behind your pages, and how consistently the rest of the web describes you. Thin or contradictory signals produce a hedged answer, and hedged answers lose calls.
We took apart the full selection process in how AI search decides which local businesses to recommend. This post is about the other side of it: what your customers do with the answer.
Passing the double-check: one story, everywhere
Here’s the mechanic worth understanding. The AI cross-references. It’s reading your Google profile, your website, your reviews, and the directories at the same time, and it notices when they disagree.
When everything tells one story, the answer comes back clean and confident, and the AI vouches for you the way a friend would. Your site says same-week appointments, your reviews say “got me in two days later,” your profile hours say you’re open when you say you are. Confidence compounds.
When the story splits, the answer says so out loud. Different service lists in different places. Hours nobody updated. A directory carrying an old address. A human shrugs past those. The AI treats them as reasons to qualify its recommendation, or to lead with the competitor whose story holds together.
The fix isn’t clever. It’s AI search visibility at its most basic: one name, one set of facts, one description of what you do, repeated everywhere, kept current.
What a mismatch costs, and why you’ll never see it
No dashboard reports “the AI hedged on you last night.” The customer at the kitchen table doesn’t email to say the recommendation went elsewhere. They asked, they read, they called someone else. From your side it’s indistinguishable from a quiet Tuesday.
That’s what makes this moment worth taking seriously now. The 6%-to-45% jump happened in one year. The double-check is quietly becoming a standard step in choosing exactly the kind of high-value work local firms live on, and the businesses winning it are the ones whose story holds up under a cross-reference they never see happen.
Questions people ask
Do people use ChatGPT to pick a lawyer or contractor? Yes, and the shift has been sudden. In BrightLocal’s 2026 Local Consumer Review Survey, 45% of consumers said they’d used ChatGPT or another AI tool for local business recommendations, up from 6% a year earlier. For people aged 30 to 44 it’s 64%. The higher the stakes, the more likely they are to want the second opinion.
What does the AI base its answer on? Public signals: your reviews and what they say, how plainly your website answers real questions, the structured facts behind your site, and whether your details match everywhere they appear. Nothing private, nothing paid. If those signals are thin or contradictory, the answer hedges.
How do I find out what ChatGPT says about my business? Ask it the questions your customers ask. Name your service and town, ask who’s recommended and whether your business is trustworthy, and read what comes back. Repeat it monthly, because the answers move as your reviews and profile change.
Where to start
The double-check runs on signals you control. The first step is seeing them the way the AI sees them. Our visibility audit includes exactly that: what ChatGPT and Gemini say right now when someone asks about a business like yours, where your story contradicts itself across the web, and which gaps would make an AI qualify its answer. Specific findings, no sales pitch.
Be the business people find, trust, and call first, even when the one running the final check isn’t a person.