Your Customers Stopped Searching. They Started Asking.
AI assistants are now the third most common way people find a local business, and they name two or three before stopping. There are exactly two ways to be one of them, and advertising drives both.
Date
Author
Jason Demek
Read
2 to 3 Min
A homeowner in Ross Township needs a roofer. Two years ago that started with a Google search and a screen full of options: a map with a dozen pins, sponsored results across the top, ten blue links underneath. The homeowner browsed, compared, and chose.
Today that same homeowner increasingly opens ChatGPT or Gemini and types a sentence. What comes back is not a list. It is a short answer naming two or three companies.
That sounds like a closing door. The data says it is closer to a front door, and the businesses that understand the difference have a real advantage in this market right now.
What actually changed
BrightLocal's 2026 Local Consumer Review Survey, fielded across a representative panel of 1,002 US adults, found that 45% of consumers used an AI tool to get a local business recommendation in the past year. The year before, that figure was 6%. AI is now the third most-used source for local recommendations, behind only Google and Facebook, and ahead of Yelp and TripAdvisor. Over the same period, Google's share of local business discovery slipped from 83% to 71%.
The shortlist those tools produce is genuinely narrow. SOCi's 2026 Local Visibility Index studied more than 350,000 locations belonging to roughly 2,700 multi-location brands, the kind of companies with real marketing budgets and full-time marketing staff. Google's local 3-pack surfaced 35.9% of those locations. ChatGPT recommended 1.2%. Gemini reached 11%, Perplexity 7.4%.
Read that carefully. Those are not mom-and-pop operations getting left out. Those are national and regional brands with dedicated teams, and the assistants still name almost none of them.
Which raises an obvious question. If the recommendation engine is that selective, how is anyone winning?
The answer is not the decision
Here is the finding that reframes the entire conversation, and it comes from the same BrightLocal research.
Among consumers who have used AI for local recommendations, 63% say they trust what it tells them. But 88% go and verify it anyway.
Read those two numbers together and the behavior is obvious once you see it. People ask the assistant, get two or three names, and then go check. They look up the business. They read reviews. They look at the website. They ask someone. They decide whether the name in front of them feels like a real, established, credible company in their area.
The AI answer is not the purchase. It is the front door, and it opens onto the same evaluation process consumers have always run. Nearly nine in ten customers are still making the actual decision somewhere you can reach them.
That is the opening, and it is where advertising does its work.
Two ways to win, and both are buyable
Survive the verification
Every business named on that shortlist then gets checked. Recognition is what determines how that check goes.
When a consumer verifies an unfamiliar name, they are starting from zero and looking for reasons to be reassured. When they verify a name they already know from Pittsburgh's Action News 4, from Friday night football, from a schedule that ran often enough to register, the check is not an investigation. It is a confirmation of something they already half believed.
This is the oldest advantage in advertising and the AI era did not weaken it. It concentrated it. The verification step is now a required stage in nearly every local purchase, and familiarity is the single biggest thing determining whether a business clears it.
Being present on WTAE.com matters here for the same reason. The verification stage happens on the open web, where consumers check the business against whatever credible local sources exist. A branded content feature or sponsorship presence on the region's news site is a credibility artifact sitting exactly where that check happens, on a name Pittsburgh already trusts.
Skip the shortlist entirely
The stronger play is to never be subject to the shortlist in the first place.
There are two kinds of local queries. An unbranded one sounds like "best HVAC company near Cranberry." That is the 1.2% problem, and it is crowded.
A branded one sounds like "is [company name] any good." That is a customer who already decided who they are interested in. The assistant is not selecting for them. It is confirming a name they walked in with.
Branded demand is created by advertising, and broadcast creates it at a speed and scale nothing else in this market matches. Every dollar of brand building you have ever done is machinery that converts unbranded lotteries into branded confirmations. That was always worth something. It is worth measurably more now that the alternative route narrowed to three names.
What a plan built for this looks like
A local media plan has three jobs, and most plans are only funding the third.
Be known. Broadcast. News, live sports, and franchise programming like Operation Football still assemble the largest single-moment audiences in this market, in an environment where people are watching rather than scrolling. This is what produces branded queries and what carries a business through the verification step.
Be findable. Station digital and your owned properties. Presence on the region's trusted news site, plus the fundamentals that cost nothing: a complete and current business profile, reviews that describe specific services rather than just awarding stars, consistent name and address data everywhere it appears. SOCi's research found that assistants answering local questions lean heavily on Google Maps, business websites, Yelp, and Facebook, so those four are not optional.
Be captured. Streaming, connected TV, targeted digital video, display, and search. This layer converts demand that already exists, and it works. It simply was never built to create the demand it captures, and a lot of local plans have spent several years asking it to.
Start here
Ask ChatGPT, Gemini, and Perplexity for the best business in your category in your part of the region. Try a few phrasings. Five minutes tells you where you stand.
Then split your search reporting into branded and unbranded volume. Most local businesses track it as one number. Branded volume is the clearest signal you have that your brand building is working, and the best predictor of how you will hold up as more discovery moves into the answer layer.
The search box did not close. It narrowed the entryway and left the decision exactly where it always was. The businesses that win are the ones customers already recognize when they get there.