Nearly half of the people hunting for local business recommendations this year did not start on Google. They asked an AI. Forty-five percent used ChatGPT, Gemini, or Perplexity to find a local business in the past year, up from six percent the year before. That is the fastest change in how customers pick anyone since online reviews became the thing everybody checked.
Here is what shifted underneath it. Ask ChatGPT for a good marketing agency or a roofer near Langley and it does not hand you ten links to sort through. It reads the web, recommends a short list, and gives you two or three names with a line on why — your AI salesforce is already describing your business to the customer, in words you never wrote. You are in that line or you are invisible. There is no page two to claw back from.
So every owner needs one answer. How does an AI decide which businesses make the list?
It is not proximity. Owners assume the closest business wins, the way Maps used to work. AI assistants do not care that your office sits one suburb over. They care whether your work is documented, consistent, and easy to check. ChatGPT pulls reviews from several sites, confirms your name, address, and phone match everywhere they appear, reads your structured data, and weighs how deep your service pages actually go. Then it cross-checks Reddit, YouTube, and industry sources before it recommends you. Since 88 percent of people fact-check what an AI tells them, a thin or contradictory footprint gets caught twice.
This is where most local businesses are quietly losing. Three short pages. Listings that disagree on the phone number. No schema telling the machine what they do or where they serve. A human visitor forgives all of it. The model reading the page does not. It moves on to the business that made itself legible.
Two things follow, and both matter for any BC owner.
The first is that being good is no longer enough by itself. The business that the AI recommends is the one whose competence is written down where the machine can read it. Detailed service pages. Consistent listings. Reviews in more than one place. Schema that names the work and the service area honestly. None of it is glamorous. All of it is what an AI decides on.
The second is that location stopped being the moat it once was. We build and run this for clients across Canada and into the US from a base in BC, and an AI has never recommended a weaker option because the team behind it sat in a different city. Across more than a thousand projects in twelve years, the rule held long before any of this: people hired the team that proved it could do the work, not the one with the nearest postal code. AI search just made it official. It decides on documentation, not distance, so which businesses get found now turns on who is demonstrably good at the job. A well-run shop an hour away beats a silent one across the street.
Want to know whether an AI would recommend your business today? Ask one. Open ChatGPT, describe your service and your area, and see if your name comes up. If it does not, the gap is almost always the same fixable set, which most owners can close in weeks: shallow content, inconsistent listings, missing structured data. That is the ground our SEO and AEO work covers, and right now it is the cheapest edge on the table, because most of your competitors have not noticed the rules changed.
The businesses that figure this out in 2026 will not be the biggest ones. They will be the ones an AI finds easiest to recommend.