How AI Picks Which Business to Recommend in a Local Area

Right, so I had a plumber in Letchworth ask me this about three weeks ago. He's been getting calls from AI searches, proper ChatGPT recommendations, but his mate down in Royston who does the same work isn't getting anything. Same Google reviews, similar websites, both been trading for years. And he wanted to know why AI was picking him and not his mate.

Fair question. And it's the same thing every business owner in North Hertfordshire is going to need to understand in 2026, because the way customers find you now is completely different to how it worked even 18 months ago.

The bit nobody's telling you about AI recommendations

When someone asks ChatGPT or Perplexity or Claude for a plumber in Stevenage, the AI isn't just doing a Google search and reading out the top three results. It's making a decision. An actual judgement call about which businesses to recommend.

And the criteria it uses? Not the same as SEO. Not even close.

I've been tracking this properly since mid-2023, running tests, watching what gets recommended and what doesn't. The patterns are pretty clear now. Some obvious, some... not what you'd expect.

Context is everything (and I mean everything)

Here's where it gets interesting. AI doesn't just look at your business in isolation. It reads the conversation. The whole thing.

Someone asks "I need a plumber in Baldock who can come out today for an emergency leak" and that's a completely different recommendation to "Looking for a reliable plumber in Baldock for bathroom renovation quotes." Same search term basically. Different context. Different businesses get recommended.

Traditional SEO doesn't work like that. You rank for "plumber Baldock" or you don't. But AI is reading intent, urgency, budget signals, complexity. All from how the question's phrased.

Which means the businesses that get recommended most often are the ones that have information covering different scenarios. Not just "we do plumbing" but actual detail about emergency callouts, about project work, about what you specialise in, response times, how you price things.

The AI needs that context to match you to the question.

Your Google reviews matter differently now

Everyone knows reviews are important. But AI reads them differently to how a human scrolls through star ratings.

It's looking for specifics. Names, details, outcomes. A review that says "John came out same day, fixed the boiler, £180, very tidy" is worth ten reviews that say "great service, very professional."

I watched this with a client in Hitchin. Electrician. Average 4.8 stars, about 90 reviews. Most of them short and generic. Wasn't getting AI recommendations. We got him to start asking customers to mention what the actual problem was and how it got solved. Two months later, ChatGPT's recommending him for specific electrical issues because it can see from the reviews that he's dealt with that exact problem before.

The AI's basically reading your reviews as case studies. If someone asks about a specific issue and your reviews mention you've solved that issue, you're in. If your reviews are all "five stars, great guy" then the AI's got nothing to pattern match against.

Being specific beats being comprehensive

You know what doesn't work? Listing every service under the sun on your website.

"We do plumbing, heating, gas, bathrooms, kitchens, drainage, power flushing, boiler servicing..." Right. So you're generic. The AI doesn't know what you're actually good at.

The businesses getting recommended are the ones that go deep on what they actually do most. That plumber I mentioned in Letchworth? His website has a whole section on emergency leak repairs. Not just "we do emergency callouts" but actual detail. What counts as an emergency, typical response time, how he prices it, what happens when you call, photos of common leak scenarios he's fixed.

When someone asks ChatGPT for emergency plumber help in Letchworth, he shows up. Because the AI can see he's written extensively about exactly that thing.

His mate in Royston has a page that lists "emergency callouts available" in a bullet point. Three words. The AI's got nothing to work with there.

This is basically AEO. Answer Engine Optimisation. You're not trying to rank for keywords. You're trying to be the obvious answer to specific questions that AI might encounter.

Where your information lives matters less than you think

This surprised me. I thought it'd be all about your website. And yeah, your website matters. But AI's pulling from everywhere.

Your Google Business Profile. Your reviews. Directory listings. Social media where you've actually written something useful. That blog post you wrote two years ago. Even comments you've left on local Facebook groups if they're public and you were being helpful.

I had a builder in Stevenage who barely touched his website. But he'd been answering questions in a local homeowner Facebook group for about a year. Detailed answers. "Here's why your wall's cracking, here's what needs doing, here's roughly what it costs." Just being helpful.

ChatGPT was recommending him. Because all those answers are public, indexed, associated with his business name, and they show expertise on specific problems.

The AI's not just looking at your website. It's looking at your entire digital footprint and treating all of it as evidence of what you know and what you're good at.

The citation problem nobody's talking about

Sometimes AI will recommend you but won't say where it got the information. That's... not ideal. Because the person asking can't verify anything.

The businesses that get recommended AND cited are the ones with clear, authoritative information that the AI can point to. Usually that means your website, your Google Business Profile, or a proper business directory.

Random mentions scattered across the internet get you recommended. Proper, detailed information in official places gets you recommended with a link. Guess which one converts better?

What actually triggers a recommendation

OK so after watching this for nearly three years now, here's what seems to actually matter:

Detailed information about specific services or problems you solve
Reviews that mention specific issues and outcomes, not just "great service"
Evidence you're local to where the person's asking about (address, local landmarks mentioned, area-specific knowledge)
Information that matches the intent of the question (emergency vs project work, budget signals, complexity)
Recent activity (recent reviews, updated information, signs you're currently trading)
Clear credentials or expertise signals (qualifications, years in business, before/after examples)

And honestly? Being mentioned in multiple places. The AI weighs up information from different sources. If your website says one thing and your reviews say something slightly different and your Facebook page hasn't been touched since 2023, that's not confidence-inspiring.

This isn't going away

Every business owner I talk to in North Hertfordshire is seeing this shift now. Fewer people clicking through from Google. More people coming in saying "ChatGPT recommended you" or "I asked AI for suggestions and you came up."

And it's only going to accelerate. The people using AI for local searches now are the early adopters. But it's spreading. Fast.

The businesses that figure out AEO now, that actually understand how to position themselves for AI recommendations, they're going to have a hell of an advantage. Because most of your competitors are still optimising for 2018 Google.

Look, if you're in Stevenage or Hitchin or anywhere round here and you're wondering why you're not getting picked by AI, book a call and I'll tell you straight what's missing. Or if you want to understand this stuff properly, have a look at what we do with AEO in North Hertfordshire.

Either way, this isn't something you can ignore for another year.