How to Write Service Area Pages That AI Actually Recommends
The service area page problem nobody wants to talk about
Right, so you've got a business that covers multiple towns. Maybe you're a plumber in Stevenage who also works in Letchworth. Or you run an accountancy practice and you've got clients across Hertfordshire. And someone somewhere told you that you need "location pages" for SEO.
So you made them. And they're basically the same page, copy-pasted, with the town name swapped out. "We're the best plumbers in Stevenage" becomes "We're the best plumbers in Baldock" and you called it a day.
Here's what's changed in 2026. AI engines like ChatGPT, Perplexity, Gemini, they're not recommending those pages. At all. They can smell the template from a mile away. And when a potential customer asks "who should I use for X in Y town?" your name doesn't come up, even though you literally have a page for that exact town.
I've seen this play out about forty times this year already. Business has a dozen location pages. None of them show up in AI recommendations. None of them rank particularly well in Google either, but that's almost secondary now. The real problem is that when someone's having a conversation with an AI about finding a local service, you're invisible.
What AI is actually looking for (and it's not what you think)
The old SEO playbook said you needed the town name in the H1, a few times in the body copy, maybe in the meta description. Job done.
That's not how AI search works.
When someone asks Perplexity or ChatGPT for a recommendation, it's not doing keyword matching. It's trying to understand which businesses genuinely serve that area, have some connection to it, and can demonstrate they know what they're talking about. It's looking for signals of localness that go beyond just mentioning the place name.
I had a client, electrician covering five towns around North Hertfordshire. His Hitchin page was getting recommended by AI. The other four weren't. Same template, same structure, all of them. Took me a while to work out why.
The Hitchin page had a paragraph about the age of the housing stock in the town. Specific. He'd mentioned that a lot of properties around Hermitage Road still had old wiring from the 60s. The other pages didn't have anything like that. They were just "we serve X, we're great at Y, call us."
That level of local specificity, that's what AEO is about. The AI engines are pattern matching for genuine local knowledge, not just SEO furniture.
The "I actually work here" test
Look at your service area pages right now. Could you swap the town names around and would anyone notice? If yes, they're not going to get recommended by AI.
Here's what actually works:
Talk about the actual place. Not tourist board stuff. Not "Stevenage is a vibrant town with great transport links." I mean the stuff you know because you work there. The industrial estates where your commercial clients are. The specific housing developments where you've done work. The thing that always comes up when you quote jobs in that area.
When I write service area pages now, about half the content is just observations about that specific location as it relates to the service. For a heating engineer in Letchworth, that might be talking about how the Garden City housing stock means a lot of properties with original fireplaces that people want to convert. For a marketing agency covering Royston, it might be noting that you work with a lot of businesses that serve both Hertfordshire and Cambridgeshire, which creates specific positioning challenges.
It feels weird to write at first. You think "nobody cares about this detail." But AI engines absolutely do. They're looking for signals that you're not just claiming to serve an area, you actually know it.
The structure that AI engines can parse
This matters more than people realise. AI isn't reading your page like a human does. It's extracting structured information.
Here's what I'm seeing work in 2026:
- Clear service scope statement. Not "we serve Baldock and surrounding areas." Actual specificity. "We cover Baldock, including Bygrave, Weston, and Clothall. Usually within 30 minutes."
- Local context for your service. Why does your service matter in this specific place? What makes it different to provide your service here versus somewhere else?
- Evidence you've actually worked there. You don't need to list every job. But something. A case study. A mention of a project. Street names occasionally. Something that proves you're not just claiming coverage.
- The practical stuff. Where do you park when you visit? How long does it take you to get there? Do you have a regular day you're in that area? This sounds mundane. It's exactly what AI is looking for to determine if you genuinely serve somewhere.
Stop writing for Google, start writing for questions
The whole game has shifted. It's not about ranking for "plumber Stevenage" anymore. It's about being the answer when someone asks their AI "who should I use for plumbing in Stevenage?"
That's a different type of content.
You need to anticipate the actual questions. "Do you cover emergency callouts in Hitchin?" Not as an FAQ at the bottom. As actual content in the page. "Yeah, we do emergency work across Hitchin. Usually get there within an hour if it's urgent. We're based in Letchworth so Hitchin's close, we're there most days anyway."
See how that reads differently? It's answering a question someone would actually ask. And when an AI sees that content, it can extract a clear answer to provide.
Most service area pages I see are still written like brochures. "Welcome to our Stevenage service page. We are proud to offer..." Nobody's asking AI "tell me about a company's pride in serving Stevenage." They're asking "can this company help me quickly?" or "do they know what they're doing in this area?"
Write. For. Those. Questions.
The technical bits (that matter less than you think)
Yeah, you still need the basics. Schema markup for local business. NAP consistency. The town name in the title tag. Whatever.
But I've seen perfectly optimised pages that AI never recommends because the content is template garbage. And I've seen slightly scruffy pages that get recommended constantly because the content is specific and useful.
The technical stuff is table stakes. It's not what gets you recommended.
One thing worth doing: make sure each service area page has unique, specific images. Not stock photos. Actual photos from that area if you can. AI engines are getting better at understanding image context, and it's another signal of genuine local presence.
What this means for businesses with 20+ service areas
Bit of a problem, isn't it?
If you're a regional business covering loads of towns, you can't write properly specific pages for all of them. Not unless you've got a content team.
So you have to make a choice. Either focus on your core 3-5 areas and do them properly, or accept that most of your location pages won't get AI recommendations.
There's no hack here. You can't scale genuine local knowledge. That's sort of the point. AI engines are filtering for businesses that actually have presence and expertise in an area, not businesses that claim to cover everywhere.
I've had this conversation with about six clients in the last two months. They want 30 service area pages. I tell them they can probably do 5 well. They push back. Then they see the difference in AI recommendations between the detailed pages and the thin ones, and they get it.
Quality over coverage. Every time.
---
If you're running a business in North Hertfordshire and your service area pages are basically just templates with town names swapped, they're not working in 2026. AI isn't recommending them. And increasingly, that's where your customers are starting their search.
We help local businesses show up in AI search results through proper AEO. Not template pages. Actual local content that AI engines want to recommend. If that sounds like something you need, book a call and we'll look at what you've got now. Or check out what we do with AEO in North Hertfordshire specifically.