How-to guide · content
How to write service pages that AI engines can recommend
When an assistant recommends a business, it is quoting a service page. Most service pages give it nothing to quote: no price, no scope, no location, no proof. Here is the nine-part structure that turns a service page into the source an engine uses.
Ask an assistant "who does commercial EPCs in East Sussex and what does it cost" and watch what it does: it finds a page that states the service, the area and a price, and quotes it. If your page says "we offer a comprehensive range of energy assessment services, contact us for a quote", it has nothing to work with and moves on to a competitor or a directory. In our audits, weak service pages are the most common content reason a business with a decent site is missing from AI recommendations.
The structure below is the same one a good salesperson uses on the phone. Put it on the page.
Step 1: One service per page, named plainly
The H1 names the service and, for local businesses, the area: "Commercial EPC assessments in East Sussex". The URL matches. A single overview page links to each service page, but each service has its own URL so an engine can match a question to a page.
Step 2: Open with the one-paragraph answer
The first paragraph answers what the service is, who it is for, where you do it and what it costs: "We carry out commercial EPC assessments for offices, shops and industrial units across East Sussex, usually within five working days of booking. Prices start at £X for units under 500 square metres." Everything a first-time visitor or an assistant needs is in the first 60 words.
Step 3: Say who it is for, and who it is not
A short block: the situations, business sizes and sectors you suit, and the ones you do not (too small, too far, outside your accreditation). Engines match on these attributes when a customer asks for something specific, and stating the "not" builds trust with both the engine and the reader.
Step 4: Publish the price or the range
A price table by size, tier or scenario, or a "from" price with the three things that move it. Include what is included and what costs extra. If you cannot publish exact numbers, publish typical ranges from real recent jobs. The engine will quote a number from somewhere; make it yours.
Step 5: Describe the process and the timeline
Numbered steps from enquiry to completion, with realistic durations: "1. Call or book online (same day). 2. Site visit (within five working days). 3. Report issued (within 48 hours of the visit)." This answers "how does it work" and "how long does it take" in one liftable block.
Step 6: Add proof that can be checked
Named clients or anonymised but specific case studies with numbers ("a 1,200 square metre warehouse in Newhaven, assessed and lodged in four days"), accreditation names and numbers, review count and average, and two or three quoted reviews. Photos of your own work, not stock. Proof is what separates a recommendation from a mention.
Step 7: Answer the phone questions in a short FAQ
Three to six questions you get on every call, answered in 40 to 90 words each, in plain HTML. "Do you charge for travel?", "Do I need to be on site?", "How long is the certificate valid?" These are the exact shape of the questions customers put to assistants.
Step 8: Add Service schema and connect it
A Service block with name, description, provider (your Organization @id), areaServed and offers with the price or price range, plus BreadcrumbList. Keep the schema consistent with the visible page.
Step 9: Link it, date it, maintain it
Link to the page from the homepage, the services overview, relevant guides and the footer. Show published and updated dates and keep them honest. Review quarterly: prices, timelines, accreditations, examples.
How to know it worked
Ask ChatGPT, Perplexity and AI Mode "how much does [service] cost in [area]" and "who does [service] in [area]". Your page should be the source, and the price quoted should be yours. Enquiries should arrive pre-qualified, already knowing the range.
Common mistakes
- "Contact us for a quote" as the entire pricing section.
- A services page that lists everything you do in one paragraph each.
- Stock photos and no named examples.
- Key content in tabs or widgets that are not in the HTML.
Frequently asked questions
Should I publish prices on service pages?
+
Yes, at least a range or a 'from' price with what moves it. When a customer asks an assistant what a service costs, the engine answers from whoever publishes a number. If that is not you, it is a competitor or a directory, and you are absent from the price conversation entirely.
How long should a service page be?
+
As long as it takes to answer every question a buyer asks before enquiring, typically 800 to 1,500 words plus a table and an FAQ. Length is not the goal; completeness is. A short page that answers what, who, where, how much and why you is better than a long one that does not.
One page per service or one page for all services?
+
One per service, with a short overview page linking to them. Engines match a page to a question; a page that covers six services is a weak match for all six.
Sources
- Google Search Central: Creating helpful, reliable, people-first content · developers.google.com
- Google Search Central: Service structured data (LocalBusiness) · developers.google.com
- Local Dominator: Why ChatGPT won't cite your website · localdominator.co
George McKenna
Co-founder, Emerging Digital Partners
George co-founded Emerging Digital Partners in Eastbourne and built the AI Search & SEO Audit you are on now. He spent around twenty years in the UK IT channel before that, most recently running solution sales teams, and now splits his time between building websites and search tooling for EDP clients and taking AI products to regulated industries through the ETT Group, where he is Chairman and CTO. He writes about answer engine optimisation the way he practises it: test it, measure it, fix what is actually broken.
Read next
How to build FAQ pages that AI engines actually use
Most FAQ pages are a list of questions the marketing team wishes customers asked. AI engines ignore them. FAQs built from the questions people really ask, answered fully in plain HTML, are among the most quoted pages on a small business site. Here is how to build one.
How to write answer-first content that AI engines quote
AI engines lift passages, not pages, and they lift from the top. A 2026 study found 44 percent of ChatGPT citations came from the first third of a page. Here is how to structure a page so the quotable part is where the engine looks first.
How to keep content fresh so AI engines keep citing it
AI engines prefer recent sources, and several 2026 analyses found recently updated pages cited several times more often than stale ones. Changing the date is not an update. Here is a refresh system that keeps your best pages current without rewriting the site every quarter.
Get in touch