For about twenty-five years, building a website meant designing for one audience: a person.
You made it look good. You made it easy to use. You put the phone number where a thumb could reach it. If a human being landed on your site and understood what you did and how to hire you, you’d done your job.
That’s over. Not in a dramatic, everything-you-know-is-wrong way. In a quieter, stranger way: you now have three audiences, and only one of them has eyes.
There’s the homeowner. There’s Google’s index. And there’s a rapidly growing third one — the AI assistant that the homeowner asks instead of opening a browser at all.
The genuinely weird part is that these three want the same thing, but they can only find it in completely different ways. And most websites in home services are currently built for exactly one of them.
The three readers, and what each one actually does
The human skims. She lands on your page at 8pm on a Tuesday with a leaking roof and a rough budget in her head. She’s not reading your About page. She’s scanning for one thing: am I in the right ballpark here, or am I wasting my evening? She gives you somewhere between eight and thirty seconds to answer.
The search engine parses. It doesn’t skim, it doesn’t feel, and it doesn’t care how nice your hero image is. It’s looking for structure — headings, entities, markup, questions with answers attached. It’s asking a very boring question: does this page actually answer what people are typing into me?
The assistant retrieves and summarises. This is the new one. Someone types “how much does it cost to replace a roof in Denver, and who’s good?” into ChatGPT, Gemini, Perplexity or the AI panel sitting on top of search. The model doesn’t browse fifteen tabs. It pulls a handful of sources it can read cleanly, extracts the specifics, and produces one answer with three names in it.
You are either in that answer or you are not. There’s no page two.
Why this feels so strange
Here’s what makes this moment genuinely odd for anyone running a service business.
For the entire history of the web, the advice was write for humans, not for search engines. Every SEO worth listening to said it. Stuffing keywords was a losing game, and rightly so.
That advice is now incomplete — because the machine isn’t just ranking you anymore. It’s speaking on your behalf.
That’s the shift. Google’s index was a librarian: it pointed at you and let the human decide. An AI assistant is a representative: it reads you, forms a view, and tells the customer what you’re like — often without the customer ever seeing your site.
So you’re no longer optimising to be found. You’re optimising to be quoted accurately by something that will never call you to check.
And the thing it most wants to quote, in home services, is the one thing most sites refuse to put in writing: the price.
The vagueness problem
A human can cope with vagueness. If your page says “every job is different, call us for a free consultation,” a motivated homeowner might still pick up the phone. Annoyed, but she might.
A machine cannot cope with vagueness. It has nothing to extract. Ask an AI assistant which local company to use and it will confidently recommend the three businesses whose pricing it could actually read — not because they’re better, but because they were legible.
This is the uncomfortable maths of the moment: ambiguity used to cost you some conversions. Now it costs you existence. You don’t rank badly in an AI answer. You’re simply absent from it, and nobody tells you.
Meanwhile Google has been moving in the same direction from the other end — adding an “online estimates” filter to local results, surfacing pages that answer cost questions directly, and rewarding content that reads like a straight answer rather than a brochure. Two very different systems, converging on the same demand: be specific, in public.
So how do you actually design for all three?
The good news is that this isn’t three websites. It’s one website built with a bit more discipline. Five things do most of the work.
1. Answer the money question above the fold. Not a number — a range, with the honest reason it’s a range. “Most kitchen refits in this area land between £12,000 and £28,000, depending on layout changes and whether you’re moving plumbing.” That single sentence serves all three readers at once: the human relaxes, the search engine gets a direct answer to a high-intent query, and the assistant gets something concrete enough to repeat.
2. Make the range interactive, not static. A price list is a snapshot. A short set of questions that narrows a range to this person’s situation does something a static page can’t: it gives the human a reason to give you their details, and it generates a real answer rather than a hedge. This is the single biggest structural change most home service sites can make — and it’s the part of the job an online estimator exists to do.
3. Structure it so a parser can find it. Real headings, not styled text. Question-shaped subheads with the answer immediately underneath. Pricing marked up in schema so it’s machine-readable rather than trapped inside an image or a PDF. If your prices live in a downloadable brochure, they don’t exist as far as any of this is concerned.
4. Write in complete, self-contained statements. Assistants extract passages, not pages. A sentence that only makes sense if you’ve read the two paragraphs above it will get lifted out of context or skipped entirely. Say the whole thing in the sentence: “Gutter cleaning for a typical two-storey semi in the Midlands runs £120–£180.” That travels. “It depends on the property” does not.
5. Don’t let the two goals fight. This is where people go wrong. They read all of the above and produce a page that’s technically perfect and completely charmless — a wall of ranges with no voice, no reassurance, no reason to trust anyone. The human is still the one who signs. Structure is what gets you into the conversation; warmth is what wins it. You need both, and they’re not actually in tension — a clear price is reassuring. Being the business that just answers the question is a personality.
The part nobody says out loud
There’s a version of this article that ends with “and that’s why you need to embrace AI.” That’s not the point.
The point is that transparency stopped being a marketing philosophy and became an infrastructure requirement. For years, showing your pricing online was a choice — a bold one, a differentiator, something you did because you’d decided it was the honest way to run a business. Plenty of good operators decided the other way and did fine.
That window is closing, and not because of a change in taste. It’s closing because the systems that now stand between you and your customer can only pass along what they can read. Being unreadable used to look like discretion. It increasingly just looks like absence.
The businesses that will do well over the next few years aren’t the ones with the best AI strategy. They’re the ones that were already willing to say what things cost — and who now get to be the only clear answer in a category full of “it depends.”
Design for the human. Structure it for the machine. Say the number either way.
Stop losing leads to competitors who answer the pricing question.
14-day free trial. No credit card required.
Priceguide helps home service businesses give an honest online price estimate before anyone has to call — built so both people and machines can read it. See how it works →
