Why Thin Coverage Keeps You Out of AI Answers

The business on the screen vs. the business in real life

Ask a business owner what they do and you'll get a long answer. Ask their website the same question and you'll get three services and a contact form.

It happens for a sensible reason. When the site was built, someone asked which services make the most money, and those got pages. Everything else became a line item on a services list, or didn't make it at all. That was fine when the job of a website was to convert a visitor who already knew who you were.

It isn't fine now, because an AI engine can only see the version of the business that's published. If seven of your ten services exist only as bullet points, an engine has no reason to name you when someone asks about any of those seven — and no way to know you're a substantial operation rather than a three-service shop.

This is Signal 3 of the 8 Signals, worth 15 of 100 points. It's the gap between what you actually do and what a machine can tell you do.

Depth isn't word count

The instinct when someone says "you need more content" is to make existing pages longer. That's usually the wrong move, and it can actively hurt you — padded pages make it harder for an engine to lift a clean answer, which costs you on extractability.

Depth means coverage, not volume. A site with twelve tight pages covering twelve real topics will outperform a site with four bloated ones every time, because it gives an engine twelve distinct things it can confidently say you handle.

The three layers of coverage

When I score this signal, I'm looking at three things.

Every service represented. Not listed — represented. Each thing you actually sell should have a page that says what it is, who it's for, what it costs or what drives the cost, and what happens when someone buys it. If a service earns you real revenue and has no page, that's the first gap to close.

The adjacent questions answered. Around every service sits a ring of questions that aren't sales questions: how long does this take, what do you need from me, what happens if it goes wrong, how do I know if I need this at all. Those are the questions people bring to an AI assistant, and they're covered in more depth under question and answer content. For this signal, what matters is that they exist on your site somewhere.

Enough subject coverage to read as a source. This is the layer most businesses skip entirely. It's the difference between a site that describes a company and a site that teaches a topic. An engine looking for an authority on roof warranties, or peer advisory groups, or commercial HVAC retrofits, is looking for a site that clearly knows the subject — not just one that sells within it.

Hub-and-spoke: how depth gets built

Depth built page-by-page in no particular order produces a pile. Depth built as a structure produces an asset.

The pattern that's worked best for us is hub-and-spoke. One page defines the whole framework and links out to each piece. Each spoke goes deep on one piece, carries its own FAQ and its own schema, and links back to the hub and across to its siblings. An engine encountering any one page can see the shape of the whole thing.

You're reading one right now. This post is a spoke; the 8 Signals is the hub. We built the same structure for a roofing client around five promises they make to every customer — a hub page defining all five, then one deep page per promise, interlinked. That's a live example you can go look at.

The reason the structure matters more than the individual pages: it tells an engine these topics belong together and belong to you. A dozen unconnected blog posts about roofing don't do that. A defined framework with a page per element does.

Where the material comes from

The objection I hear at this point is always some version of: I don't have time to write twelve pages, and I don't have twelve things to say.

You almost certainly have more than twelve. The problem is that it's stored as conversation rather than as content, which is the same problem behind the proof signal.

With one client, we recorded a single one-hour interview — me asking questions, him answering the way he would on a sales call. From that hour I've pulled fourteen distinct blog posts covering the topics he needed to strengthen, plus the complete copy for a financing page. One hour. Fourteen posts and a page.

That ratio isn't unusual. It's what happens when someone who's been doing the work for twenty years is asked specific questions instead of being handed a blank document.

This is also, more or less, what Marcus Sheridan has been arguing for years in They Ask, You Answer — that the questions your buyers actually ask are your content strategy, and that the uncomfortable ones matter most. That framework predates AI search by a decade and it turns out to be almost exactly what AI search rewards. We'd been working through those principles with Advantage Roofing before any of this AI work started, and it's a large part of why they had so much to publish once we went looking.

The pages for things you haven't sold yet

One category of gap is easy to miss: the thing you're about to start offering.

Same roofing client — they were beginning to offer financing, and had nothing on the site about it. Not a thin page, nothing. Which meant that every homeowner who asked an AI assistant whether roofers in that market offer payment plans got told about somebody else.

If you're adding a service, the page should exist before the phone starts ringing, not after. It's the cheapest coverage you'll ever build, because the material is fresh in your head from having just decided to offer it.

How much is enough?

There's no number, and I'd be inventing one if I gave you a target word count or page count.

The practical test I use: make a list of every question a serious buyer might ask you between first hearing your name and signing. Then check whether there's a page on your site that answers each one directly. The gaps on that list are your content plan, in priority order, and it's usually a lot shorter than people fear.

For most service businesses that lands somewhere between ten and twenty substantial pages. Not fifty. Not four.

What I don't know yet

I don't know whether an engine rewards a site that covers one topic exhaustively more than a site that covers ten topics adequately. I'd guess focus wins, but I haven't seen enough clean cases to say.

I don't know how the calculus changes as more businesses publish more, or whether the bar for "enough" rises as everyone's coverage gets denser. I'd expect it to.

What's held consistently: businesses whose sites represent everything they actually do get named for more things than businesses whose sites represent a slice. I'll update this as the rest gets clearer.

Frequently asked questions

Related reading

• The 8 Signals — the full framework, all eight scored categories, and what your total means.
• The Proof AI Is Looking For — Signal 1, and where the raw material for all this coverage comes from.
• How to Write Content AI Can Quote — Signal 4. Coverage gets you on the page; extractability gets you quoted.
• The Questions Your Buyers Are Asking AI — Signal 5, and the adjacent questions worth building pages around.
• Advantage Roofing case study — the client behind the fourteen-posts-from-one-hour example.

This is Signal 3 of 8. See the full framework →