Why smaller competitors beat you in AI answers (and why more content doesn’t fix it)

A large site can have hundreds of pages that each touch the subject and none that settle it, so the smaller competitor gets the citation. More content doesn’t change that. Better-aimed content does.

Why does AI mention a smaller competitor instead of us?

Because for the specific question being asked, the smaller competitor’s page is the better answer, and the assistant is choosing an answer, not a company. It retrieves a handful of pages, lifts the passages that address the question, and names whoever those passages describe.

A specialist tends to win that contest for three reasons. Their page is about one thing, opens with the answer and uses the words the customer used. Their whole site says the same thing, so the assistant sees one clear description rather than a dozen. And the rest of the web describes them the same way, as “the company that does X”, which is exactly how the question was phrased.

A larger business is usually the opposite. Its pages are built for browsing: category pages, brochures, capability overviews, news. Each one mentions the subject. None of them answers the question. The assistant, reading quickly, picks the page that does.

Does having more content help?

Not by itself, and it can make things worse. Two things happen when a site adds pages without a clear target. Pages on overlapping subjects compete with each other, so no single one becomes the obvious answer. And pages published to a calendar rather than to a question tend to describe rather than answer, which is the kind of content assistants skip.

The pattern on the citing side is unforgiving. A small number of pages get cited heavily for any given subject and most pages never get cited at all. Adding pages adds to the second group unless each one is aimed at a question and answers it better than what’s already being cited. The GEO study from Princeton and partner universities found that adding specifics to a page, such as statistics, quotations and cited sources, increased its use in AI answers, while adding keywords didn’t help. Volume wasn’t the lever. Precision was.

What are the smaller brands’ pages doing differently?

Five things, usually without a strategy behind them. They answer the question in the first paragraph. They keep to one subject per page. They’re specific, with numbers, named options and real detail. They describe the business the same way everywhere. And the description is backed up elsewhere, in reviews, directories and trade coverage that say the same thing.

None of that needs a big team. It needs someone who knows the subject writing the page a customer would want to read, then keeping it current. That’s easier for a specialist than for a marketing department managing 400 pages, which is why the specialist wins by default. It’s not an advantage a larger business can’t copy. It’s one it usually hasn’t.

How many pages does it actually take?

Fewer than you’d think, and they need to be the right ones. A subject is usually covered by a small set: one page that answers the main question, one that proves you do the work (results, cases, credentials), one that explains who you are and what you specialise in, and a few that answer the follow-up questions a customer asks next. After that, returns flatten fast.

The thing that decides whether those pages get cited isn’t how many there are. It’s whether your business is already described for that subject elsewhere on the web, and whether the pages say what the rest of the web says. A well-written page on a subject nobody else associates you with will struggle. The source ecosystem in AI search explains why the off-site part matters as much as the pages.

What should a larger business do about it?

Pick the subjects, find what already works, consolidate, rewrite the openings, and stop publishing to a calendar. In that order.

Choose the subjects that matter commercially, not every subject you could write about. Then find which of your existing pages already earn citations or rank for those subjects; Search Console and a few questions put to the assistants will tell you. Those pages are your starting point, because an assistant that already cites them will notice when they improve. Merge overlapping pages into one clear one and redirect the rest. Rewrite each page to open with the answer, in the customer’s words, with the specifics a specialist would include. Retire pages that exist only for volume. Then check that your business is described the same way across your site, your directories and your social profiles, and go and get the coverage the specialist already has.

What pages to fix first for AI search shows how to choose the starting set. What focused AI visibility work can do shows how quickly the right pages move. And if the problem isn’t a smaller rival but a competitor of similar size, why AI recommends your competitor and not you covers the wider causes.

Where to go from here

This is what our monthly plans are built for. We track the questions your customers ask across ChatGPT, Claude, Gemini, Perplexity and Google’s AI Overviews every day, show you which pages are being cited for them and whose, and turn the gaps into the content we write each month, aimed at the questions rather than the calendar. Are you free for 20 minutes in the next week or so? We’ll bring your data to the first call.