CrawlClick

Why an AI engine cited a competitor instead of you

Citation is decided in four stages, and losing at any one of them looks identical from outside. Here is how to find the stage you lost, and how to fix it.

Two pages answer the same question and only one gets named in the answer. The difference is rarely quality in the way a writer means it. Retrieval mechanics decide most of it.

Why did an AI engine cite my competitor instead of me?

Because citation is decided in four stages, and a page can fail at any one of them. Your page has to be fetched, retrieved for the question, usable as a fragment on its own, and attributable to you. A competitor clearing all four wins even when their writing is worse.

Was my page fetched at all?

Check your logs for the engine’s agent before assuming anything about writing quality. A page no crawler has requested cannot be cited, and the usual causes are access rather than prose: a robots rule, a login wall, a slow response, or a page whose text only appears once JavaScript has run. The crawler reference gives the agent name to search for per engine.

How does chunking affect whether AI quotes my page?

Chunking decides which part of your page an engine ever sees. Retrieval splits documents into passages and fetches the one that best matches the question, so your best passage competes against your competitor’s best passage rather than page against page. A page with a heading every few paragraphs divides into clean, quotable units; a wall of text divides arbitrarily and produces fragments that begin mid-argument. Chunking is the whole reason structure beats length.

How do I make my content easier for AI to quote?

Put one complete claim in each paragraph and name its subject in the first sentence. A quotable passage says what it is about, answers in sentences of roughly ten to thirty words, and needs no earlier paragraph to make sense — the test being whether it still reads correctly with everything above it deleted. Self-contained writing is the single change with the largest effect.

Why do specific claims get quoted more than general ones?

A model composing an answer needs something it can attribute. A named figure, a stated limit, a version number or a dated observation gives it a checkable claim to carry. “A leading solution for modern teams” gives it nothing to lift, so the specific version gets quoted and the vague one gets skipped.

Does ranking first in Google mean I get cited?

No. Ranking and citation are separate outcomes decided by only partly overlapping signals. A page can sit at the top of the results and never be quoted, and a page ranking modestly can be cited constantly, because retrieval matches passage meaning rather than page authority alone. The gap between the two is what generative engine optimization addresses.

What if the engine used my words but named someone else?

Attribution fails separately from retrieval. An engine that cannot confidently identify a source — no clear title, no publisher, no structured data, a headline that does not match the body — may still use the text while crediting a source it can name. Making a page easy to identify is the cheapest citation fix available.

How do I confirm which stage I lost?

Work backwards in the same order: logs first, then whether any single passage answers the question in isolation, then whether the page identifies its own source. Ask the engines the question again after each change, because share of answer is the only feedback loop that reflects what an engine actually did.

Last updated 2026-08-10.

Want this measured for your site?

The crawl report is free and needs no changes to your pages.

Request access