No account, no card. 30 days of AI Overviews tracking across six engines.
Four steps sit between a question and a citation. Knowing which one you fail at tells you what to fix.
One typed question becomes several related searches. You are competing for a set of sub-questions, not a single keyword.
Pages are pulled for each sub-question. If you are not retrievable for any of them, nothing downstream can save you.
The model summarises what it retrieved. Clear, self-contained claims survive the summary. Hedged paragraphs do not.
A handful of links are attached. This is the only step a rank tracker for AI Overviews can observe directly, so it is the one we report.
Three views on the same logged answers: how visibility moves, which queries carry it, and who sits above you.
Watch mentions move day by day instead of guessing from a single snapshot.
See the exact prompts that produced an overview naming you, ranked by how often they did.
Line up the domains cited beside you and see which ones AI reaches for first.
Six things we see in cited-source lists again and again. None of them is a trick.
Put the answer in the first two sentences, under a heading that matches how people ask it.
Engines reach for comparison pages on commercial questions. If you have not written one, a competitor's gets read.
Community threads appear constantly in cited sources. A genuinely useful reply can reach the overview.
Five thin pages on one topic split retrieval. One maintained page concentrates it.
Consistent naming, a clear about page and structured data help an engine match you to the right thing.
Your top-cited pages are your highest-leverage pages. Update those before writing anything new.
What the tracker measures, where the data comes from, and what it cannot tell you.
An AI Overviews tracker records how often Google AI Overviews and other answer engines name your domain, which queries trigger those answers, and which sources get cited alongside you. The unit of visibility is a sentence, not a position.