Track your visibility in Google AI Overviews

本文目前只有英文版。站点其余部分已汉化，这篇内容尚未翻译。

This is the AI answer surface with the largest audience by a wide margin, and it is also the one where naive measurement goes wrong most often. Both facts have the same cause: the AI Overview does not appear for every search.

How the probe works

For each tracked keyword, TriRank reads a live Google results page and pulls the AI Overview block off it, along with the organic results from the same request.

Two choices in that sentence were deliberate.

We read the results page, not the AI Mode tab. Google ships two different AI surfaces, and the vendor documentation in this space uses the two names loosely enough that it is genuinely easy to query one endpoint and report the other product's name. AI Overviews is the block above the blue links. AI Mode is the separate conversational tab. This page is about the first one only. The second has its own column and its own series, and the two are never folded together — which is the whole reason this paragraph exists, because folding them would rewrite the meaning of every historical point in this one.

The AI Overview and the organic ranking arrive together. That is what makes the citation-gap report possible — the one that surfaces keywords where you rank in the top ten and the AI Overview above those results still never mentions you. Those two facts have to come from the same page view to mean anything.

The third state, and why it matters more here

An AI Overview does not fire for every keyword. Google decides, per query, whether to summarise at all.

So a check on this surface has three possible outcomes, not two:

• Cited — an AI Overview appeared and one of its sources is your domain.
• Not cited — an AI Overview appeared and none of its sources is you.
• Not observed — no AI Overview appeared. Nothing was measured.

The third one is where most reporting on this surface quietly breaks. Collapsing "not observed" into "not cited" produces a false negative on every keyword Google chose not to summarise, and since that is a large share of keywords for most sites, it drags the resulting percentage down by an amount that has nothing to do with your content. It would make our headline number look worse, not better, which is precisely why it is worth stating: an error that flatters nobody is still an error.

Rows in that third state are dropped from both the numerator and the denominator of your citation rate. They are not failures and they are not successes; they are questions that were never asked.

What "cited" means here

A keyword counts as cited when one of the sources inside the AI Overview block resolves to your domain — host lowercased, leading www. dropped, matched only when it is your domain exactly or a subdomain of it.

As on every other engine, the summary text is never read for brand mentions. An AI Overview naming your company without linking a page of yours is not your citation.

What this cannot tell you

It is one location, one language, one moment. AI Overviews vary by locale and by the individual result page. A reading is a sample from a specific configuration, not a global truth about the keyword.

Volatility on this surface is high. Google has changed how often AI Overviews trigger, and for which query types, repeatedly. A shift in your not-observed count often reflects Google changing its mind about summarising a category rather than anything you did.

Being cited here is not the same as being clicked. An AI Overview answers above the results, which is the zero-click pattern in its most consequential form. Citation is visibility; the traffic question needs your own Search Console numbers, which is why TriRank ties both to the same page.

There is no position number here, and there is not one to have. A results page has ten slots, so a rank is a row you either occupy or you do not. A synthesized answer is prose that names a few brands and links a few pages, and there is rarely a numbered position to track inside it. What this check reports is therefore presence and citation, not a row number, and nothing on this page converts into one. Where you would want a rank, the measurement doing that job is the trend — how often this engine cited you across weeks of the same questions, rather than where you placed inside any single answer. The word "rank" is still settling in this market: we use it in the sense set out under AI search ranking, and AI search rank tracking covers what tracking it involves.

There is no published cross-brand benchmark for this engine. What runs on your account is your own sweep: your tracked queries, through this engine, on the same weekly cadence as every other engine in the sweep. What does not exist is a public study across many brands that would let you ask how this engine behaves in general — how often it cites anyone, or which kinds of sites it tends to reach for. No number on this page comes from one, and we are not putting a date on when there will be one. The distinction is worth stating plainly, because "we do not measure this engine" and "we have not published a study about this engine" are different claims and only the second is true.

What to do with the result

Start with the gap, not the rate. The most actionable output on this surface is the list of keywords where you already rank in the top ten and the AI Overview still cites somebody else. You have proven you can rank for those words; the retrieval layer is choosing other pages anyway, and the sources it chose are listed for you to read.

The usual culprit is structural rather than topical. AI Overviews pull from passages that answer the question directly and compactly, which is closer to featured snippet shape than to long-form article shape — a similarity that goes only so far, as AI Overviews vs featured snippets sets out.

The free audit runs this check against your domain and shows what came back, including which keywords produced no AI Overview at all.