How to track AI Overviews

You cannot manage what you cannot see, and AI Overviews are deliberately hard to see. Google does not hand you a report of where you appear in them, so most teams either guess or assume the worst. Neither is tracking.

To track AI Overviews properly, watch three things on a fixed set of buyer queries: whether an Overview appears at all, whether it cites or names you, and what it does to your clicks.

Search Console will not break this out for you. So tracking is part manual observation, part rank tracker, and part a read of click loss in your analytics. This is how to set it up without kidding yourself about what the numbers mean.

What tracking AI Overviews actually means

Tracking an Overview is not one measurement. It is three separate questions, and conflating them is where most attempts go wrong.

Presence is whether an Overview shows up for the query at all. Citation is whether, when it does, it names or links you. Impact is what that does to the clicks you used to get. A query can trigger an Overview that never cites you, or cite you and still cost you traffic. You need all three to know where you stand.

Fix a query set first. Pick the buyer queries you care about, the ones your category and your pages target, and keep the list stable so you are measuring change, not noise.

Track whether an Overview appears

Start with presence, because it is the cheapest to check and it frames everything else.

Run each query in a clean browser session, signed out, with personalisation off, and record whether an Overview appears above the results. Do it the same way each time, because a signed-in session with search history returns a personalised result, not the one your buyers see.

Note that presence is not stable. The same query can show an Overview one day and not the next, because they are generated fresh. That is why you record it repeatedly and read it as a frequency, not a fixed fact.

Track whether it cites you

When an Overview does appear, the question that matters is whether you are in it.

Record three things: whether you are cited or linked, whether you are merely named without a link, and which competitors appear instead. That competitor column is the most useful one, because being absent while a rival is cited on a query you both target is the clearest signal of a gap to close.

Over a few cycles this gives you a share-of-citation picture: on the queries you care about, how often you are the source Google leaned on versus how often someone else was.

Track the click impact

The third question is what the Overview costs or earns you, and this is where your own analytics comes in.

The tell-tale pattern is a page holding its ranking position while its clicks decline. That usually means an Overview is answering the query before the user reaches your link. I have set out how to confirm that pattern, and separate it from a real ranking loss, in why rankings hold while traffic falls.

Being cited in an Overview is not automatically a loss, though. On decision-stage queries a citation can keep you visible and still earn the click. The point of tracking impact is to know which queries are which, rather than treating every Overview as a threat.

Manual tracking versus tools

A small query set is easy to track by hand: a spreadsheet with a row per query and a column per check date covers presence, citation, and competitors.

Once the set grows past what you will realistically check by hand, a rank tracker or a dedicated Overview tool automates the appearance and citation checks across many queries at once. That is a convenience, not a different method. The tools answer the same three questions faster; they do not answer a question you could not.

Whatever you use, keep the analytics side in your own hands, because the click-impact read depends on your traffic data, which no external tool sees.

Set a cadence and a fixed query set

The whole exercise only works if it is repeated. One pass is a snapshot of a non-deterministic system, which is close to worthless on its own.

Re-run the same query set monthly, and more closely around a major content or site change. Keep the queries fixed so the trend is comparable. What you are building is a time series: presence, citation share, and click impact, moving over months, on the queries that matter to your pipeline.

That trend is what turns tracking into decisions. It tells you which queries to invest in, using the mechanics in how to rank in Google AI Overviews, and it belongs inside the broader habit of measuring brand visibility in AI answers across every assistant, not just Google.

The takeaway

Track AI Overviews by watching three things on a fixed query set: whether an Overview appears, whether it cites you, and what it does to your clicks. Search Console will not report it, so combine manual, signed-out checks with your own click-loss data, and re-run the same queries on a schedule so you see the trend rather than a snapshot.

If you want that measurement run for you, with a prioritised plan for the queries worth pursuing, that is what an AI visibility audit provides.

This article is part of the SEO in the AI Era: The Complete Guide guide.