Google already answers without a click: here's how to measure what AI Overviews are taking (or giving) you

The headline says it all: for over a year now, Google has been placing AI-generated answers on top of regular search results, and some of the traffic that used to click through now gets satisfied just reading the summary. The problem isn’t that it’s happening — it’s that you have no tool that tells you directly.

Why Search Console won’t tell you

GSC reports impressions and clicks by query and by page. It doesn’t report whether, at the moment of that impression, an AI Overview was sitting on top of your result. Same old limitation: Google gives you the aggregate, not the context of each individual impression.

This means if your CTR for a query has been dropping for months, you’ve got two competing hypotheses — you lost ranking, or you gained ranking but an AI Overview showed up on top and took the click — and GSC alone won’t let you tell which one it is.

The method that actually works (approximate, not perfect)

1. Segment by query type, not by page. AI Overviews show up far more on informational queries (“what is X”, “how to do Y”) than on transactional or branded ones. If your blog pages are losing CTR but your service pages aren’t, that’s a real clue.
2. Compare CTR for the same query across two time windows, not average position. Average position can stay flat while CTR drops — that’s almost always the signal of an AI Overview or some other new element on top of you, not that you dropped in rank.
3. Manually review the 10-15 highest-impression queries that lost CTR. Search them yourself (or with a rank tracker that captures the full SERP, not just position) and check whether an AI Overview shows up. It’s manual work, but it’s the only real data that exists today.
4. Apply layer 1 of your own GEO/AEO audit: if you’re already checking whether ChatGPT or Perplexity cite you, run the same test against Google’s AI Overviews for your target queries — if they cite you, that’s a sign the “lost” click traffic is being offset in authority/brand terms, even if you can’t put an exact number on it yet.

What to check today

1. Search Console → Performance → filter by query type (informational vs transactional).
2. Sort by CTR drop over the last 3 months, not by position drop.
3. Manually search the 10 highest-impression queries that lost CTR.
4. Note which ones show an AI Overview today (even if you don’t know if it was there 3 months ago).
5. Repeat the citation check (GEO/AEO layer 1) against Google, not just ChatGPT.

This isn’t a closed formula — it’s the method I use while Google doesn’t give more visibility into this in Search Console. If you want to build the full measurement layer (not just this piece) on your own site, here’s how I do it.