Organic clicks are down. Rankings look stable. Somebody in the meeting says “AI Overviews,” and everyone nods.
That explanation is mostly right, but it is also lazy. Clicks can fall for half a dozen reasons, and AI Overviews are only one of them.
If you blame the wrong cause, you spend the next quarter rewriting pages that were never the problem, while the real issue keeps costing you traffic.
The good news is that you can test the AI Overview theory with data you already have. It takes an afternoon in the Search Console and a rank tracker. And the pattern you find tells you what to fix first.
Quick Answer: AI Overviews are the likely cause of lost clicks when three things are true at once. Average position on the affected queries held steady, CTR on those queries fell, and an AI Overview now appears for them. If position has fallen too, you have a ranking problem. If impressions have fallen, you have a demand problem. Once AI Overviews are confirmed, the fix depends on whether the overview cites your page, ignores it, or cites it only some of the time.
Why Falling Clicks Don’t Automatically Mean AI Overviews
Before you look for AI Overviews, rule out the usual suspects.
Ranking losses are the obvious one. A drop from position three to position six will cut clicks whether an AI Overview appears or not.
Demand changes are the second. If impressions fall alongside clicks, fewer people are searching. Seasonal topics, product launches and news cycles all do this.
Other SERP features are the third. A new video carousel, a People Also Ask box or a block of shopping results can push organic listings down the page just as effectively as an AI Overview.
Branded query shifts are the fourth. If your brand traffic falls, the cause usually sits outside search entirely, in PR, product changes or paid campaigns.
AI Overviews produce a specific signature that none of these do. Impressions stay roughly the same. Position stays roughly the same. CTR falls.
That combination is what you are looking for.
Here’s the corrected section. The main fix is the opening claim: Search Console now has a dedicated Generative AI report, so the diagnosis is no longer fully indirect. It’s still partly indirect, because that report shows impressions but no clicks and no queries.
How to Diagnose AI Overview Click Loss in Search Console
Search Console now has a dedicated Generative AI report, found under Performance in the left-hand menu. It shows how often your pages appeared in AI Overviews and AI Mode, broken down by page, country, device and date. However, it reports impressions only. It has no clicks, no CTR and no query dimension, and its data can’t be pulled through the API. AI Overview impressions are also still counted inside your standard Performance report, so the two reports overlap rather than add up.
That means the diagnosis is still partly indirect. You use the Generative AI report to see which pages are being cited, look for the click-loss signature in the standard Performance report, then confirm AI Overview presence at query level with a tool that records it.
Here is the process:
- Pick two comparable periods. In the Performance report, use the Compare tab to set the last 28 days against the same 28 days a year earlier. Year-over-year comparisons filter out most seasonal noise. If the site is less than a year old, compare the 28 days after the drop started with the 28 days before it.
- Check the Generative AI report first. Open it for the same periods and note which pages gained AI impressions and on which dates. Copy or export the page table, since this data isn’t available through the API. A rise in AI impressions on the same dates that clicks fell is early evidence. If your site has too few AI impressions, the report may not appear at all.
- Export queries and pages for both periods. From the standard Performance report, pull clicks, impressions, CTR and position for each query and for each page into a spreadsheet. Filter to rows with enough impressions to be meaningful, so a handful of random long-tail queries does not distort the picture.
- Join pages to AI impressions. Match the page export against the Generative AI page table. For each page, divide AI impressions by total impressions to get an AI exposure share. Pages with a high share and falling CTR are your strongest candidates.
- Isolate the signature. Flag queries where impressions changed by less than about 20% and CTR fell noticeably, and where average position either moved by less than one place or improved. Include the improved ones, because an AI Overview citation can lift reported position (see the caveat below). This is your shortlist. Everything else has a different explanation.
- Split out question-style queries. Use a regex filter in Search Console, such as ^(what|why|how|when|who|is|can|does)\b, to see whether the shortlist leans toward informational questions. AI Overviews appear far more often on questions and longer queries, so a strong lean supports the theory.
- Separate branded and non-branded queries. Exclude your brand name with a second regex filter. Branded queries behave differently and can hide or exaggerate the effect.
- Confirm AI Overview presence. Load the shortlist into a rank tracker that records AI Overviews, such as Ahrefs, Semrush or SE Ranking. Check whether an overview appears for each query and whether your page is cited in it. This step is still essential because the Generative AI report has no query data.
- Spot-check the SERP by hand. Search the top ten queries yourself in an incognito window, on desktop and mobile. Tools miss things, and seeing the overview tells you how much of the screen it takes.
Two caveats are worth knowing before you read the numbers.
First, when your page is cited inside an AI Overview, Search Console assigns that link the position of the overview itself. A citation can therefore make the average position look better even while clicks fall. If a query’s position improved and its CTR dropped, check for a citation before treating the position change as a ranking gain.
Second, the Generative AI report only counts impressions when your page is linked inside the AI answer. If an AI Overview pushes your organic result down without citing you, the report shows nothing, yet you still lose clicks. Uncited losses only show up through the CTR signature and the rank tracker or manual checks, and they are often the larger share of the problem.
What the Pattern Tells You
Once the shortlist is confirmed, sort each query into one of three groups. Each group points to a different fix.
Cited but Not Clicked
Your page appears as a source in the AI Overview, but CTR still fell. This is the most common pattern for definition and how-to queries. The overview answers the question well enough that most searchers stop there, and your citation earns an impression instead of a visit.
This is not a content quality problem. Google already trusts the page. The problem is that the page gives searchers no reason to go beyond the summary.
Ranking but Not Cited
Your page ranks on page one, an AI Overview appears, and your page is not among its sources. This group usually costs the most, because you lose clicks to the overview and get no visibility inside it.
The cause is usually structural. The answer to the query is buried in a long introduction, spread across several sections, or phrased too vaguely to lift out as a passage. When this pattern holds across dozens of commercial queries rather than a few blog posts, teams usually weigh in-house content fixes against answer engine optimization services, and the deciding factor is often how many templates and pages need restructuring rather than how many articles need editing.
Mixed Results
Some queries are cited, some are not, and CTR moves inconsistently. This often means the AI Overview itself keeps changing its sources, which is common. Before acting on this group, track it for a few weeks. Changing pages in response to a moving target tends to produce noise rather than results.
What to Change First
Resist the urge to fix everything at once. Work in order of commercial impact, starting with the group that costs you the most.
Start with Commercial Pages That Are Not Cited
Pages in the “ranking but not cited” group that target commercial or comparison queries deserve attention first. These are the queries where a click is most valuable, and where losing it to a competitor’s citation hurts most.
Rewrite the section that answers the query so that the answer comes first, in two or three plain sentences, before any background. Use the exact entities searchers use, such as product names, categories and integrations. Remove hedging that makes the passage hard to quote.
Then check the page for technical barriers: content loaded only through JavaScript, answers hidden in tabs or accordions, and schema that no longer matches what the page says.
Then Give Cited Pages a Reason to Be Clicked
For pages in the “cited but not clicked” group, the goal is not to rank better. It is to offer something the summary cannot. That might be a calculator, a comparison table, a downloadable template, original data or a worked example. Make that asset visible early on the page and reference it in the title and meta description, so the listing promises more than the overview delivers.
Change What You Report
Next, adjust how success is measured. Traffic from informational queries may not come back to previous levels, and chasing it can waste effort. Report conversions per query group, citation presence on priority queries and branded search trends alongside clicks. A page that loses clicks but keeps its citation and its conversions may be performing better than the traffic chart suggests.
Re-Run the Diagnosis After Each Change
Give each batch of changes four to six weeks, then repeat the same comparison with the same query list. Look for two things: whether previously uncited pages now appear as sources, and whether CTR on cited pages has recovered at all. If neither has moved, the problem may sit elsewhere, such as a competitor with stronger coverage of the topic or an overview that now favors forums and video. Keep the original shortlist as a baseline, so each round of changes is measured against the same starting point rather than a moving one.
The Diagnosis Comes Before the Fix
AI Overviews are changing organic traffic, but they are not responsible for every drop. The signature is specific: stable position, stable impressions, falling CTR, and an overview on the query. Confirm that pattern first, sort the affected queries by whether you are cited, and fix the commercial pages that are ranking but ignored before anything else. That order turns a vague sense that AI is taking your traffic into a short, prioritized list of pages worth changing.
Also See:
