The organic click problem associated with Google AI Overviews is no longer primarily a U.S. search story. New data published by Ahrefs on September 16, 2026 finds that French websites with heavy exposure to AI Overviews experienced a median 23.1% decline in click-through rate shortly after the feature arrived in France, providing a relatively clean before-and-after view of what happens when generative answers enter a new search market.
The Ahrefs study, led by France Marketing Manager Juliette Begue with analysis by data scientist Xibeijia Guan, tracked 963 domains using aggregated Google Search Console data. Researchers compared performance during the 28 days before Google introduced AI Overviews in France on July 22, 2026 with the nine days immediately afterward. The short observation window is an important limitation, but it also gives the analysis an advantage: France had a clear rollout date, making it easier to examine the immediate change without relying entirely on long-term correlations.
The 23.1% decline applies to the most exposed sites
The headline number needs to be interpreted precisely. Ahrefs did not find that every French site lost 23.1% of its CTR. The figure describes the median raw CTR change among domains classified as highly exposed — sites for which more than 20% of tracked French queries triggered an AI Overview.
For that group, median CTR fell from 2.29% before the launch to 1.76% afterward, a 0.53 percentage-point decline that translates to 23.1% in relative terms. Ahrefs also found that 82.2% of highly exposed domains experienced a CTR decline. By contrast, its control group — domains for which AI Overviews appeared on less than 2% of queries — saw CTR increase by 2.3% during the same period.
Once that control-group movement is taken into account, Ahrefs estimates the AI Overview effect for highly exposed sites at roughly -24.8%. The company is careful to distinguish the observed 23.1% decline from that adjusted estimate, an important methodological detail when interpreting the study.
The wider market decline was smaller, but still meaningful
Across the full panel, the effect was less dramatic than the headline figure. Ahrefs reports a median CTR decline of 5.7% and a median click decline of 8.8%. The difference between the full sample and the highly exposed group reinforces the central finding: exposure appears to matter considerably more than simply operating in a market where AI Overviews are available.
The distribution of click losses is also notable. According to the study, 48% of domains lost more than 10% of their daily clicks, 19% lost more than 30%, and 5% lost more than half. France-specific .fr domains experienced a median click decline of 12.3%, compared with 8.8% across the broader panel.
Those figures do not prove that every individual decline was caused exclusively by AI Overviews. Search traffic can move for many reasons, and Ahrefs acknowledges that some lightly exposed sites performed poorly while some highly exposed sites held up well. The study instead shows a strong relationship at the group level between AI Overview exposure and declining CTR during the immediate post-launch period.
Ahrefs found an exposure-response relationship
One of the most useful findings is not the 23.1% headline but the relationship Ahrefs observed across different exposure levels. When the researchers compared each domain's change in CTR with its AI Overview rate, they calculated a slope of approximately -0.9. In practical terms, the dataset suggested that each additional percentage point of AI Overview exposure was associated with roughly one percentage point of CTR loss.
Ahrefs explicitly warns that this should not be treated as a prediction formula. A site with AI Overviews on 15% of its queries is not guaranteed to lose exactly 15% of its CTR. Query mix, ranking positions, user intent, brand strength and the type of answer Google generates can all change the outcome. The relationship is better understood as a risk indicator: as exposure rises, the probability and magnitude of click loss appear to rise with it.
That makes AI Overview exposure an increasingly useful metric for SEO forecasting. Instead of waiting for traffic to fall and then searching for an explanation, teams can measure how often their ranking queries trigger an AI Overview and identify which sections of a site face the greatest structural risk.
Rankings can remain stable while clicks disappear
The French data reinforces one of the most important changes in modern search measurement. Ahrefs found little evidence of widespread ranking losses during the period. Pages continued appearing in the results, and impressions often remained stable or increased. The deterioration happened primarily between impression and click.
That creates what the SEO industry increasingly describes as a decoupling between search visibility and search traffic. A page can preserve its ranking position and continue accumulating impressions while generating fewer visits because the AI-generated answer satisfies more of the user's informational need directly on the results page.
Ahrefs highlights an extreme example from its French panel in which impressions increased 183% while clicks declined 1.9%, causing CTR to fall sharply. The example should not be generalized to every site, but it illustrates why impression growth can now coexist with deteriorating organic performance.
For SEO reporting, this makes CTR more important. A dashboard centered on rankings and impressions can make performance appear stable even when the actual flow of users from Google is weakening. Search teams increasingly need to monitor rankings, impressions, CTR and clicks together, with AI Overview exposure layered on top.
Health was the hardest-hit sector in the study
Health stood out in Ahrefs' sector analysis. The category had an AI Overview exposure rate of 22.2% and experienced a 20.4% CTR decline. The likely explanation is not simply that Google treats health differently, but that many health searches are informational and can be answered compactly.
This points toward search intent as a more useful risk lens than industry classification. Queries asking for definitions, symptoms, basic explanations or straightforward factual information are naturally suited to an AI-generated summary. Searches that require a tool, live price, account, transaction, nuanced opinion or human experience may retain more incentive for the user to click beyond the search results.
For publishers, that distinction matters when evaluating content portfolios. Two sites in the same industry can have very different exposure depending on whether their organic traffic comes primarily from concise informational queries or from searches that require deeper interaction.
The French findings echo Ahrefs' earlier global research
The new France study is particularly significant because it aligns with earlier findings from other markets. In July, Ahrefs reported that the presence of an AI Overview was correlated with a 58% lower average CTR for the page ranking in position one in its global analysis. That research used 300,000 keywords, split between queries with AI Overviews and informational queries without them.
The methodologies are different, so the percentages should not be compared as though they measure the same thing. The global research examined CTR differences between keyword groups, while the French study follows domains across a specific rollout date. What matters is the direction: both analyses associate AI Overviews with lower organic click-through rates.
That convergence weakens the idea that AI-related click erosion is merely a quirk of mature U.S. deployment. France provides a newer market where researchers could observe the feature arriving and compare behavior around the launch itself.
Nine days of post-launch data is useful, but not the final word
The strongest reason for caution is the study's short post-launch window. Nine days can capture an immediate effect, but it cannot establish the long-term equilibrium. Google's AI Overview coverage may change, users may adapt their behavior, publishers may alter content strategies and seasonal factors can evolve over longer periods.
Ahrefs acknowledges this limitation and plans to revisit the French panel. The control-group design helps address broad late-July movement, but it cannot eliminate every confounding factor at the level of individual domains. The findings should therefore be treated as early observational evidence rather than a universal causal rule for all French search traffic.
At the same time, the short window is precisely what makes the result difficult to dismiss. The largest CTR declines concentrated among the sites with the highest AI Overview exposure, while minimally exposed sites did not show the same deterioration. That pattern appeared almost immediately after launch.
SEO's traffic model is becoming less predictable
The strategic consequence goes beyond France. For two decades, SEO could broadly assume that greater visibility created more opportunities for clicks. AI-generated search experiences weaken that relationship by inserting an answer layer between the ranking and the visit. Publishers may continue supplying the information that makes Google useful while receiving fewer sessions from the same visibility.
This does not make rankings irrelevant. A source generally needs visibility, authority and retrievability before it can influence AI-enhanced search experiences. But rankings alone are becoming a weaker proxy for business value. The more useful questions are which queries generate AI answers, how often a brand is cited or mentioned, how CTR changes when those features appear, and whether the remaining traffic converts differently.
The French rollout gives SEO teams another reason to establish those baselines now. AI Overview exposure can be measured before a site's click loss becomes obvious, allowing marketers to segment informational content by risk and prioritize pages where users still have a strong reason to continue beyond Google's answer.
The zero-click transition is international
Ahrefs' new data does not show that organic search has stopped sending traffic, nor does it establish that every AI Overview removes the same share of clicks. It shows something narrower and more actionable: in a panel of 963 domains around the French rollout, sites heavily exposed to AI Overviews experienced substantially larger CTR declines than sites with minimal exposure.
That is enough to shift the framing. The organic click erosion associated with generative search is no longer something European publishers can view primarily through U.S. case studies. France now offers its own early evidence, and the effect appeared quickly. For SEO teams, the emerging task is not simply to preserve rankings. It is to understand how much of the demand behind those rankings still produces a click once an AI-generated answer occupies the search journey between visibility and the website.