French publishers may be getting their clearest look yet at what Google AI Overviews do to organic clicks. SEO consultant Glenn Gabe analysed Google Search Console data filtered to France, covering the period after Google launched AI Overviews and AI Mode there on July 22, 2026. The picture is uneven: some sites lost nearly half or more of their daily clicks, while others barely moved at all.
Gabe examined Search Console data for five example sites, looking at clicks, click-through rate and impressions after the July 22, 2026 launch of AI Overviews and AI Mode in France. The launch date matters because it gives the analysis a clear before-and-after line, rather than a vague period when AI features may or may not have been visible to French searchers.
The most striking pattern was not a single uniform drop. Instead, the five sites split into two groups: sites that lost a large share of their clicks, and sites that kept clicks and click-through rate relatively stable even as generative-AI impressions rose sharply. That split is itself an important finding, because it suggests the effect of AI Overviews is not the same for every site, topic or query type.
Gabe also cautions that timing alone does not prove AI Overviews caused the declines. The launch coincided with the observed changes, but correlation is not causation, and other factors could have contributed. That caution should frame how publishers read the numbers that follow.
One site in the sample saw daily clicks fall from about 750 to about 400, a drop of nearly 47%. Its click-through rate fell from 2% to 1.1% over the same period, while impressions stayed stable. Stable impressions with falling clicks is a telling combination: it suggests the site was still appearing in search results about as often, but searchers were clicking through far less often.
A second site, covering YMYL topics, saw an even sharper fall. Its daily clicks dropped from about 80 to about 25, a decline of nearly 69%. YMYL stands for Your Money or Your Life, the category Google uses for topics that can affect health, finances or safety, and it is exactly the kind of topic where an AI-generated answer might satisfy a searcher without a click.
These are large movements by any standard. A loss of nearly half of daily clicks changes what a site can earn from advertising, affiliate revenue or lead generation, and a loss of nearly 69% can change whether a content programme is viable at all. But the figures come from a small sample of five sites, so they should be read as examples of what can happen, not as an average for all French sites.
Two of the five example sites kept clicks and click-through rate relatively stable despite sharp rises in generative-AI impressions. In other words, Google was showing AI features more often for their queries, but users kept clicking through to those sites at roughly the same rate as before.
The analysis as reported does not explain exactly why those two sites were more resilient, and it would be wrong to invent a reason. What the contrast does show is that exposure to AI Overviews does not automatically mean lost clicks. Something about those sites, their topics, their query mix or their position in the results appears to have protected them, at least so far.
For publishers, that difference is arguably the most useful part of the analysis. If some sites hold steady while others fall sharply, the question becomes what separates the two groups. That is a question for further analysis across more sites and more topics, rather than a conclusion that can be drawn from five examples alone.
For SEOs, the French data is a reminder to segment Search Console analysis rather than looking only at site-wide totals. A site-wide average can hide the fact that some topics are losing heavily while others are stable, and that distinction should shape where content and optimisation effort goes next. Filtering by country, as Gabe did for France, is one way to isolate the effect of a dated launch.
For link builders, the implication is more indirect but still real. If AI Overviews reduce clicks for some topics, then the value of a ranking may increasingly depend on whether the page is also cited or surfaced inside the AI answer. Links still matter for authority and discovery, but a page that ranks without being cited may earn less traffic than it used to, at least for the query types most affected here.
The honest takeaway is narrow and practical: watch your own Search Console data around AI launch dates, compare clicks against impressions rather than clicks alone, and do not assume a single industry average applies to your site. The French sample shows losses of nearly 47% and nearly 69% are possible, and it also shows stability is possible. Your own data is the only way to know which group you are in.
Sources: Search Engine Land, November 9, 2026, reporting Glenn Gabe's analysis of Google Search Console data for France.
Related reading on this site: Ziff Davis Says Organic Google Referrals Fell 37% as AI Citations Hit 2.7 Billion, Google's AI Features Sent One Site an Estimated 21% of Its Clicks, Reddit Stock Drops Over 33% as CEO Blames Google's AI Search for Falling Referrals, 3 in 4 Homepages Lost Organic Clicks, Analysis of 110 Sites Finds, Google Organic Click Share Hits a 15-Month High as Zero-Click Searches Fall, and Türkiye Opens Full Antitrust Probe Into Google AI Overviews and AI Mode.
Linkbaits helps you create content that earns citations and backlinks. Start with the toolkit and build your first linkable asset this week.
Start free trial