Google’s AI Overviews redistribute the organic traffic maps: sites not cited by the AI see their clicks drop. In this new battle for visibility, it is no longer the classic
Since Google rolled out its AI Overviews on a large scale in Europe and the United States, something has shifted in the way Internet users consume search results. Clicking on the blue link, a pillar of twenty years of natural referencing, has become an exception rather than a reflex.
An engine that responds to itself
For two decades, Google’s implicit contract was simple: I list, you click, the site wins. AI Overviews has breached this contract. From now on, for hundreds of informational queries, Google summarizes an answer directly at the top of the page, citing two or three sources, and the user no longer has any reason to go further down.
The numbers speak for themselves. Several studies conducted since the massive deployment of AI Overviews in the United States in 2024 show drops in organic click-through rates ranging from 15 to 30% on queries where an AI Overview appears. For high-volume queries with low purchase intent, some editorial sites have seen traffic losses greater than 40%. This is not a trend, it is a structural break.
The most difficult thing for marketing teams to accept is that this break does not depend on any fault on their part. A site could very well be ranked first and receive almost no clicks if an AI Overview takes up the entire screen above it. The work has been done, the algorithm recognized it, and yet the traffic is not coming.
The click doesn’t disappear, it mutates
However, it would be inaccurate to speak of the death of the click. What disappears is the comfort click, that of the Internet user who just wanted to check simple information. What remains, and which strengthens, is the click intent: the user who wants to go further, compare, buy, or read an in-depth point of view.
This natural selection reshuffles the cards in interesting ways. Sites that produced purely informative content, with no real editorial added value, now find themselves in direct competition with Google itself. On the other hand, sites that provide an original angle, embodied expertise, proprietary data or a differentiating format retain their clicks, sometimes even increasing them, because they are cited as sources in the AI Overview itself.
This is where the notion of visibility takes on a new meaning. Being visible in 2026 no longer just means appearing in the top ten search results. This means being the source that the AI chooses to mention when it constructs its response. These two goals overlap, but they are not the same, and confusing them is a mistake that many companies still make.
GEO, the new discipline
The industry’s answer is called GEO, for Generative Engine Optimization. This discipline, which did not exist under this name three years ago, consists of optimizing one’s content no longer just for traditional ranking algorithms, but to be selected and cited by generative engines, whether Google, Perplexity or ChatGPT Search.
The GEO levers partially differ from those of classic SEO. Keyword density matters less than structural clarity: explicit titles, direct answers to questions in the first paragraphs, sourced and verifiable data. The thematic authority of a site, that is to say its ability to treat a subject in depth on several linked contents, becomes a determining signal. Schema tags, consistent internal linking and fresh updates matter more than a catalog of backlinks purchased individually.
For an SME, this represents a significant editorial investment. Producing content that is both optimized for SEO, structured for GEO and truly useful for a human reader is a task that requires time and skills that most small structures do not have in-house.
SMEs in the storm
This is precisely where the divide widens between large brands, which have dedicated teams and budgets, and SMEs, which often have to manage their online visibility with limited resources. The former have already adjusted their editorial strategies. The latter, for many, have not yet taken the measure of the change.
For an SME whose 30 to 50% of traffic came from natural referencing, a 25% drop in organic clicks is not a technical incident. This has a direct impact on leads, contacts and turnover. And unlike traditional algorithm updates, there is no automatic recovery to wait: AI Overviews will not disappear, they will expand to new languages, new markets and new types of queries.
Ignoring this development in the hope of a return to normal means taking the risk of finding yourself invisible in a search engine which, in two years, will respond directly to the majority of queries without sending a single click to a third-party site. The window to adapt is open, but it won’t stay open forever.
A concrete example
It is this concrete problem that I sought to respond to by creating ForgR. The platform automates the production of content optimized for both traditional SEO and GEO, that is to say structured so as to be cited by generative engines. For an SME, this means being able to regularly publish coherent, sourced and thematically linked content, without mobilizing a full-time editorial team. Companies using ForgR aren’t looking to produce more content for the sake of it: they’re looking to produce content that Google, Perplexity or ChatGPT choose to cite when a prospect asks a question that directly concerns them.
In conclusion
The end of easy clicking is not the end of SEO. This is the start of a more demanding competition, where the quality, authority and structure of content matter more than ever. For companies that quickly understand this shift, there is a real window of opportunity: those that will be regularly cited by AI in the next twelve months will build a visibility advantage that is difficult to catch up with their competitors who have remained immobile.
The question is no longer whether your site is well ranked. It became: does the AI consider me a reliable source on my topic? And the answer to this question is being answered now, not in six months.
