In the past, the bargain between a search engine and a website was straightforward and mutually understood. Google answered the question by pointing at the pages that answered it, the searcher clicked through, and the business that had done the work of publishing something useful received the visitor. AI Overviews have altered that bargain: for a growing share of searches, the answer now appears at the top of the results page, assembled from those same sources, and the click that used to follow no longer necessarily does.
The reaction in most businesses has been somewhere between anxiety and fatalism, encouraged by a great deal of coverage announcing the end of organic search altogether. That framing is wrong, and acting on it is expensive. The accurate answer is narrower and considerably more useful: AI Overviews have taken real traffic from some businesses, have barely touched others, and which group you belong to is largely predictable from the kind of searches that bring you customers.
Has organic traffic genuinely fallen because of AI Overviews?
For informational searches, yes, and the effect is measurable rather than anecdotal. Analysis by Sistrix, reported by Forbes, found that ranking first on a results page typically earns a click around 27% of the time, but that when an AI Overview appears above it, the same first position earns roughly 11%. The page has not fallen in the rankings; the rankings have simply become worth less on those particular searches.
That is a genuine change and it deserves to be taken seriously. It is also specific. The searches most affected are broad informational ones, the "what does X cost", "how does Y work", "do I need Z" questions that a paragraph can satisfy completely. If a searcher's need is fully met by three sentences, an AI Overview meets it and the visit no longer happens.
Which businesses are actually losing traffic, and which are not?
The businesses being hit hardest are the ones whose traffic came from answering general questions, and the ones barely affected are the ones whose customers search with intent to act.
- Heavily affected: Consultancies, publishers, training providers and any business whose organic traffic was built on broad explanatory content. These are precisely the queries an AI Overview resolves in full.
- Largely unaffected so far: Local and service businesses whose customers arrive through location-based searches. Someone looking for a dentist nearby, an emergency plumber or a supplier in their city is not looking for an explanation, and a summary paragraph does not substitute for a business to contact.
- Mixed: Businesses selling considered, high-value products or services, where early research questions are absorbed by AI Overviews but the later comparison and shortlisting searches still produce clicks, because those decisions are not made from a summary.
This distinction matters more than any headline percentage, because it determines whether you are facing a genuine structural problem or reading someone else's. A manufacturer, clinic or contractor whose enquiries come from location-based and product-specific searches is very likely watching an industry panic that has not reached them, and rebuilding a content strategy in response to it would be solving a problem they do not have.
What actually changes about how content should be written?
Content built to be the answer keeps working; content built to be the first paragraph of an answer does not. If your traffic came from posts explaining a concept in general terms, that traffic is genuinely at risk, because the explanation itself is now delivered on the results page. What survives is content that a summary cannot replace: specific pricing, real methodology, sourced figures, comparisons made with a point of view, and the details of how a decision is actually made in your category.
There is also a shift in what a "result" means. Being cited inside an AI Overview places your business in front of the searcher even when no click follows, which is worth having and is earned in a fairly specific way. We have set out that mechanism in full in how to write one blog post that ranks and gets cited by AI, and the same structural habits, direct answers to real questions, clear headings and genuinely checkable sourcing, serve both purposes at once.