Ahrefs says AI search and query fan-out are reshaping SEO

Ahrefs is drawing a hard line under the old SEO playbook: if you only optimize for the visible keyword, you miss the hidden work AI search is already doing. The shift is query fan-out, where one prompt expands into a cluster of related searches, and that changes how brands earn visibility inside answers.

Why query fan-out is the real shift

Search demand for query fan-out has surged 2,550% year over year, which is a useful clue that marketers are finally putting a name to a behavior AI systems have been using behind the scenes. Ahrefs’ view is that AI search is now a brand-new awareness and growth channel inside SEO, not a side quest, and teams are still figuring out what compounds and what does not.

The mechanics matter. Ahrefs defines query fan-out as one prompt expanding into multiple related sub-queries, and says AI Mode typically makes 5 to 11 searches for a simple shopping query. In one example, ChatGPT Deep Research made 420 searches, which shows how much retrieval work can happen before a user ever sees an answer. Google has described the same basic behavior as multiple related searches running across subtopics and data sources, which means the system is not looking for a single perfect page. It is assembling an answer from a web of material.

That is why the old instinct to chase one exact keyword looks increasingly thin. The prompt on the screen is only the beginning; the hidden query cluster is where visibility gets decided. If your content only covers the head term and nothing around it, you may never be present when the answer engine fans out into the neighboring questions that actually shape what gets cited.

What to stop doing, and what to build instead

The clearest practical lesson is that topical coverage beats narrow keyword chasing. Ahrefs argues that if you want to show up, you need to rank for the head term and the fan-out topics around it, which is really classic topic clustering updated for AI-era retrieval. The smart move is not to chase every probabilistic sub-query one by one, because those can change with each generation. It is to build a topic cluster that covers the full space so AI systems can retrieve you from multiple angles.

That change should force a hard audit of content strategy. The old top-of-funnel informational model looks weaker, especially for affiliates and other publishers that depended on generic how-to content. Generic explainers built to catch easy clicks are much less defensible when search engines are doing more of the synthesis themselves. Distinctive expertise, original evidence, and a wider cluster around the subject matter now matter more than volume for its own sake.

A practical reset looks like this:

• Stop treating a single page as the whole strategy.
• Stop publishing generic how-to content that only serves one visible query.
• Stop assuming exact-match fan-out keywords are stable enough to target individually.

Start doing this instead:

• Build topic clusters that cover the head term, the adjacent questions, and the follow-up intents AI systems are likely to generate.
• Add material that is hard to replicate, including original examples, stronger explanations, and more distinctive media.
• Treat the goal as becoming the source AI search can pull from repeatedly, not just the page that happens to rank once.

That is the difference between old-school SEO and visibility in answer engines. One is built around matching a phrase. The other is built around owning a subject.

Why this matters at scale

This is not a niche behavior hiding at the edges of search. Google says AI Overviews now has more than 2.5 billion monthly active users, and AI Mode has surpassed one billion monthly users. That scale turns what used to look experimental into something every serious search team has to account for, because visibility inside AI answers now sits in front of a very large audience.

The publisher side of the story is even more sobering. Reuters Institute research, based on 280 news executives across 51 countries and territories, says publishers expect search traffic to almost halve, or fall 43%, over the next three years as search engines become AI-driven answer engines. The fear is straightforward: if the answer is increasingly delivered on the search page itself, referral traffic could dry up. Many publishers are responding by emphasizing distinctive content, more video, and a more human face, which is a reminder that AI-era search rewards identity as much as information.

That is the key strategic lesson buried inside all the noise. AI search does not just change distribution. It changes what looks worth citing. If your content feels interchangeable, it is easier for an answer engine to pass over you. If it feels unmistakably yours, it is easier to pull you into the answer.

Google is making the new visibility measurable

For years, one of the biggest frustrations around AI search was opacity. Google is starting to soften that problem with Search Console generative AI performance reports for Search and Discover, which include impressions for AI Overviews and AI Mode. That does not make the system fully transparent, but it does give teams a first-party way to see whether they are appearing in generative surfaces at all.

Google has also introduced Preferred Sources in AI Overviews and AI Mode. The company says users are twice as likely to click through to a Preferred Source, and more than 345,000 unique sources had already been selected. It is also adding more inline links and website previews, which suggests that even inside AI-generated answers, Google still wants to send people out to the web.

That makes the measurement brief much sharper. Rankings alone are no longer enough. Teams now need to watch impressions in AI Overviews and AI Mode, track whether their domains are being selected or preferred, and compare that visibility against referral traffic. If the search result is becoming an answer, then the right metric is no longer only where the page sits. It is whether the brand is being surfaced, cited, and clicked from inside the answer layer.

The practical field guide for the next phase

The hype version of this story says SEO is dead. The more useful version says classic SEO is no longer sufficient on its own. Query fan-out, topic clustering, and AI answer surfaces are changing the way visibility gets assigned, but the durable work still looks familiar: build depth, earn trust, and cover a subject so thoroughly that the engine has to keep coming back to you.

The brands that win now will not be the ones that chase every fleeting sub-query. They will be the ones that build enough authority across the whole topic space that AI systems can retrieve them from multiple directions, prefer them when given the choice, and cite them when the answer is assembled.