AI Overviews reshape search behavior, extend time on results pages

The search page is becoming the product

AI Overviews are changing search in a way that goes beyond traffic loss. Kevin Indig’s memo argues that the old idea that intent predicts behavior is breaking down, because AI Overviews flatten different query types into the same reading pattern on the results page itself. In practice, that means navigational, local, informational, and transactional searches no longer behave as separate journeys once an AI answer appears.

The shift is not subtle. Search Engine Land’s analysis of about 846,000 U.S.-based Google search sessions from February and March 2026 found that time on the search results page can stretch to roughly four times what it used to be when AI Overviews are present. That is the clearest signal yet that users are not just clicking less or more. They are spending longer inside Google, reading, evaluating, and revisiting the question before they ever leave the results experience.

Why intent no longer tells the full story

Classic search strategy has always leaned on intent mapping. A person looking for a nearby service behaves differently from someone comparing products or looking up a definition, and content teams have built entire page structures around that distinction. AI Overviews blur that logic by presenting a synthesized answer that pulls the user into a single on-SERP reading mode, even when the original query would normally lead to very different behavior.

That matters because the old funnel assumption is weaker now. The user is no longer moving cleanly from query to click; the AI answer reduces the original question first, then leaves a second decision point behind. Indig calls that the second impression, and that is the part brands have to win if they want to stay relevant after the summary has done its work.

The numbers show the habit is already changing

Google launched Search Generative Experience in May 2023, then rebranded the feature as AI Overviews at Google I/O on May 14, 2024. Google also said at the time that it was bringing AI Overviews to everyone in the U.S., and later described the feature as something used by more than a billion people. It has also said that the links inside AI Overviews can earn more clicks than the same page would have received as a traditional listing for that query.

The user behavior side is more sobering. Pew Research Center found that in March 2025, about 58% of respondents used a search that produced an AI-generated summary. In that same analysis, Google users were less likely to click on links when an AI summary appeared, and they very rarely clicked the sources cited in those summaries. Pew also found that one-in-five U.S. adults said AI summaries were extremely or very useful, while more than half said they were only somewhat useful.

That split is important. Users do not need to love AI summaries for the behavior to change. They only need to trust them enough to keep reading on the page. Even before AI Overviews became widespread, SparkToro’s 2024 zero-click study estimated that 360 out of every 1,000 U.S. Google searches sent a click to the open web. AI Overviews do not create zero-click behavior from scratch. They intensify a trend that was already well underway.

What the “second impression” means for content

If the search page is now a reading surface, content has to earn visibility in two places at once. First, it has to be legible to the model so it can be summarized. Second, it has to be compelling enough that the user still wants more after the summary has answered the easiest version of the question.

That changes how you structure pages, pages sections, and even the language inside them. The strongest material is tight enough to be summarized cleanly, but not so flattened that it loses the detail a serious searcher still wants. In practical terms, that means:

• Put the answer up front, then widen out into nuance, examples, and comparison points.
• Use clear subheads that break a topic into discrete chunks an AI system can parse.
• Write definitions, steps, and feature explanations in plain language that can stand alone.
• Include specific facts, numbers, and distinctions that make the page worth opening after the summary.

This is not about writing less. It is about writing with a cleaner hierarchy, so the page can serve both the summary layer and the reader who arrives ready to go deeper.

What Google is optimizing for, and why that matters

Google’s own messaging makes the direction of travel pretty clear. The company has said people are asking longer questions, diving deeper into complex subjects, and uncovering new perspectives. It has also expanded AI Overviews with more planning and research capabilities, and it has begun layering in AI-organized search results and follow-up conversational behavior through AI Mode.

That creates a very specific challenge for publishers and brands. Google is trying to keep users inside its own results experience longer, while still framing the system as useful for exploration. If the search page becomes the place where research starts and continues, then publishers are competing not just for a click, but for the trust and attention that used to belong to the open web by default.

How to adapt without chasing ghosts

The answer is not to abandon search traffic strategy. It is to stop treating blue-link traffic as the only win condition. AI Search Visibility now depends on whether your content can survive summary extraction and still invite the next action, whether that next action is a click, a deeper browse, or a follow-up search.

The pages that hold up best tend to do three things well. They state the core answer fast, they prove it with specifics, and they leave enough structure in place for the user to keep moving. That might mean tighter intros, more explicit headings, sharper comparisons, or stronger fact patterns around the sections most likely to be summarized.

For brands, the practical goal is no longer just to be found. It is to be remembered inside the results page. That is the real pressure point in the AI Overview era: search is becoming less like a doorway and more like a destination, and the pages that adapt to that reality will keep earning attention even when the first answer comes before the click.