AI Overviews turn search into reading sessions

Search intent still determines what content companies create. Once an AI Overview appears in the Google SERP, however, users no longer behave like they do in classic search. The results page becomes a reading session: direct answers capture attention, time on page rises, and the usual gaps between navigational, informational, transactional, local, and video intent blur. For SEO and content teams, that means a strategic shift away from click-only focus toward visibility within the SERP itself.

The article summarizes insights from a growth memo based on anonymized clickstream data. Together with Eric Van Buskirk of Clickstream Solutions, roughly 846,000 U.S. Google search sessions were analyzed. Core thesis: with an AI Overview, time on the SERP barely depends on intent anymore—instead, a uniform reading pattern emerges across intent types.

The new mental model of search intent

For decades, SEO practice assumed navigational is fast, informational slower, and local stays longer because of maps and listings. SERPs without an AI Overview still show this pattern clearly—about 12 percent of navigational searchers remain on the SERP after 21 seconds, versus 32 percent for local.

With an AI Overview, that logic collapses. After 21 seconds, all five major intents sit between 41.9 and 48.5 percent time on SERP; the spread shrinks to about six percentage points. Informational, local, navigational, transactional, and video behave almost the same on the SERP. The AI Overview compresses different search intents into shared reading behavior.

• Old model: Time on SERP follows intent; navigational leaves quickly, local stays longer.
• New model: With an AI Overview, intent differences on the SERP are barely measurable; users read the answer surface.

Longer sessions and more context

Average SERP sessions grow roughly fourfold with an AI Overview. Direct answers carry more context and need more reading time; the original query intent fades into the background. Users increasingly validate information on the SERP before clicking—or skip the click entirely. For reporting, classic intent segments explain dwell time on AI SERPs poorly; dashboards should treat AIO presence as a dimension.

The analysis explicitly compares sessions with and without an AI Overview. Without one, the old intent pattern remains visible and serves as a baseline for classic optimization. With an overview, that separation fades—a finding that challenges a decades-old SEO assumption that time on SERP is a reliable proxy for intent and for snippet and content prioritization.

That widens the gap between links and answers. Google used to serve ten blue links; users were responsible for verification and depth after the click, and Google learned from click behavior. When the answer engine synthesizes the response, correctness and completeness sit with the system. Bing describes this shift as evolving the index: from ranking documents to supporting reliable, verifiable answers—from fetching the best pages to fetching the best information to synthesize.

The second impression on the SERP

Classic SEO metrics often optimize the first impression: title, snippet, position. With AI Overviews, a second impression gains weight—how product, category, and blog pages are perceived in the answer environment, in citations, and in the remaining link list after users have read the overview. Teams should check whether content answers questions the overview leaves open and whether meta descriptions hold up against competitors on the same SERP.

The memo headline “reading sessions” captures the shift: search behaves more like a short read on the results page than a fast handoff. Product pages must convince after the overview; category pages need clear value in the snippet; blog posts need precise subheads and fact-based paragraphs that stay citable in summaries. Teams that maximize organic clicks without measuring SERP readability optimize the wrong layer.

Related work also notes that users behave differently in AI Overviews than in separate AI Mode—both surfaces should not share one report. For GEO and content strategy, track each channel separately and run snippet tests specifically on SERPs with an AIO present.

Search intent remains relevant for content planning, but the SERP surface redefines user behavior. Teams that only optimize landing pages without modeling the reading session on the SERP underestimate users who decide before they click.

Branded prompts and monitoring

Alongside classic brand-keyword defense in ads, disciplined monitoring of branded prompts in LLMs and AI surfaces pays off. Companies should verify which facts, products, and narratives models return for their brand—not only for generic product or pain-point queries. Missing or wrong brand representation in AI answers becomes a reputation and conversion risk before any click happens.

Action items for SEO teams

• Track SERP dwell time and intent clustering with AI Overviews in analytics and Search Console.
• Structure content to complement overview questions and supply verifiable facts.
• Audit meta descriptions and snippets against competitors on the same SERP regularly.
• Track branded prompts in AI answers and align with PR, product, and legal teams.
• Align stakeholders with data on intent compression and longer reading sessions on the SERP.

AI Overviews turn search from a pure navigation ritual into a reading session on the results page. Teams that adapt intent models, content formats, and reporting accordingly stay actionable in generative search surfaces.