What to do now that AI Overviews turned search into reading sessions

The New Mental Model of Search Intent

For more than two decades, search engine optimization operated on a fundamental, predictable law: user intent dictates user behavior on the search engine results page (SERP). If a user typed a brand name into Google, they were executing a navigational search. They knew exactly where they wanted to go, resulting in a rapid, friction-free exit from the SERP. If they typed a broad query like “how to repair a running toilet,” they were in informational mode, settling in for a slower, multi-page discovery process.

That paradigm has officially broken. The catalyst for this shift is the widespread integration of Google’s AI Overviews (AIO). By rendering a cohesive, synthesized block of text at the very top of the organic search results, Google has fundamentally altered how users interact with information. Instead of treating the SERP as a mere launchpad of links, users are now treating it as the destination itself. Search has transformed into a reading session.

To understand the depth of this transformation, Eric Van Buskirk of Clickstream Solutions analyzed anonymized clickstream data from approximately 846,000 U.S.-based Google search sessions. The findings reveal a stark divergence in user behavior depending on whether an AI Overview is present on the SERP.

The Flattening and Compression of Intent

In a traditional search environment without an AI Overview, time-on-page scales predictably with user intent. The differences between how different searchers interact with the page are distinct:

• Navigational searchers are highly efficient. After 21 seconds, only 12% of these users remain on the search results page. They find their target link immediately and click away.
• Local searchers linger much longer. Because local results are densely packed with maps, reviews, operational hours, and address details, 32% of these searchers are still evaluating their options on the SERP after 21 seconds.
• Informational searchers sit comfortably in the middle, scanning organic snippets before committing to a click.

However, when an AI Overview is present, this behavioral spread completely collapses. The distinctive signatures of different search intents disappear, compressing into a tight, uniform cluster.

With an AIO active on the page, the percentage of users still on the SERP after 21 seconds across all five primary search intents—informational, local, navigational, transactional, and video—concentrates remarkably between 41.9% and 48.5%. The variation between a fast-paced navigational search and a complex local search shrinks to a mere six percentage points.

This means that search sessions on pages featuring an AI Overview are, on average, nearly four times longer for quick-intent queries. The presence of the AI block arrests the user’s journey, turning passive scrollers into active readers regardless of their initial goal.

Why Searchers Stay: The Grounding of Answers

This dramatic expansion of SERP dwell time is driven by the density of the information provided. Rather than forcing users to click through to three different websites to compare facts, Google’s AI Overview performs the heavy lifting of aggregation, synthesis, and summarization directly on the search page. The searcher stays because they are reading a custom-generated answer.

This shift from indexing pages to synthesizing answers represents a structural evolution in how search engines work. Bing outlined this transition clearly in an industry publication titled “Evolving the role of the index: From ranking pages to supporting answers.” As search engines evolve from simple indexers to proactive answer engines, the core engineering constraint changes:

“Grounding an AI-generated answer introduces a fundamentally different constraint: The system is no longer just pointing to information, it is using it. The goal shifts from ‘fetch the best documents’ to ‘fetch the best information to synthesize into a reliable, verifiable answer.'”

In the classic search model, search engines bore little responsibility for the accuracy of individual landing pages. They merely provided a ranked list of “ten blue links” and monitored user clicks to refine those rankings. If a user clicked a link and found poor information, they bounced back to the search page, signaling to the algorithm that the destination was low quality.

Under the new AI-driven model, the responsibility for accuracy shifts to the search engine. To generate a synthesized answer, the engine must extract factual statements from underlying web documents, verify their accuracy against known entities, and present them cohesively. The engine is no longer just a signpost; it is an author. Consequently, the user spends their time scrutinizing the engine’s output before—or instead of—navigating to an external source.

This behavioral change is not opt-in. Most Google users are not actively seeking out AI tools; they are simply using the default search interface as they have for decades. As Google continuously rolls out these formats, users are guided into AI-centric reading patterns. This passive, forced adoption has met some resistance, as evidenced by a 30% surge in installations of privacy-focused alternatives like DuckDuckGo. Nevertheless, with Google reporting that over 1.5 billion people actively interact with AI Overviews, this is the standard operating environment for modern organic search.

Winning the “Second Impression”

If users are spending more time reading the SERP, how do brands secure their attention and drive traffic to their websites? The answer lies in optimizing for the “second impression.”

The second impression occurs during the back-scroll. When a user lands on a SERP containing an AIO, they naturally focus on the large AI-generated block at the top. They read the summary, absorb the primary takeaways, and then begin scrolling down to view the traditional organic results. If they do not find an immediate click, or if they seek to verify the AIO’s assertions, they scroll back up. This second pass is the equivalent of a consumer re-evaluating options on a retail shelf. It is the moment where credibility, structural clarity, and visual cues determine which link earns the click.

To win this crucial micro-moment, websites must optimize their search listings based on the specific page templates they are trying to rank. The strategy requires a tailored approach across three primary page types: Product Detail Pages (PDP), Category Detail Pages (CDP), and Blog Content.

Product Detail Pages (PDP)

When users scroll past an AI Overview to evaluate specific product pages, they are looking for rapid, verifiable commercial signals. In this context, basic metadata is insufficient. To capture the second impression, brands must focus on three primary elements outside of standard title and description tags:

• Robust Product Schema Markup: Ensure that your schema implementation is flawless, containing accurate fields for aggregateRating, review, offers, and availability. If your structured data lacks any of these elements, search engines will display a bare-bones listing next to a competitor’s fully populated rich snippet, severely damaging your click-through rate.
• Prioritizing Review Velocity and Volume: Review counts act as a powerful trust signal. If your product listing displays 47 reviews while a competitor’s listing displays 2,300, you will likely lose the second-impression click even if your meta description is expertly written. Elevating review acquisition must be treated as a core search engine optimization objective.
• Schema Image Arrays: Populate your product schema with multiple high-quality image URLs. This provides the search engine with the necessary assets to render diverse visual layouts, such as thumbnail grids or product carousels, directly in the organic search results.

Category Detail Pages (CDP)

Category pages face a unique challenge in an AI-dominated SERP. Because an AI Overview often lists and briefly describes the top product recommendations for a given category, your category page must position itself as the ultimate portal for deeper comparison and selection.

• Implementing ItemList Schema: Utilize ItemList structured data to map the products featured on your category page. This allows search engines to pull your inventory into organic product carousels, expanding your listing’s vertical footprint on the SERP and capturing more attention during the back-scroll.
• Surfacing Filter and Sort UI: Build clean, logical internal linking structures for your category filters (such as “shop by price,” “shop by brand,” or “top-rated”). This increases the likelihood that Google will surface these facets as clickable sitelinks beneath your main listing, demonstrating immediate utility to the user.
• Maximizing Inventory Depth: A category page showcasing only a dozen products compares poorly against a competitor listing that references hundreds of options. Users executing a second-pass evaluation seek comprehensiveness. Ensure your category previews convey depth and variety.

Blog and Informational Content

When users search for informational topics, the AI Overview often provides the direct answer, satisfying their immediate curiosity. To earn a “validation click”—where a user clicks through to verify the AI’s claims or to seek deeper analysis—your informational content must project immediate authority and timeliness.

• Managing Date Snippets: Ensure that your datePublished and dateModified schema fields are accurate and visible on the page. In time-sensitive industries, an article clearly marked with a current year will consistently outperform a listing displaying an outdated publication date.
• Author Entity Resolution: Leverage Article schema with fully populated author profiles. Link the author field to a sameAs URL, such as a verified LinkedIn profile or an authoritative author biography page. This helps search engines resolve the writer’s identity as a recognized entity, boosting E-E-A-T (Experience, Expertise, Authoritativeness, and Trustworthiness) metrics and reinforcing credibility on the SERP.

What Intent Compression Means for Operators

The discovery that AI Overviews compress search intent behavior does not mean that historical search engine optimization strategies are obsolete. The extensive work digital marketing and content teams have dedicated to mapping search intent, clustering keywords, and creating highly targeted landing pages remains the foundation of a successful organic search presence.

The primary change lies in the prediction layer built on top of that content strategy. While user search intent still dictates *what* content you need to write, it no longer dictates *how* those users will engage with the search results page. The traditional metrics of search behavior—such as expected click-through rates by ranking position and average time-on-SERP—must be re-evaluated through the lens of AIO presence.

Consequently, optimization work must expand to include on-SERP performance. SEO professionals must dedicate more resources to analyzing how their listings appear in direct proximity to AI-generated text blocks, rather than simply tracking raw keyword rankings. This involves continuous testing of meta descriptions, title tags, schema markups, and rich media assets to maximize clickability on the second impression.

By adapting to this reading-centric search environment, brands can protect their organic visibility, capture high-value validation clicks, and continue to drive meaningful engagement in an increasingly AI-driven landscape.