Google AI Overviews are making a visible change to how people use search. A UX study described in the source article found that when an AI Overview appears, many users treat it as the answer and stop before visiting outside websites.
The result is a practical shift for the web: search is becoming less about clicking through a list of links and more about absorbing information directly on the results page.
Clicks fall when AI answers appear
The study’s central finding is direct: when Google’s AI Overview appears in search results, outbound clicks decline sharply. On desktop, clicks to external websites drop by about two-thirds. On mobile, the drop is nearly 50 percent.
Even the links inside the AI Overview do not receive much attention. The study found that users clicked those links in 7.4 percent of desktop sessions and 19 percent of mobile sessions.
That behavior matters because Google, Apple, OpenAI, and others are presenting the future of search as a move toward answers rather than links. The study shows that this is not just a product vision. Users are already responding to AI answers as if they can replace part of the traditional search journey.
For websites, especially publishers, the problem is straightforward. If users get enough information from the AI Overview, the original destination may lose the visit, even when its information helps shape what appears on the page.
Users still look for trust signals
The study does not suggest that users accept every answer blindly. It found that people generally approach search in two steps. First, they judge whether a source seems trustworthy. Then they decide whether the content fits what they were looking for.
Recognizable brands and .gov/.edu domains had a clear advantage. When those sources appeared, they were clicked first in 58 percent of cases.
Some users also looked beyond the first answer. 18% opened a second source for validation, such as a Reddit thread, a YouTube video, or another organic result.
But even when users interacted with the AI Overview, they often stayed near the top. While 88 percent tapped "Show more," the median scroll depth was just 30 percent. The study found that 86 percent mostly skimmed the AI Overview, and only a small number reached the bottom.
This means the top third of the AI answer carries much of the attention and trust. In that way, AI Overviews behave somewhat like traditional snippets: what appears early is far more likely to be noticed.
Personal and high-stakes searches get more scrutiny
Search behavior changed depending on the subject. For high-stakes queries, including health or finance, users scrolled more deeply, with average scroll depth rising above 50 percent.
For lower-risk tasks, such as searching for coupons, scroll depth dropped to just 34 percent. The more personal or consequential the topic, the more likely users were to examine the AI answer and seek confirmation.
In 38 percent of AIO sessions on sensitive topics, users opened a second source to verify the information they were seeing. That is an important distinction. AI Overviews may reduce clicks overall, but users still appear more cautious when the decision feels important.
When users did leave the AI Overview, they often chose platforms that provide social proof or demonstration. Reddit, YouTube, and other forums served as validation sources, especially for younger users.
DIY searches showed this clearly. Many users skipped the AI Overview in favor of videos that showed real-world demonstrations. On average, users spent 37 seconds watching videos, compared with 31 seconds inside AI Overviews.
Age, device, and shopping searches change the pattern
The study also found differences by age and device. Younger users and people with higher digital literacy were more likely to treat AI answers as final. Among users aged 25 to 34, half accepted AI responses without looking further.
Mobile users also behaved differently from desktop users. They scrolled further through AI Overviews, with 54 percent scroll depth on mobile compared with 29 percent on desktop.
Older users still preferred traditional blue links. That suggests AI search adoption is uneven, with different habits depending on user group and search context.
Commercial queries followed another pattern. For product comparisons and shopping recommendations, users looked at AI Overviews but continued relying on other platforms. Amazon, in particular, acted like a meta search engine.
In shopping-related searches, traditional organic results had less influence. Paid ads and Google-curated verticals received seven out of ten clicks. AI Overviews sent very little traffic, partly because users preferred visual elements, star ratings, and user reviews when making buying decisions.
Search marketing shifts toward presence
For marketers, the study points to a change in measurement. Classic rankings and click volume may become less useful when users stay inside the results page. Visibility inside AI Overviews may become the more important signal.
That could mean tracking how often a brand appears in AI answers, where it appears, and how prominently it is cited. The source article mentions citation rank and "share of voice" as examples of metrics that may matter more as search changes.
This shift will not affect every site equally. Sites that depend on direct traffic, especially publishers, face the largest risk. Some may end up buying traffic back from platforms such as Google.
The broader implication is more platform power. The source article points to Google, Meta, and eventually Apple as companies that could gain more control as users spend less time clicking out to independent websites.
Google recently announced plans to roll out AI Overviews globally. A more advanced version, called AI Mode, is already in development and is designed to keep users inside the Google ecosystem longer.
The study was carried out by Kevin Indig in collaboration with Eric van Buskirk and his team, and published on Growth Memo. It involved 69 English-speaking users in the U.S., including 42 on mobile and 27 on desktop.
Participants completed eight real Google searches, six of which included an AI Overview. Topics included shopping, finance, health, and DIY. Sessions were recorded using UXtweak, capturing screen activity, scrolling, mouse movement, and audio commentary through "think aloud" protocols.
Three trained analysts reviewed over 500 videos frame by frame. They logged scroll depth, dwell time, click behavior, user reactions, and final decisions, then analyzed trust scores, click rates, and scroll behavior.
The picture that emerges is simple but significant: Google AI Overviews are not only changing what appears in search. They are changing what users consider necessary after the answer appears.