# Study finds Google AI Overviews cut outbound clicks by about 40%

Google’s AI Overviews reduced clicks from search results to outside websites by about 40% in a randomized experiment involving US desktop users, according to an analysis published on November 18.

The experiment assigned Chrome users to groups that either saw Google search normally or had AI Overviews hidden. During two weeks of searches in January and February 2026, queries that would ordinarily trigger an Overview generated an average of 0.37 outbound clicks when the summary was visible, compared with 0.62 when it was hidden. The researchers calculated the difference as a 39.8% reduction.

Search sessions ending without an external click also increased sharply, from 54% among users who did not see the summaries to 73% among those who did. The study found no corresponding change in advertising clicks. Its statistically significant result applied to informational searches; the samples for navigational and transactional searches were too small to support firm conclusions.

The findings carry important limits. The research is a working paper that has not yet completed peer review, covered desktop use only and observed participants for a relatively short period. Even so, random assignment gives the experiment a stronger basis for identifying cause and effect than comparisons between different kinds of searches.

The result aligns with other evidence that generated summaries change browsing behavior. A separate preregistered experiment involving 1,100 participants also found that removing AI Overviews and AI Mode increased publisher clicks. Earlier observational work from Pew Research Center reported that users clicked a conventional result on 8% of visits to pages containing an AI summary, compared with 15% on pages without one. That comparison was not controlled for differences between queries.

The potential impact is growing because AI-generated answers now appear much more frequently. DataForSEO’s panel showed AI Overviews on 61.43% of tracked US search keywords in August 2026, up from 5.93% in September 2024. Coverage varied by sector, reaching roughly 70% for health and finance keywords while remaining lower for dining and real estate.

For publishers, the experiment suggests that maintaining a high search position may no longer deliver the same traffic when a generated answer appears above conventional links. The analysis found little evidence that so-called generative engine optimization provides a separate remedy. It concluded that being discoverable in AI answers still largely depends on familiar search signals, presence on third-party sources and pages written clearly enough to be cited.

Across all searches, including those without an Overview, the researchers estimated the feature reduced organic Google clicks to websites by about 18.5%. That broader estimate could rise if generated summaries continue expanding, but the study’s scope means its precise effect outside US desktop informational searches remains unresolved.