The AI Search Disruption Study

Abstract

An analysis of how Google's AI Overviews, rolled out to all U.S. users in May 2024, are quietly draining clicks from local businesses before a user ever taps a result. Combining a randomized field experiment with longitudinal keyword tracking and primary North Atlanta field research, the study documents a measurable, causal drop in outbound clicks each time an AI Overview appears, and explains why most small business owners cannot see the loss in their analytics.

Methodology

The study synthesizes a 2026 randomized controlled experiment on click behavior, Semrush's longitudinal tracking panel of more than 10 million U.S. keywords, BrightEdge's commercial-industry query data, and primary field research conducted directly on real North Atlanta service searches. The randomized experiment was designed to isolate the effect of AI Overviews under controlled conditions, producing a causal rather than correlational result.

Findings

A 2026 randomized field experiment measured a 39.8% drop in outbound clicks every time an AI Overview appears, with some publishers in heavily AI-saturated categories reporting losses as steep as 89% on affected queries. AI Overview presence is expanding rapidly, doubling from 6.49% of queries in January 2025 to 13.14% by March 2025, and reaching 48% of tracked commercial queries by February 2026. The searches hit hardest are high-intent commercial queries that turn into phone calls and booked jobs. Traditional rankings offer little protection, because 66% of URLs cited inside AI Overviews do not rank in the traditional top 10 for the same query.

Overview

Something fundamental changed in Google search in May 2024, when AI Overviews rolled out to every U.S. user. Since then a quiet traffic drain has been pulling clicks away from local businesses before a single result is tapped. The damage is real and measurable, yet most small business owners cannot see it in their analytics. This study documents what is happening, why it matters, and what North Atlanta businesses can do before the window to adapt closes.

A measurable, causal traffic drop

A 2026 randomized field experiment put a hard number on the damage: a 39.8% drop in outbound clicks every time an AI Overview appears. Some publishers in heavily AI-saturated categories report losses as steep as 89% on affected queries. Because the experiment was randomized and controlled, the result is causal, not correlational. The reason owners miss it is that standard analytics count clicks that occur, not searches that end before a click is ever made.

The expansion is fast and hitting commercial queries

Semrush tracked more than 10 million U.S. keywords and watched AI Overview presence double in eight weeks, from 6.49% of queries in January 2025 to 13.14% by March 2025. By February 2026, BrightEdge recorded AI Overviews on 48% of tracked commercial queries, a 58% year-over-year jump. Those commercial searches, the ones that turn into phone calls and booked jobs, are exactly the high-value queries most aggressively affected.

Rankings no longer guarantee visibility

The study’s most significant structural finding is that 66% of URLs cited inside AI Overviews do not rank in the traditional top 10 for the same query. A business can hold a top-five organic ranking and still be entirely absent from the AI summary most users read first. Ranking well and being cited by AI are now two separate challenges requiring different strategies.

Practical implications

• Audit your most important commercial queries to see whether AI Overviews appear and which sources they cite.
• Watch Google Search Console for rising impressions paired with flat or declining clicks, a reliable signal that AI Overviews are intercepting traffic.
• Restructure content to answer specific customer questions directly, with FAQ-style formatting, structured data markup, and clear demonstrations of expertise.
• Diversify beyond Google organic to reduce dependence on clicks that AI is increasingly capturing, and move quickly while the landscape is still forming.

These findings directly inform our AI Search Optimization service, connect to our broader SEO work, and expand the AI Search and GEO guide.