Google’s AI Overviews now appear in roughly one out of every five searches, based on new data from Ahrefs that examined 146 million results. The analysis mapped 86 traits across queries to understand what prompts Google’s AI-generated summaries and where they remain absent.
The findings show a clear divide. AI Overviews most often appear when users type longer or question-style queries. They rarely surface for news, shopping, or local searches. Informational queries dominate the landscape, accounting for almost all appearances.
Ahrefs measured a 20.5 percent overall trigger rate across the dataset. But within that average, the spread is wide. Some categories, such as medical and science-related topics, reach nearly double the baseline, while others barely cross one percent.
Query structure plays a large part. Single-word searches activate AI Overviews in fewer than ten percent of cases. Queries stretching to seven words or more push that rate past 46 percent. Question formats drive the strongest correlation. Nearly six out of ten question-based searches include an AI Overview, compared with only fifteen percent for non-question phrases.
Intent matters too. Keywords categorized as informational account for 99.9 percent of all AI Overviews. Navigational, transactional, and commercial intents trail far behind, together representing only a fraction of one percent. The data suggests Google deploys AI most confidently where the purpose is to explain, not to sell or direct users elsewhere.
Industry differences stand out. Science queries show a 43.6 percent likelihood of an AI Overview, with health close behind at 43 percent. Pets, society, and technology categories also register above average. In contrast, shopping and real estate searches sit at the bottom, at 3.2 and 5.8 percent respectively. Sports and news perform modestly, around 15 percent.
The numbers within Google’s “Your Money or Your Life” (YMYL) category are notable. Medical searches lead the group, showing AI-generated answers 44 percent of the time. Financial queries follow at 23 percent, and safety topics at 31 percent. This is surprising given Google’s emphasis on reliability in areas that influence health or finances.
Non-branded searches are also more exposed. AI Overviews appear almost twice as often on general queries as they do for branded terms. Roughly 25 percent of generic searches trigger an AI summary, compared with 13 percent of branded ones.
Location and timeliness seem to suppress AI activity. Local queries, such as “dentist near me,” show AI Overviews in only eight percent of results. Highly time-sensitive or “newsy” searches drop to six percent. Google appears to limit automated summaries where freshness and real-world accuracy carry greater weight.
Adult and sensitive subjects remain mostly untouched. NSFW topics trigger AI Overviews just four percent of the time, with gambling, drugs, and violence-related terms staying below average.
Cost-per-click data offers no clear link. Across all CPC ranges, from near zero to premium ad categories, the rate of AI Overviews fluctuates within a narrow band of roughly twelve to twenty-eight percent.
Overall, the trend indicates that Google’s AI Overviews are tuned for comprehension-driven queries rather than navigation or purchase intent. For publishers, the exposure risk varies sharply by content type. Informational and educational sites face the highest overlap with AI-generated summaries, while ecommerce and news organizations remain relatively shielded.
For search strategists, the signal is plain: length, intent, and context determine visibility within Google’s AI layer. The 21 percent figure might look moderate, but behind it lies a growing pattern of machine-written summaries redefining what users see first.
Notes: This post was edited/created using GenAI tools.
Read next: Andrew Ng Breaks Down Vibe Coding and Its Impact on Programming Practices
The findings show a clear divide. AI Overviews most often appear when users type longer or question-style queries. They rarely surface for news, shopping, or local searches. Informational queries dominate the landscape, accounting for almost all appearances.
Ahrefs measured a 20.5 percent overall trigger rate across the dataset. But within that average, the spread is wide. Some categories, such as medical and science-related topics, reach nearly double the baseline, while others barely cross one percent.
Query structure plays a large part. Single-word searches activate AI Overviews in fewer than ten percent of cases. Queries stretching to seven words or more push that rate past 46 percent. Question formats drive the strongest correlation. Nearly six out of ten question-based searches include an AI Overview, compared with only fifteen percent for non-question phrases.
Intent matters too. Keywords categorized as informational account for 99.9 percent of all AI Overviews. Navigational, transactional, and commercial intents trail far behind, together representing only a fraction of one percent. The data suggests Google deploys AI most confidently where the purpose is to explain, not to sell or direct users elsewhere.
Industry differences stand out. Science queries show a 43.6 percent likelihood of an AI Overview, with health close behind at 43 percent. Pets, society, and technology categories also register above average. In contrast, shopping and real estate searches sit at the bottom, at 3.2 and 5.8 percent respectively. Sports and news perform modestly, around 15 percent.
The numbers within Google’s “Your Money or Your Life” (YMYL) category are notable. Medical searches lead the group, showing AI-generated answers 44 percent of the time. Financial queries follow at 23 percent, and safety topics at 31 percent. This is surprising given Google’s emphasis on reliability in areas that influence health or finances.
Non-branded searches are also more exposed. AI Overviews appear almost twice as often on general queries as they do for branded terms. Roughly 25 percent of generic searches trigger an AI summary, compared with 13 percent of branded ones.
Location and timeliness seem to suppress AI activity. Local queries, such as “dentist near me,” show AI Overviews in only eight percent of results. Highly time-sensitive or “newsy” searches drop to six percent. Google appears to limit automated summaries where freshness and real-world accuracy carry greater weight.
Adult and sensitive subjects remain mostly untouched. NSFW topics trigger AI Overviews just four percent of the time, with gambling, drugs, and violence-related terms staying below average.
Cost-per-click data offers no clear link. Across all CPC ranges, from near zero to premium ad categories, the rate of AI Overviews fluctuates within a narrow band of roughly twelve to twenty-eight percent.
Overall, the trend indicates that Google’s AI Overviews are tuned for comprehension-driven queries rather than navigation or purchase intent. For publishers, the exposure risk varies sharply by content type. Informational and educational sites face the highest overlap with AI-generated summaries, while ecommerce and news organizations remain relatively shielded.
For search strategists, the signal is plain: length, intent, and context determine visibility within Google’s AI layer. The 21 percent figure might look moderate, but behind it lies a growing pattern of machine-written summaries redefining what users see first.
Notes: This post was edited/created using GenAI tools.
Read next: Andrew Ng Breaks Down Vibe Coding and Its Impact on Programming Practices
