Last year, SEO fundamentally changed as traffic dropped across the board, with sites losing up to 58% of clicks when AI overviews appear—up from 35% in April.
AI assistants research like humans—pulling from YouTube, Reddit, Quora, publishers, and niche sites—and cite different sources: Google AI overviews favor YouTube and Reddit, ChatGPT prefers news outlets, and Perplexity cites regional and niche domains.
AI uses 'query fan-out', breaking broad queries like 'best smartphone for photography' into dozens of subqueries (e.g., 'iPhone 17 Pro vs Samsung Galaxy S25 Ultra camera', 'best phone for low light photos'), then stitching answers from pages ranking for those subqueries—making branded mentions across those pages the strongest predictor of AI overview visibility.
For action-driven queries like 'backlink checker', 'snow removal service', or 'mortgage calculator', AI cannot satisfy users’ intent to act—so organic clicks still flow to traditional SEO, where content, links, and technical optimization remain highly effective.
To adapt, marketers must reverse-engineer AI citations using tools like Ahrefs Brand Radar, target predictable fan-out queries (e.g., containing 'Best', 'Versus', 'Review'), and prioritize transactional or tool-related keywords that resist AI summarization.
Brief overview
SEO has shifted into three distinct buckets: AI-driven discovery requiring brand mentions across niche-relevant pages, query fan-out behavior demanding coverage of subqueries, and action-oriented organic queries where traditional SEO still
AI overviews steal clicks—but brand presence across cited pages winsBranded mentions on listicles, comparisons, and reviews—not backlinks or domain rating—are the top correlate for visibility in Google's AI overviews.
Query fan-out means SEO now covers subqueries, not just head termsAI breaks 'best smartphone for photography' into dozens of specific subqueries and pulls from pages ranking for each—so you need mentions across all those pages.
Action queries still drive organic clicks—traditional SEO isn’t deadQueries like 'mortgage calculator' or 'snow removal service' force clicks because AI can’t complete actions—so content, links, and technical SEO still win there.
Questions this recording answers
7 questions, each answered where it is said
What percentage of clicks does the number one ranking page lose when an AI overview appears?
The number one ranking page loses about 35% of its clicks when an AI overview appears, and that loss has increased to 58% in the latest study.
Which sources do different AI assistants prefer for citations?
Google AI overviews lean heavily on YouTube, Reddit, and Quora; ChatGPT prefers publishers and news outlets; Perplexity tends to cite niche and regional sites.
What is the strongest predictor of visibility in Google's AI overviews according to the research?
Branded mentions have the strongest correlation with visibility in Google's AI overviews, more than backlinks, referring domains, and domain rating.
What are the three key shifts in search behavior revealed by last year's SEO turmoil?
The three key shifts are: (1) AI assistants now research like humans across YouTube, Reddit, forums, and review sites; (2) query fan-out—AI breaks queries into subqueries and pulls from pages ranking for each; (3) organic clicks are flowing to action-oriented, transactional queries where AI cannot satisfy the user.
What types of queries still drive organic clicks despite AI overviews?
Action-oriented queries like 'backlink checker', 'snow removal service', or 'mortgage calculator'—and transactional queries—still drive organic clicks because AI cannot fulfill the user's need to complete an action.
How can you identify the exact pages worth getting mentioned on for AI visibility?
Use Ahrefs Brand Radar with your brand and competitors, open the Cited Pages report, filter for queries containing words like 'Best', 'Top', 'Versus', 'Review', or 'Alternative'—these are predictable fan-out queries AI uses.
What is query fan-out and how does it affect SEO strategy?
Query fan-out is when AI breaks a single query like 'best smartphone for photography' into dozens of subqueries (e.g., 'iPhone 17 Pro vs Samsung Galaxy S25 Ultra camera'), then pulls from pages ranking for each subquery—so brands need mentions across those specific pages to appear in AI answers.
Key Quote
“Our research from April showed that when an AI overview appears in search results, the number one ranking page loses about 35% of its clicks.”
— 0
Key Quote
“And we just re-ran that study, and it's now a 58% loss.”
— 0
Key Quote
“86% of the top-cited sources are unique to each AI assistant.”