Google's AI Overviews — the AI-generated answer that appears at the top of many search results pages — have become a fixture on commercial queries since the global rollout in 2024. BrightEdge's tracking shows they now appear on 75%+ of healthcare queries, 70%+ of e-commerce queries, and a rapidly growing share of local service queries.
When AI Overviews appear, traditional results get pushed below the fold on most devices. The businesses cited inside the AI Overview capture a disproportionate share of clicks; everyone else fights for the scraps.
How AI Overviews choose what to cite
Semrush's 2024 study of 8,000 AI Overviews found that the cited sources skewed heavily toward: pages that directly answered a question in the first paragraph, FAQ-formatted content, well-structured H2/H3 hierarchies, and pages with clear schema markup. Domain authority mattered, but less than content structure.
Pages that 'self-contain' an answer — meaning the first 100-200 words fully address the query — were cited far more often than longer articles that buried the answer.
How to structure for AI citation
Write content in a Q&A format whenever it fits. 'What does AC repair cost in Dayton?' as an H2, followed by a clear answer in the first paragraph.
Use bulleted lists for steps, criteria, and comparisons. AI engines lift these almost verbatim into answers.
Add schema: FAQPage, HowTo, and Article markup all increase the probability of citation.
Keep paragraphs short — 2-4 sentences. The atomic unit of an AI Overview is the lift-able paragraph, not the long-form article.
Track impact in Search Console
Search Console doesn't yet break out AI Overview impressions cleanly, but you can identify them by sudden impression spikes on informational queries with stable or declining clicks. That's the signature of an AI Overview eating your clicks.
The fix is to be the cited source rather than the unfeatured 7th result. Aggressive structure and answer-first writing are the path.