To rank in Google AI Overviews, a page must first appear in Google's top-ranked results for the query, then satisfy a second layer of criteria: a direct answer in the opening lines, specific supporting data, and clear structure that lets Google's synthesis layer extract the key claim without ambiguity.
Google AI Overviews (formerly Search Generative Experience, or SGE) are synthesised answer boxes that appear at the top of search results for a significant portion of informational queries. They are generated by a large language model that draws on pages Google has already indexed and ranked. The model selects sources, extracts relevant passages, and composes a summary with citations shown as cards beneath the text.
Critically, AI Overviews are not a separate index. They draw on the same index and broadly the same quality signals as traditional search. A page that is not indexed, or that ranks outside the top results for a query, is unlikely to appear in that query's AI Overview. Earning a position in an AI Overview is therefore a two-stage problem: first rank, then satisfy the synthesis layer.
Yes, with a caveat. Google has stated that its ranking systems evaluate content quality rather than method of production. AI-generated content can rank and can appear in AI Overviews if it meets Google's quality criteria: genuine helpfulness, original analysis, accurate information, and evidence of expertise. What Google's systems target is content that is generated at scale without meaningful human review, lacks original insight, and exists primarily to capture keyword traffic. That class of content is at risk of manual action under the scaled content abuse policy, regardless of whether a human or a machine wrote it. The gate is quality, not authorship.
Quarrybank's approach reflects this directly: every page is drafted, checked against a quality gate, and approved by a human editor before it is published. Volume without that gate is a liability. Volume with it is a compounding asset.
Traditional ranking remains the entry ticket. Studies of AI Overview source pages consistently find that the majority come from the top five organic results. Getting into an AI Overview without a solid SEO foundation is very unlikely. The baseline requirements: indexed, crawlable, authoritative enough to rank on the first page.
Answer positioning is decisive. Google's synthesis layer pulls passages, not full pages. If your direct answer to the query does not appear in the first paragraph of the relevant section, Google may cite a page that answers more directly rather than scrolling further into yours. This is a structural change: the convention of building up to a conclusion, common in long-form SEO content, actively hurts AI Overview performance.
Specificity and first-hand data. Analysis of pages cited in AI Overviews finds they contain measurably more specific statistics, named methodologies, and verifiable claims than pages that rank but are not cited. A sentence like "in our test of 200 pages, 68% that led with a direct answer were cited in AI Overviews vs 21% that did not" is more likely to be extracted than "pages with clear answers perform better." Specificity is the signal.
E-E-A-T signals. Google's quality rater guidelines, which inform the models behind AI Overviews, weight experience, expertise, authoritativeness, and trustworthiness. Bylines, About pages, cited sources within the body, and demonstrable first-hand experience all contribute. A page that reads as if it could have been written by anyone about anything is lower-trust than a page that is clearly written by someone with direct experience of the topic.
Schema markup. FAQ and HowTo schema do not guarantee an AI Overview appearance, but they make the structure of a page easier for Google's extraction layer to parse. They are a low-cost, high-reliability improvement worth adding to any informational page.
Keyword stuffing for AI Overviews does not work. Unlike traditional search, where keyword density was a meaningful signal in the early years, AI synthesis looks for coherent, extractable answers. A page that repeats the query phrase throughout every paragraph without adding new information will not be cited in preference to a page that answers the question once, directly and specifically.
Thin pages that aggregate information available elsewhere without adding original analysis are also deprioritised. AI Overviews consistently cite pages with unique angles, proprietary data, or expert commentary that cannot be found on a dozen similar pages.
AI Overview appearances became reportable in Google Search Console in 2024 under the "AI Overviews" filter in the Search results report. The data is available but often misread: an AI Overview appearance does not always drive clicks, because the Overview may fully resolve the query. The metric to track is impression share in AI Overviews, not just clicks. A brand that appears consistently in AI Overviews for its core topics is building trust-signal recall even when the user does not click through.
Related: What is answer engine optimisation? · How to get cited by ChatGPT · AEO vs SEO
Quarrybank identifies queries where AI Overview appearances are winnable, drafts pages that satisfy both the ranking and citation criteria, and tracks the results.
Request founding access