A smaller competitor is quoted in the AI Overview while your firm ranks number one because Google does not build the Overview from the typed query's ranking. It splits the question into sub-queries, then cites pages that surface across them. A firm with one strong page for the head term competes in one of those searches. A smaller firm with narrow pages competes in several.
Earlier Firm AEO posts covered the 38% top-10 figure, how fan-out should shape a site's page structure, and how to rewrite a single passage. This post is the diagnostic for one complaint, built on a dataset none of them used: Surfer's measurement of which ranking pattern actually gets cited, plus Pew's count of government sources. It ends in five checks, run in order.
- Ahrefs' March 2026 study of 863,000 keyword results and 4 million AI Overview URLs found 37.9% of cited pages ranked in Google's top 10, down from about 76% in July 2025, and Ahrefs attributed the drop to query fan-out.
- In Surfer's study of 10,000 keywords and 33,000 fan-out queries, pages ranking for the main query and at least one fan-out query made up 51.2% of AI Overview citations, against 19.6% for pages ranking only for the main query.
- Surfer measured a Spearman correlation of 0.77 between the number of fan-out queries a page ranks for and its likelihood of being cited, and stated plainly that the finding is a correlation, not proof of cause.
- Pew Research Center found .gov sites made up 6% of AI Overview sources against 2% of standard search results, so the page quoted above a law firm is sometimes a court or agency page rather than a competitor.
Why is a smaller competitor quoted in the AI Overview while we rank number one?
Because position one for the typed query is now a minority route into the citation set. Ahrefs analysed 863,000 keyword results and 4 million AI Overview URLs in March 2026 and found 37.9% of cited pages ranked in the top 10, 31.2% ranked between 11 and 100, and 31.0% ranked beyond 100. In July 2025 the top-10 share had been about 76%.
Ahrefs' explanation was query fan-out: when an AI Overview is triggered, Google splits the original query into related sub-queries and draws on the results for those. Firm size, ad spend and domain age do not appear in that mechanism. A two-lawyer firm with a page that answers "does my car insurance cover a rideshare accident in Texas" can surface in a sub-query that a 40-lawyer firm's general car accident page never ranks for.
What does ranking for fan-out queries have to do with being quoted?
Almost everything that Surfer could measure. Surfer took the top 10 pages for 10,000 keywords, found 76% carried an AI Overview, used Gemini to extract 33,000 fan-out queries, and scraped the top 10 for those as well, 173,902 URLs in all. Among cited pages that ranked anywhere in that set, pages ranking for both the main query and a fan-out query took the largest share.
Two caveats travel with the numbers. Surfer stated that the result is a correlation, not proof that fan-out rankings cause citations. Surfer also found fan-out queries were consistent across repeated searches only about 27% of the time, which means the sub-queries, and so the cited set, move between searches. A firm can be quoted on Tuesday and missing on Wednesday without anything on its site changing.
| Ranking pattern of the cited page | Share of citations | What it means for a firm in position one |
|---|---|---|
| Ranks for the main query and at least one fan-out query | 51.2% | The strongest position; a head-term page that also answers sub-questions |
| Ranks only for one or more fan-out queries | 29.2% | Where the smaller competitor usually sits |
| Ranks only for the main query | 19.6% | Where a firm with one strong head-term page sits |
| Ranking for fan-outs versus main query alone | 49% more likely to be cited | Breadth across sub-questions beats depth on one term |
| Correlation, fan-outs ranked vs citation likelihood | Spearman 0.77 | Strong association, not demonstrated cause |
Is the page quoted above us always a competitor?
No, and checking that first saves wasted work. Pew Research Center's study of 68,879 searches by 900 US adults found .gov sites made up 6% of AI Overview sources against 2% of standard results. For a legal question about a deadline or a filing fee, the source above a firm is often a court self-help page or a state agency, which no firm should expect to displace.
Video is the other non-competitor. Ahrefs found YouTube made up 18.2% of the AI Overview citations that did not rank for the query and 5.6% of all citations, making it the most-cited domain in the study. Surfer reported Google's own properties and YouTube together at roughly 40% of AI Overview citations. When the quoted source is a court page or a video, the firm's realistic goal is to be the firm-side source cited beside it, not the only one.
How do you diagnose why your page was skipped?
Run five checks in a fixed order, because each one rules out a cause the next one depends on. The table lists them with the evidence each produces and a typical time to run it. The times are illustrative estimates from Firm AEO's own audits, not measured benchmarks, and a firm with dozens of practice pages should sample its top five by revenue first.
| Check | What to look at | What a failure means | Typical time |
|---|---|---|---|
| 1. Is the Overview firing on your query? | Search the exact term you rank for, logged out, then its question form | You rank for a head term; the Overview fires on the question version | 15 minutes |
| 2. Is the page eligible? | Page source for nosnippet, data-nosnippet, max-snippet, noindex | The page has opted its text out of being quoted | 5 minutes per page |
| 3. Who is quoted, and what kind of source? | Each cited link: court, agency, video, directory or firm | A .gov or video source is a co-citation target, not a rival | 20 minutes |
| 4. Which sub-questions do you rank for? | Write the six to eight sub-questions the Overview answers; check your rank for each | You cover the head term and none of the sub-questions | 1 to 2 hours |
| 5. Does the page state the answer in one sentence? | The paragraph under each heading, read alone | No passage an engine can lift without rewriting | 30 minutes per page |
What should a firm that ranks first do next?
Keep the ranking page and build outward from it, in the order the checks failed. Google Search Central states that a page needs only to be indexed and eligible for a snippet to appear in AI features, with no special markup or AI text file required, so nothing in the fix list is a technical trick. The order below is the one Firm AEO uses.
One marketing constraint applies throughout. A firm that is later quoted should not advertise that it is "featured by Google AI" or "the top-ranked" firm without a verifiable basis, since ABA Model Rule 7.1 prohibits misleading communications and the citation set changes between searches. The firm's ethics counsel decides whether any such statement can be made.
- 1. Clear any eligibility block found in check two before writing anything new.
- 2. Add a one-sentence, jurisdiction-qualified answer under each existing heading of the ranking page.
- 3. Publish narrow pages for the two or three sub-questions where a firm, not a court, is the natural source.
- 4. Link each new page from the ranking page with the sub-question as the anchor text.
- 5. Re-run checks one, three and four monthly, logged out, three runs per query, and record the cited set rather than a single result.
Do more backlinks or a bigger site help a firm get into the AI Overview?
Not directly by any published measure. The Ahrefs and Surfer studies tie AI Overview citations to ranking across sub-queries, and neither reports site size as a factor. Links still help pages rank for those sub-queries, which is the indirect route. A larger firm's advantage is that it can publish more sub-question pages, not that the Overview prefers it.
Why does the firm quoted in the AI Overview change from one day to the next?
Because the sub-queries change. Surfer found fan-out queries were consistent across repeated searches only about 27% of the time, so the pool of candidate pages shifts between searches. A single screenshot proves little. Firm AEO records the cited set over three logged-out runs per query each month and reports the share of runs in which the firm appears.
Should a firm stop optimizing the page that ranks number one?
No. In Surfer's data the largest share of citations, 51.2%, went to pages ranking for the main query and at least one fan-out query, so the ranking page is the base the rest is built on. The change is adding sub-question coverage around it rather than spending more effort on the head term alone.
- 1.Ahrefs, Only 38% of AI Overview citations rank in the top 10 (March 2, 2026)
- 2.Surfer, Ranking for multiple fan-out queries increases your chances of getting cited in AIOs (173,902 URLs studied)
- 3.Surfer, 67.82% of AIO citations don't rank in Google's top 10. So where do they come from?
- 4.Search Engine Land, AI Overview fan-out rankings boost citation odds by 161%: Study
- 5.Pew Research Center, Google users are less likely to click on links when an AI summary appears (July 22, 2025)
- 6.Google Search Central, AI features and your website
- 7.ABA Model Rule 7.1, Communications Concerning a Lawyer's Services