Google AI Overview Citations: Why Being #1 on Google No Longer Guarantees Inclusion

Most businesses assume there is a simple relationship between Google rankings and AI Overviews: rank high, get cited. New data from Ahrefs shows that assumption is no longer accurate, and the gap is widening fast.

In July 2025, around 76% of pages cited in AI Overviews also ranked in Google's top 10 for the same query. That number is now 38%, based on Ahrefs' updated analysis of 863,000 keyword SERPs and 4 million AI Overview URLs. The connection between ranking and being cited has been cut roughly in half within eight months.

Key Takeaway: Only 38% of Google AI Overview citations now come from top-10 ranked pages, down from 76% in July 2025. The driver is query fan-out: AI generates sub-queries and pulls sources from across those results. For SMEs, topical depth, structured content, and FAQ coverage now matter more than holding a single page one ranking.

Written by Derek Chua, digital marketing consultant and founder of Magnified Technologies. Derek works with Singapore SMEs on GEO content strategy and has been tracking AI search developments since Google AI Overviews launched in 2024.

For businesses that have spent years chasing page one rankings, this is both unsettling and, if you read it correctly, an opportunity.

What the Data Actually Shows

Ahrefs ran two versions of the analysis. The first included all result types: ads, featured snippets, People Also Ask boxes, video packs, and organic listings. The second looked only at standard organic links.

The findings were consistent across both tests.

Including all SERP features:

• 37.9% of cited pages ranked in positions 1-10
• 31.2% ranked in positions 11-100
• 31.0% ranked beyond position 100

Organic results only:

• 37.1% of cited pages ranked in the top 10
• 26.2% ranked in positions 11-100
• 36.7% ranked outside the top 100

The organic-only split is notable. When SERP features are removed from the picture, the proportion of citations from beyond position 100 actually increases slightly. This suggests some AI Overview citations come from pages that appear in SERP features (featured snippets, image packs, video results) rather than in standard organic rankings at all.

Why Rankings Have Become a Weaker Predictor

The mechanism behind this shift is Google's query fan-out process. When a user submits a query that triggers an AI Overview, Google does not just look at the standard results for that one query. It splits the original query into multiple related sub-queries, then identifies pages appearing most frequently across those sub-query results.

Those pages get cited in the final AI Overview, regardless of whether they ranked in the top 10 for the original query.

Ahrefs attributes part of the data shift to improved parsing methodology: they can now detect more AI Overview citations than was possible in July 2025. So the two studies are not perfectly comparable. But even accounting for methodology changes, the directional trend is clear: fan-out queries are playing a larger role in source selection.

Google upgraded AI Overviews to Gemini 3 globally in January 2026, which Ahrefs notes as relevant timing. Gemini 3's ability to handle more complex, long-tail queries may be expanding how aggressively fan-out is applied. Google has not confirmed specific changes to fan-out behavior.

The trend line reinforces this reading. A late 2024 study showed around 75% of AI Overview citations from top-12 pages. BrightEdge reported 54% overlap in October 2025 (using different methodology). Ahrefs now puts the figure at 38%. The direction is consistent regardless of which data source you use.

YouTube: The Most Cited Domain You're Not Thinking About

Among AI Overview citations that did not rank in Google's top 100 for the same keyword, 18.2% were YouTube URLs. YouTube accounted for 5.6% of all AI Overview citations in the Ahrefs dataset.

According to Ahrefs' Brand Radar data, YouTube is now the most-cited domain in AI Overviews overall and has grown 34% in the past six months.

A separate Ahrefs study of 75,000 brands found that YouTube mentions, including video titles, transcripts, and descriptions, are the strongest correlating factor with AI Overview visibility, stronger even than organic search rankings.

The implication for any business producing long-form content: if that content exists only as a written article, you are missing a significant AI citation channel. A short explainer video, even a repurposed version of your written content, gives AI a second format through which to discover and cite your material.

What Actually Drives Citation Selection

The fan-out model shifts the question from "can I rank for this keyword?" to "is my content citation-worthy across related queries?" These require different strategies.

Topic breadth beats single-keyword depth. Fan-out queries evaluate content across related sub-queries. An article covering only one narrow aspect of a topic is less likely to appear across those sub-queries than a piece addressing the topic from multiple angles. Topic clusters, meaning groups of connected articles covering a subject from different entry points, become more valuable under this model.

Structure signals citation-worthiness. The pages most likely to be cited share a common characteristic: AI can extract a clear, structured answer from them. This means using declarative headings that function as standalone answers, including FAQ sections that directly address question-phrased queries, and opening sections with concise, quotable summaries. These are not cosmetic choices. They directly affect whether AI can extract your content for a fan-out sub-query.

Schema reinforces discoverability. FAQ schema, How-To schema, and Article schema help AI identify and use the right sections of your content in response to specific sub-queries. Structured data is not a ranking hack at this point. It is a basic requirement for content that wants to be cited across AI systems.

Domain authority is less protective than it was. In the old model, large brand rankings served as a proxy for AI Overview inclusion. That buffer is eroding. Smaller, more authoritative sites covering specific topics thoroughly can now displace larger sites with thin coverage of those topics.

The Good News for Smaller Sites

If your business operates in a specific niche, a particular type of professional service, a defined local market, or a specialist product category, the fan-out model may work in your favour.

Fan-out queries evaluate topic authority at scale. A boutique accounting firm that publishes thorough, well-structured content on corporate secretarial requirements, GST registration, or Employment Pass compliance may find its content surfaced across multiple related sub-queries, even against pages from larger regional firms.

The question is not "can I outrank this competitor?" It is "can I produce content that is more citation-worthy on this specific topic?" Those are different challenges, and the second one is more achievable for most SMEs.

How We Approach This with Magnified Clients

At Magnified, we have been advising clients for several months that treating AI search as a "rank first, get cited second" problem is the wrong frame. We see this reflected in our own content performance: articles that get picked up across AI Overviews are not always the ones with the highest organic rankings. They are the ones with the clearest structure, the most direct answers to question-phrased queries, and the strongest topical coherence within a cluster.

When we build content strategies now, we treat every piece as a node in a topic cluster. The goal is not one article ranking for one keyword. The goal is a set of interconnected articles that collectively establish authority across a topic, so that when AI evaluates multiple sub-queries within that domain, our client's content appears consistently across them.

This connects directly to what we covered in our Google AI Overviews growth article, the CORE research on GEO optimisation, and our guide on what AI actually sees when it visits your website. The citation data from Ahrefs is the mechanism behind all of it: AI is not just reading your top page, it is traversing your topic cluster.

Practical Actions to Take Now

1. Audit your topic coverage, not just your rankings. Look at the full set of questions your target audience asks around your core service or product. Identify gaps, meaning topics where you have no published content, and prioritise those over ranking improvements on existing pages.

2. Add FAQ sections to existing pages. Question-phrased queries are precisely what fan-out generates. A FAQ section covering the five to eight most common questions related to a topic converts an article from a ranking asset into a citation asset.

3. Implement FAQ schema. Structure your FAQs in proper schema markup so AI can parse them precisely. This is a half-day technical implementation for most sites.

4. Review your heading structure. Each H2 should be able to stand alone as an answer to a specific question. If your headings are generic or decorative, they are not contributing to citation-worthiness.

5. Consider video for your highest-value topics. YouTube's 34% growth in AI Overview citations is not coincidental. If your most valuable content exists only as written text, a short video version opens a second citation pathway.

---

Frequently Asked Questions

Does this mean SEO is no longer important for AI Overview visibility?

No. Organic rankings still correlate with citation. The point is that rankings are no longer sufficient on their own. A page can rank in the top 10 and still not be cited in an AI Overview, while a page ranking in position 30 can be cited if it is more citation-worthy. SEO and GEO work together, but GEO requires additional structural and topical investment beyond traditional ranking optimisation.

What is query fan-out and how does it affect my content?

Query fan-out is Google's process of splitting an original user query into multiple related sub-queries when generating an AI Overview. Pages cited in the final Overview are those appearing most consistently across those sub-query results. This means content covering a topic from multiple angles, rather than targeting a single keyword, is more likely to appear across fan-out results.

Do I need to rank in the top 100 to be cited in AI Overviews?

No. The Ahrefs data shows roughly 31% of all AI Overview citations come from pages outside the top 100 for the original query. The citation selection process pulls from fan-out sub-query results, which can surface pages that do not rank for the primary keyword at all.

Why is YouTube appearing so frequently in AI Overview citations?

YouTube is the most-cited domain in AI Overviews, accounting for 5.6% of all citations in Ahrefs' dataset. AI appears to treat video content with descriptive titles, transcripts, and descriptions as authoritative and citation-worthy. YouTube's engagement and content clarity signals may align well with the factors AI uses to evaluate whether a source is worth citing.

How many articles do I need to build topical authority for AI search?

There is no fixed number, but the pattern that works is a cluster of five to eight interconnected articles covering a topic from different angles and intent types. This gives fan-out queries enough content to pull from across sub-queries. A single comprehensive article rarely establishes the same topical signal as a well-structured cluster.

---

If you want to understand how your current content performs against AI citation criteria and identify the gaps creating the most risk, our team offers GEO content audits tailored to your business. Get in touch with Magnified to find out where your content stands.

Sources: Ahrefs blog, March 2026 (updated AI Overview citations analysis, 863K keywords, 4M URLs); Search Engine Journal, March 3, 2026 (reporting on Ahrefs analysis); Ahrefs Brand Radar data on YouTube citation growth.