How Can Your Page Get Cited in AI Overviews Without Top Rankings?

What changed in AI Overview citations?

The link between a top-10 ranking and an AI Overview citation has weakened in recent measurements. Ahrefs reported that 37.9% of URLs cited in AI Overviews appeared in the first 10 SERP blocks for the same query in a March 2026 analysis of 863,000 keyword SERPs and 4 million AI Overview URLs. The other cited URLs were split almost evenly between positions 11 to 100 and beyond the top 100. Ahrefs reported 31.2% in positions 11 to 100 and 31.0% beyond the top 100 blocks.

That is a major contrast with Ahrefs' July 2025 study, which examined 1.9 million citations from 1 million AI Overviews and found that 76.1% of cited pages ranked in the top 10. The before-and-after is useful, but it is not proof of a clean algorithm change. Ahrefs says its citation-parsing methodology improved between the studies, so better detection may explain part of the difference. Search Engine Journal makes the same caution explicit.

The defensible conclusion is narrower than 'rankings no longer matter.' Rankings and citations still overlap. What changed is that a ranking for the exact visible query is a less reliable proxy for citation inclusion than it once appeared to be. A page outside page one can now be part of the source set, and a page at the top can be absent from the answer.

This is why the old reporting model breaks down. A team that tracks only target-keyword rank can see an apparent win while losing visibility inside the answer that many searchers read first. Citation Share measures a different question: what percentage of relevant AI answers cite your brand. Both measures matter, but neither substitutes for the other.

1. July 2025: 76.1% of AI Overview-cited pages ranked in the top 10 in Ahrefs' 1.9 million-citation study.
2. March 2026: 37.9% of cited URLs appeared in the first 10 SERP blocks in Ahrefs' 863,000-SERP analysis.
3. What to do: treat the comparison as a signal to measure citations directly, not as a reason to abandon organic SEO.

Why can a page outside page one get cited?

A page can get cited because AI Overviews may retrieve answers for related subtopics, not only the exact query shown in the search box. Google states that AI Overviews and AI Mode may use query fan-out, issuing multiple related searches across subtopics and data sources to develop a response. Google Search Central also says the systems can identify more supporting web pages while the response is generated.

That means an AI Overview is not simply a new presentation layer for the first page of a single SERP. A buyer may ask a broad question about a product category. The generated answer may need evidence about implementation, compatibility, regional coverage, pricing model, security, use case, or a narrower comparison. A page can be the clearest source for one of those needs even when it does not rank strongly for the broad opening query.

Google does not publish the individual fan-out searches behind every answer. It also says AI Overviews and AI Mode may use different models and techniques, so their responses and linked pages can vary. Do not claim that a single heading, schema field, or formatting pattern forces a citation. That would turn a changing retrieval system into a fake checklist.

The useful editorial implication is coverage with purpose. Build pages that answer distinct buyer questions fully, make the answer easy to verify, and connect related pages with clear internal links. That is compatible with GEO and conventional SEO because Google says the same foundational SEO best practices remain relevant for AI features.

1. Map the buyer question, then identify the connected questions required to answer it accurately.
2. Give each important question a direct answer near the relevant heading.
3. Use sources, original research, product documentation, or first-hand evidence when the claim requires proof.
4. Link related pages so readers and crawlers can understand the topic relationship.

Who does this affect most?

This change affects teams that use page-one ranking as their only measure of search visibility. A B2B SaaS company may rank for a category term yet fail to appear when an AI answer compares has or explains an implementation constraint. A local service business may have weak visibility for a broad service phrase but hold useful evidence for a precise location and service question.

It also affects publishers that chase broad keywords while leaving the questions beneath them unanswered. A page optimized around a head term can be relevant, but it may not provide the passage, source, or detail needed for a related retrieval path. The work is not to stuff more variants into one page. It is to publish an answer a reader can inspect and an AI system can support with a citation.

Smaller specialist sites have a real opening here, although not a free pass. Google says AI has can create opportunities for more types of sites to appear, and that supporting links help people explore content they may not have discovered before. Eligibility still requires a page to be indexed and eligible to appear in Google Search with a snippet. A page that cannot be crawled, indexed, or rendered properly cannot win on a clever content angle.

The hard part is measurement. Google includes AI-has traffic in the Search Console Performance report under the Web search type, rather than supplying a separate universal AI Overview report. That makes conventional traffic analysis necessary but incomplete. Teams need a recurring prompt set and a citation record alongside rank, organic traffic, and conversion data. That is the difference between monitoring activity and measuring Answer Engine Optimization.

1. B2B teams competing on category, alternative, comparison, and use-case questions.
2. Local and multi-location businesses competing on service and geographic intent.
3. Specialist publishers with credible depth on a narrow part of a broader topic.
4. SEO teams whose dashboards report rankings but do not record AI answer presence.

How should your AI citation strategy respond?

Your AI citation strategy should keep ranking work, then add a citation layer built around the full question universe. Start with the questions buyers actually need answered before they choose. Separate broad category questions from comparison questions, technical questions, local questions, and decision-stage questions. The goal is not to predict every hidden fan-out query. It is to remove obvious gaps in the evidence a good answer requires.

Each page should answer its central question early, then support that answer with useful detail. Use question-shaped headings, named sources, dates for time-sensitive information, and tables when a reader needs to compare options. The structure helps a person scan the page and creates a clearer evidence surface than a vague block of promotional copy. This is Citation Engineering in practice: quality, coverage, and freshness, not an attempt to manipulate a model.

Do not create pages only because they contain a keyword. Create them because the page can add a specific answer, original data point, documented experience, or decision aid. Google recommends helpful, reliable, people-first content for AI has and says there are no additional technical requirements or special AI markup required for eligibility. A new AI text file or a made-up schema tactic is not the answer.

Keep the technical foundation intact. Google says a supporting link must be indexed and eligible to appear in Google Search with a snippet. Review crawl access, indexability, canonical handling, internal links, mobile usability, and the page experience before treating a citation gap as an editorial problem. Then use Search Console and analytics to understand traffic and conversion outcomes, while tracking citations separately across the questions that matter.

Finally, measure the result at the answer level. Record the prompt, engine, date, cited domains, your presence, and whether the citation supports a commercial decision or a supporting fact. Track Citation Share for the defined prompt set, Citation Count per day for volume, and Answer Presence for breadth. A rank report tells you where a URL sits. A citation report tells you whether the answer chose you.

1. Retain technical SEO and ranking work as the eligibility foundation.
2. Map category, comparison, implementation, local, and evidence questions around each buyer decision.
3. Publish distinct pages and sections that answer those questions directly.
4. Add dated sources and original evidence where the topic calls for proof.
5. Track Citation Share and Answer Presence alongside rankings, traffic, and conversions.

What should you stop doing?

Stop treating a move from rank eight to rank three as proof that you will be cited. It may improve your odds, and it may improve click visibility, but the March 2026 Ahrefs data shows that most cited URLs in its analysis were not in the first 10 SERP blocks for the same query. The page-one report is no longer a complete answer to the citation question.

Stop treating pages outside page one as disposable. Review them against the questions they answer, their evidence, their technical eligibility, and their connections to the rest of the site. A page at position 35 may be irrelevant, thin, or broken. It may also be the only precise source on a subtopic. The point is not to celebrate low rankings. It is to evaluate the page against the job it can do.

Stop promising a mechanical route into AI Overviews. Google says there is no special schema.org structured data or additional technical requirement for these features, and it does not guarantee indexing or serving. Anyone selling a single technical trick as an AI Overview switch is replacing uncertainty with theatre.

The better discipline is simple. Improve what a buyer can verify, keep the site eligible for Search, and watch the answers themselves. Rankings got you found. Citations get you chosen. In an AI answer, there is no page two.

How do you know whether the strategy is working?

The strategy is working when your brand appears more often in the defined AI answers that matter to your buyers, not merely when a broader keyword report improves. Establish a stable prompt set by intent and revisit it on a regular cadence. Keep the prompt wording, engine, date, response, and cited URLs so that changes can be audited instead of guessed.

Compare citation performance with organic data, but keep the metrics separate. Google says AI-has impressions and clicks are reported within Search Console's Web search type. Use that data for traffic trends, then use analytics for engagement and conversions. Do not assign every organic session to an AI Overview or treat every AI citation as a conversion. The evidence needs to stay honest.

A strong reporting view has two layers. The first is visibility: Citation Share, Citation Count per day, Answer Presence, and Share of Voice against named competitors. The second is business impact: visits, qualified leads, calls, bookings, or signups where the available attribution supports the conclusion. The first layer tells you whether the answer cites you. The second tells you whether that visibility is useful.

This is an operating discipline, not a one-off audit. AI answers can vary by query and change over time. Keep publishing, refresh dated claims, repair technical issues, and use the citation record to decide which gaps deserve the next piece of editorial work. The result is a strategy grounded in evidence rather than a race for a rank that may not decide the answer.

Sources

Source: Ahrefs

Ahrefs found that 37.9% of URLs cited in AI Overviews appeared within the first 10 SERP blocks for the same query, based on 863,000 keyword SERPs and 4 million AI Overview URLs. Ahrefs, 2026-03-02

Source: Ahrefs

Ahrefs' July 2025 study found that 76.1% of AI Overview-cited pages ranked in the top 10, based on 1.9 million citations from 1 million AI Overviews. Ahrefs, 2025-07-21

Source: Google Search Central

Google states that AI Overviews and AI Mode may use query fan-out and that the same foundational SEO best practices remain relevant for AI features. Google Search Central, 2025-12-10

Source: Search Engine Journal

Search Engine Journal reported Ahrefs' methodology caution and the March 2026 distribution of cited pages across top-10, positions 11 to 100, and beyond the top 100. Search Engine Journal, 2026-03-02

Source: SEO Turtle

SEO Turtle framed the news reaction around AI Overviews citing pages that do not rank on page one and linked to the before-and-after research. SEO Turtle, 2026-09-02