Does ranking in Google's top 10 still get you into AI Overviews?

The short answer: ranking helps, but it is not the gate

A top-10 ranking does not decide whether a page is cited in the AI Overview for the same query. In an Ahrefs study of 863,000 search results and 4 million cited URLs, published in March 2026 with the data period not stated, 38% of pages cited in AI Overviews ranked in Google's top 10 for that query.¹ The other cited pages came from further down, or ranked for different queries.

That figure is often quoted as "ranking no longer matters". It says something narrower. Ahrefs sells SEO and AI visibility tools, and its figure measures overlap for one query at a time. Ranking work still decides whether a page can be cited at all and how many clicks it keeps. What changes is the target: the answer is built from more than one search.

This post sets out what the overlap figure measures, where AI Overviews appear most, and when ranking work still pays for AI visibility.

What does the overlap figure actually measure?

It measures the share of cited pages that also rank in the top 10 for the exact query that triggered the AI Overview. It does not measure the chance that a top-10 page is cited, and it does not say the cited pages rank poorly in general.

Two readings are easy to confuse. "38% of cited pages rank in the top 10" is a statement about citations, from a study whose data period is not stated.¹ "38% of top-10 pages get cited" would be a statement about rankings, and the study does not make it. A top-10 page can be passed over, and a page ranked twentieth for the query can be cited because it answers one part of the question better.

Citation overlap

The share of pages cited in an AI Overview that also rank in Google's top 10 organic results for the same query, measured query by query.

Our view: the useful lesson is not that rankings stopped mattering. It is that a single keyword is the wrong unit for planning AI visibility.

Why the overlap is low: answers are built from several searches

Google says so directly. Its documentation for site owners states that AI Overviews and AI Mode may use a "query fan-out" technique, issuing multiple related searches across subtopics and data sources to develop a response.² A page can be cited because it ranks for one of those related searches, not the query the user typed.

Take a buyer who asks which project management tool suits a small agency. The answer may draw on searches about pricing tiers, client portals and time tracking. A vendor's page on client portals can be cited even if it ranks nowhere for the original question, because it is the best result for one sub-question.

That explains the low overlap without any mystery about AI favouring obscure pages. It also tells you where to work: on the sub-questions, each with a page that answers it clearly in its first lines.

How one query becomes several searches

1. Buyer's query
2. Related searches by subtopic
3. Supporting pages found
4. Answer written
5. Links shown

Where do AI Overviews appear most?

They appear most on questions, health queries and comparisons. In an Ahrefs study of 146 million desktop search results from September 2025, data now ten months old, AI Overviews appeared on 20.5% of results overall, 44.1% of medical queries and 57.9% of question queries.⁴

Comparisons are the most exposed. In Seer Interactive's panel of 53 brands, tracked from January 2025 to February 2026, 95.4% of "X vs Y" queries showed an AI Overview.³ Seer is a search agency, the panel is 53 brands it tracks and the comparison set is 280 queries, so treat that as a strong signal rather than a market average.

For a buyer comparing vendors, the AI Overview is now the default first screen. That is where citation matters more than position, and where the overlap question has real money attached.

Where ranking work still pays

It pays in three places: eligibility, clicks and the sub-questions themselves. None of them goes away because the overlap figure is low.

Eligibility

Google states that to be shown as a supporting link in AI Overviews or AI Mode, a page must be indexed and eligible to be shown in Search with a snippet, and that there are no additional technical requirements.² Technical SEO, indexing and snippet controls are therefore the entry ticket. A page that is blocked, noindexed or set to nosnippet cannot be cited.

Clicks

Being cited is worth having. In Seer's panel (January 2025–February 2026), pages cited in an AI Overview earned about 120% more organic clicks per impression than when they were not cited, and organic click-through rate on AI Overview queries rose from 1.3% in December 2025 to 2.4% in February 2026.³ A ranked page that is also cited keeps far more of the remaining clicks. Our guide to which AI search numbers belong in a board report treats cited clicks as one of the four lines worth tracking.

The sub-questions

Ranking for the related searches is how a page gets into the fan-out. That is still ranking work, aimed at a different list of queries.

Ranking work vs citation work

Ranking work

• Target: the head query
• Unit: one keyword, one page
• Win: a position on the results page
• Still needed for indexing, snippets and clicks

Citation work

• Target: the sub-questions behind the answer
• Unit: a prompt and its related searches
• Win: being named or linked in the answer
• Measured as share of answer across engines

What should you change in your plan?

Change the unit of planning from keywords to prompts, and keep the ranking basics. In practice that is four steps.

1. For each priority query, write down the three to five sub-questions a full answer would need: price, fit, alternatives, setup, risks.
2. Check which of your pages rank for those sub-questions, and which pages are cited in the AI Overview today.
3. Give each sub-question one page, or one clearly headed section, that answers it in the first two sentences.
4. Track citations on the prompt, not positions on the keyword, and repeat the check on several days, because answers vary.

Our view: if a provider tells you rankings no longer matter, ask how they plan to get your pages indexed and eligible for snippets. If a provider reports only rankings, ask how they will know whether you are cited. The evidence on schema, llms.txt and other citation tactics points the same way: clear pages that answer specific questions do more than markup.

An AI Visibility Audit from Sigzen AI measures citations across six engines, including Google's AI Overviews and AI Mode, on 150 buyer prompts run on three separate days. The audit page lists what it covers.