How to Rank in Google AI Overviews: What the Data Shows Works

Key takeaways

• Organic rankings are still the entry ticket, but not the whole game. Ahrefs found 76% of AI Overview citations came from top-10 pages in July 2025; its updated analysis puts that at 38% — structure now decides the rest.
• Pages built around question headings with 30–50 word direct answers are the easiest format for Google's AI to extract and cite.
• Being cited pays off. Seer Interactive found brands cited in AI Overviews earn roughly 35% higher organic CTR than brands that are not cited.
• Statistics with named sources and visible freshness signals (published dates, recent updates) raise your odds of being quoted.
• Search Console does not label AI Overview traffic. You measure impact indirectly through query-level CTR shifts and dedicated AI-visibility tracking.

Do AI Overviews really cite pages that already rank organically?

Mostly yes — but less than they used to. When Ahrefs analyzed AI Overview citations in July 2025, 76% of cited pages also ranked in the top 10 for the same query. Its updated study, covering 863,000 keywords and roughly 4 million AI Overview URLs, found that figure dropped to 38%. Ranking well helps, but it no longer guarantees a citation.

The updated Ahrefs data shows where the rest of the citations come from, and it should change how you think about this:

BrightEdge saw the overlap moving the other way in its own tracking: its 16-month study found citations increasingly matching organic rankings, growing from 32% to 54% overlap. Different datasets, different methods — but one conclusion holds in both. Strong organic rankings make you eligible. They do not make you quotable.

Why the loosening? Ahrefs points to Google's "query fan-out": the AI breaks one search into several sub-queries and pulls sources for each. A page that answers a specific sub-question clearly can get cited even if it would never crack the top 10 for the head term. That is exactly the opening a smaller business site can exploit.

How much traffic do AI Overviews actually cost you?

A lot, if you only watch clicks. Seer Interactive tracked 3,119 search terms across 42 organizations — about 25.1 million organic impressions — from June 2024 to September 2025. Organic CTR on queries with AI Overviews fell from 1.76% to 0.61%, a decline of roughly two-thirds. Queries without AI Overviews dropped too, from 2.72% to 1.62%.

Pew Research Center's browsing study of real users found the same pattern from the other side: people clicked a traditional result in just 8% of sessions where an AI summary appeared, versus 15% of sessions without one. Only 1% of sessions with an AI summary produced a click on a source cited inside the summary itself.

Read those numbers together and two things are true at once. First, informational traffic is shrinking whether you play or not. Second, the citation slot is the only position on the page that is gaining relative value. Seer Interactive's data shows brands cited inside AI Overviews earn about 35% higher organic CTR than brands that are not cited — being the quoted source is the new position one.

So the goal shifts: from "get the click" to "get the citation, then earn the click from the people who want depth." The visitors who do click through after reading an AI summary are further along in their decision. Treat that traffic as smaller but better.

How do you structure content so Google's AI cites it?

Format every section as a question with a self-contained answer. Put the question in an H2 or H3, then answer it directly in 30–50 words in the very first paragraph below it. Only after that, expand with evidence, examples, and nuance. Google's AI extracts passages, not pages — make each passage quotable on its own.

This is not theory. It mirrors how featured snippets have worked for years, and AI Overviews use similar passage-level extraction. The practical rules we apply on client sites:

• Question headings. Write H2s the way a real person types them: "How much does X cost?", "Is Y worth it?", "How long does Z take?" Avoid clever headings the AI cannot map to a query.
• Front-loaded answers. The first paragraph under each heading must answer the question completely in 30–50 words. No wind-up, no "it depends" without the dependency named immediately.
• Short sentences and plain definitions. If you use a term, define it the first time in one sentence. AI systems prefer passages they can quote without editing.
• Lists and tables for comparisons. Structured elements survive extraction better than dense prose.
• One idea per section. If a passage needs three paragraphs of context to make sense, it will not be quoted.

This question-and-answer architecture is also the foundation of generative engine optimization more broadly — the same format that wins AI Overviews gets you cited by ChatGPT and Perplexity. We cover the full framework in our generative engine optimization guide for businesses.

Does FAQPage schema still matter for AI Overviews?

Yes — but for a different reason than most people think. Google cut FAQ rich results back in August 2023, limiting them to well-known government and health sites, so FAQPage markup rarely earns you visible dropdowns anymore. What it still does is tell machines, explicitly and unambiguously, "this text is a question, and this text is its answer." That is exactly the mapping AI systems need.

Think of schema as labeling, not decoration. FAQPage and Article markup will not rescue weak content, and Google has been clear that no special markup is required to appear in AI Overviews. But when two pages answer the same question equally well, the one whose structure is machine-readable is the safer bet for extraction.

Our position: keep FAQPage schema on any page with a genuine FAQ section, write the questions the way users phrase them, and keep answers to two to four sentences. Skip markup games — marking up content that is not visibly on the page violates Google's guidelines and risks a manual action.

What role do statistics, sources, and freshness play?

A big one. AI systems building an answer prefer passages that contain specific, attributable facts — "38% of citations come from top-10 pages, per Ahrefs" beats "many citations come from top pages" every time. Named sources give the AI something concrete to pass along, and they give human readers a reason to trust you.

Freshness matters in two ways. Google has confirmed that AI Overviews favor up-to-date content for queries where recency is relevant, and the citation data reflects it — fast-moving topics see cited pages rotate constantly. Practically, that means:

• Show a real published date and a real "last updated" date. Do not fake the update — change the content when you change the date.
• Refresh statistics on money pages and top blog posts at least twice a year. Replace old figures with newer studies and say so in the text.
• Add new sections as the topic evolves rather than publishing thin duplicate posts.

This is one area where small businesses can genuinely outmaneuver bigger competitors. A 50-page site can be fully refreshed in a week. A 5,000-page publisher cannot.

How do you measure AI Overview performance in Search Console?

Indirectly — that is the honest answer. Google lumps AI Overview impressions and clicks into the same Search Console buckets as everything else, with no filter or breakdown. There is no "AI Overviews" row to check. You have to infer impact from query-level patterns instead.

The workflow that works:

1. Export your informational queries (how, what, why, best, vs.) from the last 16 months and compare CTR year over year, not just clicks. A impressions-up, CTR-down pattern on how-to queries is the classic AI Overview signature.
2. Track average position alongside CTR. If position holds steady while CTR falls, an AI Overview is likely absorbing clicks above you.
3. Watch brand queries separately. A rising share of branded impressions often means non-brand informational queries are being answered without clicks — your brand becomes the discovery channel.
4. Spot-check manually. Search your 20 most important queries in an incognito window monthly and record whether an AI Overview appears and who gets cited. It is tedious and it is still the only ground truth.

Because manual tracking does not scale, we built citation monitoring into our own SEO automation platform, AutoRankFlow, which tracks where your pages are quoted across AI answers alongside classic rankings. Whether you use a tool or a spreadsheet, the point stands: if you are not tracking citations, you are flying blind on the fastest-growing part of the results page.

How long does it take to start getting cited?

For a page that already ranks on page one, restructuring it into question-and-answer format can produce citations within a few weeks of recrawling. For a brand-new page on a newer domain, expect the same runway as organic SEO generally — typically three to six months before rankings stabilize enough to earn consistent citations.

The order of operations matters. Do not chase AI citations before you have organic traction on a query cluster; the Ahrefs data shows top-10 rankings still account for the single largest slice of citations. Earn the ranking first, then optimize the format. If you are starting from zero, our breakdown of how long SEO takes to work in 2026 sets realistic month-by-month expectations.

If you would rather have a team run this end to end — technical fixes, content restructuring, schema, and citation tracking — that is exactly what our SEO and GEO services for US businesses cover.

Frequently asked questions

Can I opt out of Google AI Overviews?

Not cleanly. You can block AI Overviews specifically with nosnippet or max-snippet directives, but those same controls suppress your regular snippets and featured snippets too. For most businesses the trade-off is not worth it — you lose more than you protect.

Do AI Overviews show up for every search?

No. Semrush's analysis of more than 10 million keywords found AI Overviews on 13.14% of US desktop queries in March 2025, up from 6.49% in January 2025. They concentrate on informational queries and rarely appear for transactional or local searches, though coverage keeps expanding.

Is ranking #1 enough to get cited in an AI Overview?

No. Ahrefs' 2026 analysis found only 38% of cited pages ranked in the top 10 at all. Rankings determine eligibility; extractable formatting — question headings, direct answers, sourced statistics — determines whether you actually get quoted.

Does FAQ schema guarantee I'll appear in AI Overviews?

No markup guarantees anything, and Google says no special schema is required for AI Overviews. FAQPage structured data simply makes your question-answer pairs unambiguous to machines, which supports extraction. Treat it as table stakes, not a shortcut.

How often should I update content to stay cited?

For competitive informational topics, review every three to six months. Update statistics, add sections for new sub-questions, and change the "last updated" date only when the content genuinely changes. Fast-moving topics may need monthly refreshes to hold citation slots.

Do AI Overviews matter for local service businesses?

Less than for informational sites — for now. AI Overviews rarely trigger on "near me" or high-intent local queries, where the map pack still dominates. They matter most for the research-stage content that brings local customers to you earlier in their decision.

What is the difference between SEO and GEO?

SEO optimizes pages to rank in a list of links; GEO (generative engine optimization) optimizes content to be quoted inside AI-generated answers. In practice they share the same foundation — technical health, authority, clear structure — and GEO adds extractable formatting and citation tracking on top.