How to Rank in AI Overviews: Win the Fan-Out, Not the Keyword

Last updated: October 2026. This guide explains how to rank in AI Overviews for B2B SaaS teams: why the old "get into the top 10 first" advice no longer matches the data, how to map and win the fan-out, what Google says you do not need, and how to measure the result in Search Console now that Google reports AI impressions separately.

To rank in Google AI Overviews, stop treating the head keyword as the target. Google builds each AI Overview by running a "query fan-out", a set of related sub-searches, and cites the pages that answer those sub-questions best. So map the sub-questions behind your topic, answer each one clearly on an indexable page, and measure impressions, not clicks.

Key takeaways

• Google says AI Overviews and AI Mode may use a "query fan-out" technique, issuing multiple related searches across subtopics to build one answer.
• In a March 2026 study of 863,000 keyword SERPs, Ahrefs found that only 37.9% of pages cited in AI Overviews also ranked in the top 10 for the same query.
• Surfer's study of 173,902 URLs found that pages ranking for both the main query and its fan-out queries were 161% more likely to be cited than pages ranking for the main query alone.
• Google's AI features documentation says you do not need llms.txt, AI text files or special schema to appear in AI Overviews, but the page must be indexed and eligible for a snippet.
• Ahrefs reported that AI Overviews trigger on 57.9% of question queries and that 99.9% of AI Overview keywords are informational.
• The Search Console generative AI performance report shows AI Overview and AI Mode impressions by page, country, device and date, but no queries, clicks or position.

How does Google decide which sources to cite in AI Overviews?

Google cites the pages that keep proving useful across the related sub-searches it runs while writing the answer. There is no position one inside an AI Overview. There is a short list of supporting links, and Google picks them while it writes. Google's own documentation for site owners, AI features and your website, describes the mechanism directly: both 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.

In plain terms, when someone types a long question, Google quietly breaks it into smaller searches. Google looks at what ranks for each of those, and it cites the pages that keep showing up as useful across the set. You are not competing for one query. You are competing to be the best answer to several hidden ones.

Three facts make this concrete:

• Top 10 is no longer the gate. In a March 2026 update covering 863,000 keyword SERPs and 4 million AI Overview URLs, Ahrefs found that only 37.9% of cited pages also ranked in the top 10 for the same query. 31.2% ranked in positions 11 to 100 and 31.0% did not rank in the top 100 at all. Ahrefs notes its detection improved since its July 2025 study (which found about 76%), so the two numbers are not a perfect like-for-like comparison, and it points to fan-out as the likely reason.
• Ranking for the sub-queries is what correlates with citations. Surfer's study of 173,902 URLs across 10,000 keywords (33,000 fan-out queries extracted with Gemini) found a Spearman correlation of 0.77 between the number of fan-out queries a page ranks for and its likelihood of being cited. Pages that ranked for the main query and fan-outs were 161% more likely to be cited than pages that ranked for the main query alone.
• Google says the answer is not fixed. Google states that AI Overviews and AI Mode may use different models and techniques, so the responses and links vary, and that AI Overviews only show when Google decides they add something beyond classic results.

So the useful mental model is not "rank higher". It is "be the page Google keeps finding while it researches the answer".

How to rank in AI Overviews without a top-10 ranking

You can rank in AI Overviews without a top-10 position by ranking for the narrower sub-questions Google fans out to. Most guides on this topic, including good ones, were written when the top-10 overlap looked like three in four citations. That made "just do normal SEO and rank for the keyword" sound like enough. The newer data says roughly six in ten citations come from pages that are not in the top 10 for the query the user typed.

That does not mean classic rankings stopped mattering. Google is clear that a page must be indexed and eligible to show with a snippet to appear as a supporting link, and that normal SEO best practices still apply. The ranking that matters has moved: you still need to rank, just for the sub-questions Google spins off, which are often narrower, longer and less competitive than the head term. It is one more sign that traditional SEO alone is no longer enough.

For a SaaS company with modest domain authority, that is good news. You may never outrank a publisher with a DR in the 90s for "ai overviews seo". You can absolutely own "how long does it take to show up in AI Overviews after publishing" or "does schema markup help AI Overviews for B2B sites", and those are exactly the kinds of sub-questions Google fans out to.

Which searches trigger AI Overviews for a B2B SaaS company?

AI Overviews mostly trigger on informational, question-shaped searches, so B2B SaaS teams should aim at research-stage questions rather than "best software" head terms. In its analysis of 146 million SERPs, Ahrefs reported that AI Overviews trigger on 21% of all keywords, on 57.9% of question queries and on 46.4% of queries with seven or more words. "Why" questions triggered them 59.8% of the time, the highest of any query type Ahrefs measured, and 99.9% of AI Overview keywords were informational.

That has a direct consequence for SaaS teams. The AI Overview battleground is the research stage: problems, causes, definitions, how-tos and "is this normal" questions. It is much less about "best [category] software" head terms, where you usually see classic results, ads and listicles. For those comparison and shortlist queries, AI Mode is often where the conversation continues, and we cover that separately in our guide to Google AI Mode for SaaS.

A real example from our own Search Console: one of the queries this page earns impressions for is "why is my saas company not ranking well in google ai overviews". The query is long, starts with "why", and asks for a diagnosis. That is the shape of query where AI Overviews live, and it is the shape your content should be built to answer.

How to rank in AI Overviews: the fan-out method in five steps

The fan-out method is the process we use for B2B SaaS sites. It is built around sub-questions instead of keywords, which is the part most checklists skip. The table summarises the full workflow at a glance, from query mapping to tracking, and the five steps below go deeper on the fan-out work.

Step 1: Pick the parent question, not the keyword

Start from a question a buyer actually asks during research, written the way they would type it into Google. "Why do our SOC 2 audits take so long" is a parent question. "SOC 2 software" is a keyword. Parent questions are what trigger AI Overviews and what Google decomposes into sub-searches.

Good sources for parent questions are your sales call notes, support tickets, and the long queries already appearing in your Search Console Performance report. If your buyers ask it on calls, someone is asking Google the same thing.

Step 2: Map the fan-out for each parent question

Google does not publish its sub-queries, so you reconstruct them. Four practical ways:

• Read the AI Overview itself. Search the parent question in a clean, logged-out browser. Each paragraph or bullet in the answer usually maps to one sub-question. Write them down.
• Collect People Also Ask and follow-ups. Expand the People Also Ask box and note the AI Mode follow-up suggestions. These are close cousins of the fan-out.
• Mine long queries in Search Console. Filter the Performance report for queries of eight or more words (the regex is in the measurement section below). Long, prompt-style queries with impressions but few clicks are often fan-out traffic.
• Use a B2B sub-question checklist. For most SaaS topics the fan-out falls into the same buckets: what it is, why it happens, how to fix it, how long it takes, what it costs, how the options compare, whether it fits a specific company type, what it integrates with, and what the risks are.

You should end up with 8 to 15 sub-questions per parent question. That list is your real target set.

Step 3: Give every sub-question one owner page

Now map each sub-question to exactly one URL on your site. Some will fit as an H2 on your main guide. Some deserve their own page because they carry their own search demand. A few will already be covered by an existing post, and the fix is to sharpen that section rather than write something new. Commercial sub-questions often belong on product pages, which is why AI search engines often prefer product and landing pages for them.

Two rules keep this clean:

• One owner per sub-question. If three posts half-answer the same sub-question, none of them is the clear best answer. Consolidate, and link the others to the owner.
• Link the cluster. Google's guidance lists making content easy to find through internal links as a fundamental. Link the parent guide to each owner page and back, with descriptive anchors that match the sub-question.

This is where most SaaS blogs fail. They have one long "ultimate guide" that mentions everything once and answers nothing in a way that can be lifted cleanly. Fan-out rewards coverage across pages, not length on one page.

Step 4: Write passages that survive being quoted

An AI Overview cites a page because a specific passage supports a specific sentence in the answer. Write every section so that passage can stand on its own:

• Question-shaped heading, answer first. Put the sub-question in the H2 or H3 and answer it in the first one or two sentences below it. Context and nuance come after.
• Name the thing. Say "Google AI Overviews" or your product category by name instead of "it" or "this". A lifted paragraph loses the surrounding context.
• Use dated, sourced numbers. "In a March 2026 study of 863,000 SERPs" is quotable. "Studies show" is not.
• Use lists for steps and options. Short bullets map neatly onto the list-style answers AI Overviews often produce.
• Keep key content in HTML text. Google asks that important content be available in textual form. Answers hidden in images, tabs that load late or client-side widgets are harder to use.

Google also says it wants "unique, non-commodity content". For SaaS, the easiest way to be non-commodity is to use what only you have: product usage patterns, anonymized customer benchmarks, implementation timelines from real projects, and your team's point of view on the trade-offs. Original research works the same way, which is why we published our SaaS brand AI visibility study. Specific, sourced claims are also one of the five signals AI engines use to recommend brands.

Step 5: Show up in the fan-out results you do not own

Some sub-questions will always be answered by pages you do not control: "best of" roundups, review sites, YouTube videos and community threads. Those pages rank for fan-out queries too, so being mentioned on them puts your brand inside the answer even when your own URL is not cited. Our research on the most influential domains in AI search rankings shows which kinds of sites recur.

YouTube is the clearest example. Ahrefs found that YouTube URLs made up 5.6% of all AI Overview citations in its March 2026 dataset and that YouTube is the most cited domain in its tracking. A short product walkthrough or explainer video that answers one sub-question well can earn citations your blog cannot.

For the rest, the work is ordinary and slow: accurate listings on review sites, fair inclusion in category roundups, and useful answers in the places your buyers ask questions, such as the communities in our Reddit guide for SaaS. Describe your product consistently everywhere, so Google sees the same facts about you across sources.

Why is my SaaS company not showing up in Google AI Overviews?

A SaaS company usually misses AI Overviews because it targets queries that do not trigger them, its pages are not eligible, or its answers are buried. The full list, roughly in order of how often we see each cause, overlaps with the reasons in why your brand does not appear in AI answers:

• You are targeting queries that do not trigger AI Overviews. Commercial head terms like "[category] software" often show no AI Overview at all. Check the SERP before you plan content.
• Your page is not eligible. Google requires the page to be indexed and eligible for a snippet. A stray noindex, a nosnippet or restrictive max-snippet directive, a blocked Googlebot in robots.txt or a CDN rule, or a non-200 status all remove you from consideration. Use URL Inspection in Search Console to see what Googlebot actually received.
• Your answers are buried. The answer to the question sits in paragraph six, after a product pitch. Move it to the top of the section.
• You have one page for everything. A single guide cannot be the best answer to fifteen sub-questions. Split out the sub-questions with their own demand.
• Your content is marketing, not information. AI Overviews appear on informational queries. Pages that only describe features rarely support an informational answer.
• Nobody else mentions you. If your brand appears only on your own site, you are missing from the roundups, reviews and videos that also rank in the fan-out.
• Your facts are stale. Old pricing, discontinued features or 2024 statistics give Google a reason to prefer a fresher source.

If you want a quick read on where you stand across AI engines before you dig into Search Console, run your domain through the free AI visibility checker.

What does not matter for Google AI Overviews?

Google says llms.txt files, AI text files and special schema do not matter for AI Overviews. Google's AI features documentation states there are no additional requirements to appear in AI Overviews or AI Mode, and that you do not need to create new machine-readable files, AI text files or special markup. Google also says there is no special schema.org structured data you need to add. Structured data should match the visible content of the page, as always.

That matters because a lot of AI SEO advice now centers on llms.txt files and extra schema. Our own data shows why the confusion exists. In the Arobis AI Readiness Report, built from 1,450 site scans between June and September 2026, 45.6% of sites had an llms.txt file and 75.3% shipped JSON-LD structured data. Sites with both averaged a readiness score of 93.2, compared with 47.2 for sites with neither.

Two honest caveats. First, the sample is self-selected: these are sites whose owners chose to run an AI readiness scan, so they skew toward teams that already care about AI search. Second, the score measures readiness signals, not Google citations. The gap most likely reflects teams that take technical quality seriously across the board. The data is not evidence that llms.txt gets you into AI Overviews, and Google says it does not.

Our takeaway: keep clean structured data because it helps Google understand the page and supports rich results, and treat llms.txt as optional housekeeping for other AI tools. Neither is your AI Overview strategy. Fan-out coverage is.

Is an AI Overview citation worth anything if nobody clicks?

Yes, an AI Overview citation is worth something, but you have to measure it by visibility and demand rather than clicks. The click data is sobering. Ahrefs' February 2026 update, based on 300,000 keywords, found that an AI Overview correlated with a 58% lower click-through rate for the top-ranking page on informational queries, up from the 34.5% it reported in April 2025. Pew Research, tracking the browsing of 900 US adults in March 2025, found people clicked a traditional result in 8% of visits when an AI summary appeared versus 15% without one, and clicked a link inside the summary in just 1% of visits. Our roundup of AI search statistics collects more of these numbers.

We see the same pattern on our own site. Some of our pages lost clicks on long-tail queries during a period when AI Overviews were citing them. The citation replaced the blue-link click: the buyer got the answer, with our name attached, without visiting.

For B2B SaaS that trade is not all bad. Your buyer is doing research weeks or months before a demo request. Being the named source in the answer to "why is this happening" builds familiarity at exactly the stage where shortlists form, which is central to how AI search changes SaaS marketing. Google also says clicks from results pages with AI Overviews tend to be higher quality, with visitors more likely to spend more time on the site. That is Google's claim rather than independent data, so check it against your own engagement numbers.

The practical shift is to judge AI Overview work by visibility and downstream demand, such as branded search, direct traffic and "how did you hear about us" answers, and not by organic clicks alone.

How do you measure AI Overview visibility in Search Console?

You measure AI Overview visibility with the Search Console generative AI performance report, a long-query regex filter in the Performance report, and logged-out spot checks. Here is how to use what exists:

• Open the generative AI performance report. The report shows how often your URLs appeared in AI Overviews and AI Mode, broken down by page, country, device and date. The report does not show queries, clicks or position, so it tells you whether you are visible, not what you earned.
• Compare AI impressions with total impressions. The same impressions are still included in the main Performance report. Divide one by the other to get your AI share. Our own recent figures are in the dated section below.
• Isolate prompt-style queries. In the Performance report, add a query filter, choose Custom (regex) and use ^(\S+\s+){7,}\S+$ to show queries of eight words or more. These long, prompt-style queries are the kind AI Mode generates as fan-out sub-searches, and they cannot be clicked in the classic sense. Expect high impressions and very low CTR here, and do not treat that as failure.
• Watch pages, not just totals. Pages with rising AI impressions and falling clicks are being cited. Pages with falling impressions on both sides are losing the fan-out, and those are the ones to refresh first.
• Use the 28-day or 3-month view. Daily data is noisy, and AI answers change from one refresh to the next. Trends over weeks are what count.
• Spot-check the SERP by hand. Search your parent questions in a logged-out browser set to your target country and note which of your pages are cited. Personalized results skew what you see when you are signed in.

To connect this to visits and pipeline in GA4, see our guide on how to track AI traffic to your website. If you want a done-for-you version, with the fan-out mapped for your category, the citation gaps against competitors and a prioritized fix list, that is what our AI visibility audit covers, and our SaaS AI search case study shows the kind of work involved.

Should you optimize for Google AI Overviews separately from other AI search engines?

Partly. The core work carries over, but each engine picks sources differently.

AI Overviews appear inside normal Google results, mostly on informational questions, with a short answer and a handful of links. AI Mode is a separate, conversational Google experience for deeper comparisons and follow-ups, and Google says it may use different models and show different links. ChatGPT, Claude and Gemini run their own retrieval and weigh sources in their own ways, as we explain in ChatGPT vs Claude vs Gemini search behavior and how Perplexity, ChatGPT, Claude and Gemini choose which brands to mention.

The fan-out idea carries across all of them: these systems break questions into sub-questions and assemble answers from sources that cover the pieces well. The details differ enough that each engine deserves its own plan. Our AI visibility guide covers the shared basics, and the engine guides cover how to rank in ChatGPT, how to rank in Claude and how to rank in Google Gemini.

What has changed in AI Overviews recently? (Checked October 2026)

AI Overview reporting and click data have changed quickly, so the time-sensitive facts sit in this one section. We re-check it every month.

• June 3, 2026: Google launched the Search generative AI performance reports in Search Console.
• August 31, 2026: the generative AI performance reports became available to all websites.
• Our own numbers, September 2026: for arobis.ai, the report attributed about 19% of impressions to AI features, and about 35% of impressions came from long, prompt-style queries of the kind AI Mode generates as fan-out sub-searches.
• Study dates: the Ahrefs citation overlap figure (37.9%) is from March 2026, the Ahrefs click study (58% lower CTR) from February 2026, and the Pew click data from March 2025.

Where to start this week

Start with one parent question your buyers ask. Search it logged out, and write down every sub-question the AI Overview answers. Check which of those your site answers clearly, which it answers badly, and which it does not answer at all. That single page of notes is a better AI Overview plan than any generic checklist, because it is built from the fan-out Google is actually running for your market.

Frequently asked questions

How is ranking in Google AI Overviews different from ranking in organic search?

Organic ranking is a position for one query. An AI Overview citation is a pick Google makes while answering several related sub-queries. Ahrefs' March 2026 study found only 37.9% of cited pages ranked in the top 10 for the query the user typed, so covering the sub-questions matters more than one head-term position.

Do I need high domain authority to appear in Google AI Overviews?

No. About 62% of AI Overview citations in Ahrefs' March 2026 data came from pages outside the top 10, and the narrower sub-questions Google fans out to are often far less competitive than the head term. Your page must still be indexed by Google and eligible to show with a snippet.

What is query fan-out in Google AI Overviews?

Query fan-out is Google's technique of issuing multiple related searches across subtopics to build one AI response. Google confirms both AI Overviews and AI Mode may use it. The practical effect is that pages answering the sub-questions behind a query can be cited even when they do not rank for the main query.

Do I need llms.txt or special schema to get into AI Overviews?

No. Google's AI features documentation says you do not need new machine-readable files, AI text files or special schema.org markup to appear in AI Overviews or AI Mode. Standard structured data that matches the visible content of the page is still good practice, and the page must be indexed and snippet-eligible.

How do I know if my brand is appearing in Google AI Overviews?

Open the generative AI performance report in Google Search Console, which shows how often your pages appeared in AI Overviews and AI Mode by page, country, device and date. The report does not show queries, clicks or position, so add logged-out spot checks of your key buyer questions.

How long does it take to appear in AI Overviews?

There is no fixed timeline. A page first has to be crawled and indexed, and Google says recrawling can take anywhere from several days to several months. After that, citations depend on how well the page answers the sub-questions, and they can change between refreshes, so judge progress over 28 days or more.

Is AI Overview optimization the same as GEO or AEO?

AI Overview optimization is one Google-specific part of generative engine optimization, which covers earning citations in all AI-generated answers. It also relies on answer engine optimization, the practice of writing direct, self-contained answers that Google AI Overviews can quote.