How to Rank in AI Overviews: The 2026 Guide

To rank in AI Overviews, your page must be indexed by Google, eligible to show with a snippet, and included in Search generative AI features in Search Console. Beyond that, Google says there are no special AI optimisations. What actually wins citations in 2026 is answer first content, first hand expertise, coverage of the sub questions Google generates through query fan out, and consistent brand presence across the sources Google already trusts.

That is the compressed version. The rest of this guide covers the mechanics, the current data, the tactics that work, and the popular advice that Google has explicitly told you to ignore.

Key numbers at a glance

What AI Overviews are in 2026

AI Overviews are the AI generated answer blocks that appear at the top of Google's results page, with clickable links to the pages that support the response. They launched in the US in May 2024 and now run globally.

Two things changed materially in 2026 and both affect how you optimise:

The model changed twice. Google made Gemini 3 the default model for AI Overviews globally on 27 January 2026, bringing the same reasoning that already powered AI Mode into the results page. At I/O on 19 May 2026, Google upgraded AI Mode to Gemini 3.5 Flash globally and shipped what it called the biggest upgrade to the Search box in over 25 years. Model changes shift citation patterns, which is a large part of why the 2025 playbooks stopped working.

The two surfaces converged. You can now ask a follow up question directly from an AI Overview and carry the context into AI Mode. Google was clear at I/O that classic results have not gone away, stating that people continue to get a range of results from Search. But the AI answer is increasingly the first thing anyone sees.

For SEO purposes, AI Overviews and AI Mode share the same retrieval machinery, so most of what follows applies to both. They do not, however, cite the same pages. Ahrefs found only 13.7% URL overlap between the two surfaces in December 2025, and SE Ranking put URL overlap at 10.7% in August 2025. Treat them as related but separate visibility layers.

How Google actually selects sources for AI Overviews

Google published a dedicated guide, Optimizing your website for generative AI features on Google Search, in May 2026 and last updated it on 10 July 2026. It is the closest thing to an official answer and it names two mechanisms.

1. Retrieval augmented generation (grounding)

Google's AI features do not invent answers from model memory. They retrieve relevant, current pages from the Search index using core Search ranking systems, then generate a response grounded in what those pages say, with clickable links back to them.

The practical consequence: if you are not indexed and retrievable, you cannot be cited. Ranking systems remain the gatekeeper.

2. Query fan out

This is the mechanism most guides underweight. Google splits the original query into a set of concurrent related sub queries and fetches results for each one. Google's own example: for "how to fix a lawn that's full of weeds", fan out queries might include best herbicides for lawns, removing weeds without chemicals, and preventing weeds in lawn.

The AI Overview then draws from the pages that perform well across those sub queries. This is why a page can be cited for a term it does not rank for, and why ranking first for the headline term no longer guarantees inclusion.

3. Eligibility requirements

Google's AI features and your website documentation states the technical bar plainly: to appear as a supporting link, a page must be indexed and eligible to be shown in Google Search with a snippet. There are no additional technical requirements.

Since June 2026 there is one more condition. Your site must be included in Search generative AI features in Search Console. This is on by default, but it is now a setting you can accidentally switch off. More on that below.

The biggest shift of 2026: rankings stopped predicting citations

If you take one thing from this guide, take this.

In July 2025, Ahrefs analysed 1.9 million citations and found 76% of cited pages ranked in the organic top 10 for the same query. The consensus at the time was that AI Overview optimisation and normal SEO were the same job.

Ahrefs re ran the study across 863,000 keywords and 4 million AI Overview URLs and published the update in February 2026. The figure had fallen to 38%. The remaining citations split almost evenly between positions 11 to 100 (31.2%) and pages that do not appear in the top 100 at all (31.0%). Filtering to organic results only produced similar numbers: 37% in the top 10, 36% outside the top 100.

A separate BrightEdge analysis published on 12 February 2026 put top 10 overlap even lower, at roughly 17%, using different methodology and a different dataset.

Two honest caveats, because this stat gets misquoted constantly:

1. Ahrefs itself notes its parsing methodology improved between the two studies, so it can now detect citations it previously missed. The two datasets are not directly comparable and some of the drop reflects better detection rather than pure behaviour change.
2. Prevalence and overlap figures vary wildly between trackers because keyword sets, geographies and detection methods differ. Treat every number as a range, not a constant.

Even with those caveats, the direction is unambiguous, and Google's own fan out documentation explains the mechanism. Ranking well still helps. It is no longer sufficient, and it is no longer necessary.

One more finding worth acting on: among Ahrefs' cited pages that did not rank in Google's top 100 for the keyword, 18.2% were YouTube URLs. YouTube accounted for 5.6% of all AI Overview citations in the dataset and is now the most cited domain in AI Overviews according to Ahrefs Brand Radar.

How to rank in AI Overviews: the 12 step playbook

1. Confirm you are actually eligible

Before optimising anything, check three things:

• The page is indexed and returns a normal snippet in Search
• Googlebot is not blocked in robots.txt, by your CDN, or by your hosting infrastructure
• In Search Console, go to Settings > Search generative AI and confirm the property is set to Include

That third one is new. Google began rolling out the Search generative AI control on 3 June 2026 and started honouring it on 17 June 2026. It is a property level toggle with Include, Exclude and Inherit from parent options. Excluding removes your content and links from AI Overviews, AI Mode and generative features in Discover, and stops your crawled content being used to ground responses. It does not affect classic organic rankings. It rolled out to UK properties first following a Competition and Markets Authority conduct requirement, then expanded further.

2. Answer the question in the first two or three sentences

Grounding systems extract passages. A page that buries its answer under 400 words of preamble gives the model nothing clean to lift.

Put a direct, self contained answer to the page's core question at the top, then expand. Write each key passage so it makes sense in isolation: avoid opening sentences with "this approach" or "these factors" that only resolve if you read the paragraph above.

Note the nuance here. Google explicitly says you do not need to chunk your content into tiny pieces, and there is no ideal page length. Answer first structure is good writing for humans, which is the actual bar. Mechanical chunking is not.

3. Build for fan out, not for one keyword

Since Google is fetching results for sub queries you never targeted, coverage across the facets of a topic beats depth on one exact phrase.

Practical method:

• Take your target query and list the 8 to 15 questions a searcher would logically ask next
• Check which of those you already cover, on this page or elsewhere on the site
• Fill the gaps with genuinely useful sections or supporting pages, and interlink them

The caution: Google's scaled content abuse policy explicitly covers creating separate pages for every possible query variation, including fan out queries, primarily to manipulate rankings. Cover the topic properly. Do not spin out 200 near duplicate pages.

4. Publish non commodity content

Google's guide draws a sharp line and gives real examples. Commodity content, such as "7 Tips for First-Time Homebuyers", is common knowledge that could have come from anyone. Non commodity content, such as "Why We Waived the Inspection & Saved Money: A Look Inside the Sewer Line", carries an expert or experienced take that goes beyond the ordinary.

Google says creating content people find unique, compelling and useful will influence your presence in generative AI search more than anything else in its guide. That is not a platitude. Because AI systems review many sources, a distinctive viewpoint is what makes yours worth citing rather than the fifth restatement of the same facts.

Things that qualify:

• Original data, surveys or benchmarks you ran
• First hand reviews and tests, with photos and specifics
• Case studies with real numbers and named outcomes
• Expert commentary that takes a position other sources do not

5. Put first hand experience on the page and put your name on it

On 6 May 2026, Google shipped five link changes to AI Overviews and AI Mode, announced by Hema Budaraju, VP of Product Management for Search:

• Suggested "explore new angles" links to in depth articles at the end of many responses
• Highlighting of links from your own news subscriptions
• Previews of perspectives from public discussions, social media and other first hand sources, with the creator's name, handle or community name shown
• More links placed inline, directly next to the text they support
• Hover previews of the source site on desktop

The third and fourth points matter most. First hand experience is no longer just an E-E-A-T signal. It is now a labelled, clickable surface inside the AI answer. Named authors with real credentials, visible bylines, author pages, and clearly attributable expertise are now directly monetisable in AI search.

On 27 May 2026 Google went further, bringing Preferred Sources to AI Overviews and AI Mode and adding a "Highly Cited" badge that flags articles many other stories reference. Original reporting now has a visible marker attached to it.

6. Get video into your content mix

The YouTube data is too strong to ignore. It is the most cited domain in AI Overviews, it supplied 18.2% of citations from pages ranking outside the top 100, and Google's own guidance says its generative AI features can bring in relevant images and video, which means more opportunities to appear beyond web page links.

Practical steps: add genuinely useful video to key pages, publish it on YouTube with accurate titles, descriptions and full transcripts, and follow Google's video SEO documentation. Google is explicit that if you already follow its image and video SEO best practices, you are already optimising for generative AI search.

7. Show up where AI already looks

AI answers increasingly cite third party sources rather than brand sites. Across the major engines, Reddit, Wikipedia, YouTube, LinkedIn and Forbes dominate citation share, and Google AI Overviews cite Reddit and YouTube heavily.

The right response is participation, not manipulation. Contribute real expertise in the communities relevant to your category, under a real identity. Earn coverage in the review sites, roundups and industry publications that already rank for your category terms.

Google is blunt about the wrong version of this: seeking inauthentic mentions across the web is not as helpful as it might seem, because core ranking systems focus on high quality content while other systems block spam, and generative AI features depend on both. Paid mention schemes and coordinated posting are a fast route to a spam action.

8. Keep your facts consistent and verifiable

Grounding systems cross reference. If your stated pricing, founding date, headquarters, product specs or headline statistics conflict with what authoritative sources say elsewhere, you become a risky citation.

Audit for consistency across your site, your Google Business Profile, Merchant Center, Wikipedia if you have an entry, LinkedIn, and major directories. Cite your own sources with dates and link out to primary research. Update stale figures rather than leaving 2023 numbers on a page.

9. Fix crawling, rendering and page experience

Google's technical guidance for generative AI search is the same technical guidance as always:

• Allow crawling in robots.txt and at CDN and hosting level
• Make important content available in textual form, not locked behind interactions
• Follow JavaScript SEO best practices if your site is framework based, since Google can process JavaScript content but it adds complexity
• Make content findable through internal links
• Reduce duplicate content so crawl budget is not wasted
• Provide a good page experience across devices

Google says perfect semantic HTML is not required, though it helps screen readers and, increasingly, browser agents parsing your accessibility tree.

10. Use structured data, but for the right reason

Google states directly that structured data is not required for generative AI search and there is no special schema.org markup you need to add. Any blog claiming a specific multiplier for FAQPage schema in AI Overviews is quoting a number Google does not support.

Keep using it anyway. It qualifies you for rich results in classic Search, it disambiguates entities, and Google asks that your structured data matches the visible text on the page. Organisation, Article, Product and LocalBusiness markup remain worth having. Just do not expect it to be the lever that gets you cited.

11. Sort out local and commerce data

Generative AI responses can include product listings, product information and local business details. Keep Merchant Center feeds and your Google Business Profile accurate and current. For retailers, this is often higher leverage than any content work.

12. Prepare for agents

Google's guide now includes a section on agentic experiences. Browser agents access sites by analysing screenshots, inspecting the DOM and interpreting the accessibility tree. Google points site owners to the agent friendly website best practices guide and flags emerging protocols such as the Universal Commerce Protocol.

This is a "if you have extra time" item for most sites in 2026. For ecommerce, it is worth starting now.

How to measure whether you rank in AI Overviews

On 3 June 2026, Google launched Search Generative AI performance reports in Search Console, announced by Hillel Maoz, Search Ecosystem Engineering Manager, and Moshe Samet, Product Manager Lead for Search Console. This is the first official measurement of AI visibility.

What it gives you:

• Impressions within generative AI features on Search (AI Overviews and AI Mode) and in Discover
• Breakdowns by page, country, device and date, with hourly through monthly granularity
• A view separate from, but also included in, the overall Performance report

What it does not give you yet:

• Clicks, CTR or query data
• Historical data before roughly 18 May 2026
• Access for every property, since it rolled out to a subset of sites starting in the UK before expanding

Pair it with GA4 for referral traffic from other AI platforms, since Search Console only measures Google Search. Traffic from ChatGPT, Perplexity, Copilot and Claude arrives as ordinary referral traffic and has to be tracked separately.

One warning straight from Google's guide: be wary of third party tools that promise ranking success or claim to use internal Google metrics, because no third party tool has access to Google's internal ranking or AI systems. Rank tracking tools for AI Overviews are useful for directional monitoring. They are not ground truth.

What you can safely ignore

Google's guide includes a mythbusting section. These are the things it says you can ignore for Google Search:

Google also has dedicated guidance on evaluating third party SEO advice, which is worth reading before you buy an "AEO" or "GEO" package. Google's own framing is that from Search's perspective, optimising for generative AI search is optimising for the search experience, and therefore still SEO.

Should you opt out of AI Overviews?

For almost every commercial site, no.

The case for staying in: AI Overviews reach 2.5 billion monthly users, AI Mode passed 1 billion, and Seer Interactive found that being cited inside an AI Overview delivered around 35% more organic clicks than not being cited on the same results page. Google also reports that clicks from results pages with AI Overviews tend to be higher quality, with users spending more time on site.

The case some publishers make for opting out: when an AI Overview is present, Pew Research found users clicked an organic result on 8% of visits compared with 15% without one, and clicked a link inside the summary itself only about 1% of the time. Seer measured organic CTR falling from 1.76% to 0.61% on AI Overview queries. If your entire model is ad impressions on informational content, the maths can look different.

Opting out is a blunt instrument. It removes your links, not just your text, from AI answers, and forfeits all impressions and traffic from those surfaces. Reporting suggests page level controls are planned so publishers can exclude specific URLs rather than whole properties, which would make selective opt out far more practical. Until that lands, treat exclusion as a decision for a small number of publishers with a specific commercial reason, not a default.

Frequently asked questions

Is ranking in the top 10 enough to appear in AI Overviews? No. Ahrefs found 38% of cited pages ranked in the top 10 in February 2026, down from 76% in July 2025, and BrightEdge put the figure closer to 17%. Around a third of citations come from pages that do not rank in the top 100 at all. Ranking helps but does not guarantee citation.

How long does it take to appear in an AI Overview? There is no fixed timeline. Your page has to be crawled, indexed and then selected during grounding, and Google warns that crawling can take anywhere from several days to several months. Requesting reindexing after a significant update speeds up the first step only.

Do AI Overviews use the same sources as AI Mode? Largely no. Ahrefs found only 13.7% URL overlap between AI Overviews and AI Mode in December 2025. Google confirms the two features may use different models and techniques, so the links they show vary.

Does schema markup help you rank in AI Overviews? Not directly. Google states structured data is not required for generative AI search and there is no special schema you need. It remains worthwhile for rich results and entity clarity in classic Search.

Can you track AI Overview impressions in Search Console? Yes, since 3 June 2026. The Search Generative AI performance report shows impressions by page, country, device and date. Clicks, CTR and query data are not included yet, and access is still rolling out.

Does appearing in an AI Overview actually send traffic? Less than a top organic position used to, but more than being absent. Seer Interactive measured roughly 35% more organic clicks for cited sites versus uncited sites on the same query, and Google's May 2026 link updates put more citations inline next to the text they support, which should improve click through over time.

What is query fan out? Query fan out is Google generating a set of concurrent, related sub queries from the user's original question and fetching additional search results for each. The AI Overview is assembled from pages that answer those sub queries, which is why coverage across a topic beats optimising for a single phrase.

The bottom line

The honest summary is that there is no AI Overview ranking hack, and Google has now said so in writing three separate times. What has changed is where the leverage sits.

The leverage used to be in ranking position. It has moved to topical coverage across the sub questions Google generates, first hand expertise that is attributable to a named person, presence on the platforms AI already trusts, and technical hygiene that keeps you retrievable. Ranking is the entry ticket. Being the most useful, most distinctive, most quotable source on a specific question is what actually gets you cited.