Back to Orbit

Back to Orbit

SEO & AEO

How to rank in AI search results: what the research actually shows

We read the studies behind the AI SEO hype: Google's own guide, Pew, Ahrefs and the Princeton GEO paper. Here is what actually gets a page cited.

Arshia Nemati
Arshia Nemati · 11 min read

Key takeaways

  • Google says optimizing for its AI features is still SEO: indexed, crawlable, genuinely useful pages come first.
  • Ranking in the top 10 helps, but it is no longer enough. Ahrefs now finds only about 38% of AI Overview citations rank on page one.
  • Clicks are scarcer when an AI summary appears (8% versus 15% in Pew's data), so being the cited source matters more than ever.
  • The Princeton GEO study found statistics, quotations and named sources lifted visibility by up to 40%. Keyword stuffing did not help.
  • Cover a topic from several angles and formats, including video, because AI answers pull from many related searches at once.

To rank in AI search results, do the SEO basics well and then make your pages worth quoting. Google's own guide says its AI features run on the same index and quality systems as normal search, so a page that is not indexed and crawlable cannot be cited. On top of that, the research points the same way: pages that answer clearly, back claims with real numbers and named sources, and cover a topic from several angles get picked far more often. There is no secret file or trick that skips this.

There is a lot of loud advice about AI SEO right now, and most of it is opinion. So we went to the primary sources: Google's documentation, a browsing study from Pew Research Center, two large citation studies from Ahrefs, and the peer reviewed GEO paper from Princeton and IIT Delhi. Below is what each one actually found, where they disagree, and the plan we use on our own site and for clients.

What is AI SEO, in plain words?

AI SEO is the work of getting your pages chosen as sources inside AI generated answers: Google AI Overviews and AI Mode, ChatGPT search, Perplexity, Gemini and Copilot. Classic SEO aims for a position in a list of links. AI SEO aims for a citation inside the answer itself, which is often the only thing the searcher reads. If you want the full map of the terms, our guide to SEO vs AEO vs GEO explains how they fit together.

What does Google itself say about ranking in AI Overviews and AI Mode?

In May 2026 Google published its first official guide on optimizing for generative AI features in Search. Its message is blunt: from Google's point of view, optimizing for generative AI search is still SEO. The features are rooted in the same ranking and quality systems as the rest of Search, and Google says unique, compelling and useful content will matter more over time than any other suggestion in the guide.

  • The technical gate is simple. A page must be indexed and eligible to show with a snippet. Google also asks that your site is included in generative AI features in Search Console.
  • Crawling matters. Do not block Googlebot in robots.txt, at your firewall or in your CDN settings.
  • Images and video count. AI features can show images and video, so image and video SEO are part of AI SEO.
  • Some popular tactics are unnecessary for Google. The guide lists ideas you can ignore for Google Search, and a special llms.txt file is not needed to appear in AI Overviews or AI Mode. Other AI engines may treat such files differently.

Google also explains that AI Overviews and AI Mode may use a technique it calls query fan out: the system runs several related searches across subtopics and sources, then builds one answer. That single detail explains a lot of what the independent studies found.

Do you need to rank on page one to be cited by AI?

It helps, but less than it used to. In July 2025 Ahrefs studied 1.9 million citations from 1 million AI Overviews and found that 76% of cited pages also ranked in the top 10 for the same search. In March 2026 Ahrefs repeated the study with better tracking across 863,000 keywords and found only about 38% did. The rest were split almost evenly between pages ranking 11 to 100 and pages outside the top 100 entirely.

Where pages cited in Google AI Overviews rank in normal search
StudyCited pages in the top 10Positions 11 to 100Outside the top 100
Ahrefs, July 2025 (1.9M citations)76%About 10%About 14%
Ahrefs, March 2026 (863K keywords, 4M URLs)About 38%About 31%About 31%
Ahrefs changed its tracking method between the two studies, so the numbers are not a perfect like for like comparison. The direction is still clear: page one is no longer the whole story.

Why would Google cite a page that does not rank? Query fan out. The answer is built from many related searches, so a page that ranks well for a narrower follow up question can be cited for the broad one. Ahrefs also found YouTube is now the most cited domain in AI Overviews, and that 18% of the cited pages that did not rank in the top 100 were YouTube videos.

38%of pages cited in Google AI Overviews also rank in the top 10 for the same search (Ahrefs, March 2026)

Does AI search still send traffic to websites?

Less of it, which raises the value of being the source that gets named. Pew Research Center tracked the real browsing of 900 adults in the United States in March 2025. When a Google search page showed an AI summary, people clicked a normal result in 8% of visits. Without a summary, they clicked in 15%. Only 1% of visits included a click on a link inside the summary itself.

Pew also found AI summaries appeared on about 18% of searches, and far more often on long questions: just 8% of one or two word searches triggered one, against 53% of searches with ten or more words. Google disputed the method and says clicks from AI results tend to be higher quality. Both can be true: fewer clicks, but the people who click are ready to act. Either way, if your brand is named in the answer, you win attention even when nobody clicks.

8% vs 15%share of Google visits with a click on a normal result, with and without an AI summary on the page (Pew Research Center, 2025)

What makes AI engines choose one source over another?

The best evidence comes from GEO: Generative Engine Optimization, a peer reviewed paper from researchers at Princeton, IIT Delhi and partners, presented at KDD 2024. They rewrote source pages in different ways and measured how visible each version became inside AI generated answers, across thousands of queries, then confirmed the results on Perplexity.

What the Princeton GEO study found
Change to the pageEffect on visibility in AI answers
Adding relevant statisticsStrong lift, among the best single changes
Adding quotations from credible sourcesStrong lift, up to about 40% in the best cases
Citing sources clearlyStrong lift, and the biggest gains for lower ranked pages
Clearer, more fluent writingModerate lift
Keyword stuffingNo help, and worse than doing nothing on Perplexity
Results varied by topic: the paper stresses that what works for history questions differs from what works for opinion or factual ones. Treat these as tested tendencies, not guarantees.

Two findings stand out for small and growing businesses. First, the old habit of repeating keywords is dead weight in AI answers. Second, the biggest gains went to pages that did not already rank at the top. That is good news if you are the challenger: evidence and clarity can beat size.

Including citations, quotations from relevant sources, and statistics can significantly boost source visibility.

Aggarwal et al., GEO: Generative Engine Optimization (KDD 2024)

How do you rank in AI search, step by step?

We turned the research into a five rung ladder. Each rung depends on the one below it, so work from the bottom up.

  1. Findable. Make sure Google, Bing and the AI search crawlers can reach every important page. Check robots.txt, your firewall and CDN bot settings, submit a sitemap, and confirm pages are indexed in Search Console and Bing Webmaster Tools.
  2. Readable. Put the answer in the first two or three sentences, use question style headings, and keep pages fast on mobile. Use real HTML tables for comparisons and prices so they can be quoted cleanly.
  3. Quotable. Back every important claim with a number, a named source or a short quotation from someone credible. Write sentences that still make sense when lifted out on their own.
  4. Complete. Cover the follow up questions a buyer would ask next, in linked articles or sections, so you show up across the fan out searches. Add a short video where a demo helps, since YouTube is heavily cited.
  5. Trusted. Keep your business facts identical everywhere (site, Google Business Profile, directories, social profiles) and earn mentions and reviews on other sites. AI engines trust what others say about you more than what you say about yourself.

Is AI SEO different from traditional SEO?

Mostly it is traditional SEO done properly, with a sharper focus on evidence and clarity. The foundations are identical: crawlable, indexed, fast, helpful pages. What changes is the reward. A traditional ranking pays off in clicks. An AI citation pays off in being named and trusted inside the answer, often before anyone visits a website.

Traditional SEO and AI SEO side by side
Traditional SEOAI SEO
GoalA high position in the list of linksBeing cited or named inside the answer
What winsRelevance, links, page experienceThe same, plus clear answers, statistics, quotations and named sources
Unit of competitionOne keyword, one pageA topic, across many related searches
FormatsMostly web pagesWeb pages, video, images, reviews and listings
How to measureRankings, clicks, conversionsCitations and mentions in AI answers, AI referral visits, branded searches

Will AI replace SEO?

No. AI changes where the results show up, not the need to be found. Every major AI search product still has to retrieve pages from the web before it writes an answer, and Google says its AI features rely on its normal Search index. The businesses that lose are the ones that stop maintaining the basics because they assume the rules were rewritten. The rules mostly stayed. The bar for quality went up.

How do you measure your AI search visibility?

  • Search Console. Visits from AI Overviews and AI Mode are counted in your normal Search Console performance data, and Google's guide points to a generative AI performance report as well.
  • Analytics referrals. Watch for visits from chatgpt.com, perplexity.ai, gemini.google.com and copilot.microsoft.com.
  • A monthly prompt check. Run the same set of customer questions through the main assistants each month and log whether you are cited or named.
  • Ask every new lead. "How did you find us?" now often has the answer "ChatGPT told me".

We build every site with this ladder in mind from the first line of code: open crawlers, full structured data, answer first pages, quotable tables and fast mobile performance. Then we grow the evidence around your brand month by month. If you want the practical side of being named by assistants, read how to get recommended by ChatGPT. If you are weighing budgets, see what this typically costs in our 2026 website cost guide.

Frequently asked

Make sure your pages are indexed and crawlable, put clear answers at the top, support claims with statistics, quotations and named sources, cover the follow up questions in related pages, and keep your business facts consistent across the web. Google says this is still SEO, and independent studies agree on what earns citations.

AI SEO is the work of getting your pages chosen as sources inside AI generated answers, such as Google AI Overviews, AI Mode, ChatGPT search and Perplexity. It builds on traditional SEO and adds a focus on quotable, well evidenced content.

No. Google's guide says a special file like llms.txt is not needed for Google Search or its AI features. Other AI engines may use such files differently, so it does no harm, but it is not a shortcut.

It helps, but it is not required. Ahrefs found in March 2026 that only about 38% of cited pages also ranked in the top 10 for the same search, because AI Overviews draw on many related searches at once.

No. In the Princeton GEO study, stuffing keywords did not improve visibility in AI answers and performed worse than the original page on Perplexity.

Fixing crawl and indexing problems can show results within weeks, because AI search tools read the live web. Earning citations for competitive questions usually takes several months of steady, well evidenced content and mentions.

No. AI search still retrieves pages from the web before it writes an answer, and Google's AI features use its normal Search index. SEO remains the foundation; the quality bar is simply higher.

Sources

  1. Google Search Central: Optimizing your website for generative AI features on Google Search (May 2026)
  2. Google Search Central Blog: A new resource for optimizing for generative AI in Google Search
  3. Google Search Central: AI features and your website
  4. Pew Research Center: Google users are less likely to click on links when an AI summary appears (July 2025)
  5. Ahrefs: 76% of AI Overview citations pull from the top 10 (July 2025)
  6. Ahrefs: Update, 38% of AI Overview citations pull from the top 10 (March 2026)
  7. Aggarwal et al.: GEO, Generative Engine Optimization (KDD 2024, arXiv 2311.09735)
TopicsAI SEOAI searchAI OverviewsGEOSEO
Arshia Nemati

Written by

Arshia Nemati

Founder & CEO at Webco

Meet the team

Share this article

Follow the studio on InstagramBehind the scenes, launches and quick SEO and AI search tips from the studio.@webcostudiosFollow

Work with Webco

Want this done for your business?

We design, build and grow websites with SEO, AEO and GEO built in from the first line of code.

Start a project