If you want to know how to get cited in Google AI Overviews and AI Mode, the short answer is SEO done well. Google’s own documentation says there are “no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary.” A page has to be indexed and eligible to show with a snippet, and then the pages that get picked tend to be the ones that add something the rest of the web doesn’t.
That’s less of a shortcut than people hope for, and more useful. Below: how both features choose their sources, what Google’s May 2026 guide actually asks for, what independent studies add, the myths you can skip, a checklist, and how to measure progress.
How AI Overviews and AI Mode choose their sources
Both features are built on top of Google Search, not next to it. In its guide to optimizing for generative AI features, Google explains that they “are rooted in our core Search ranking and quality systems” and rely on two techniques:
- Retrieval-augmented generation (grounding): Google’s ranking systems retrieve relevant, up-to-date pages from the Search index, and the response shows links to the pages that support it.
- Query fan-out: the model runs “a set of concurrent, related queries” to fetch more results. Google’s example: for “how to fix a lawn that’s full of weeds,” fan-out queries might include “best herbicides for lawns” and “remove weeds without chemicals.”
So a citation isn’t only decided by the query someone typed. It can come from any of the related searches behind it.
One question, many searches
Question
how to fix a lawn that’s full of weeds
Fan-out searches
best herbicides for lawns
remove weeds without chemicals
…and more related searches
Answer + links
Sources
AI Overviews vs AI Mode
AI Overviews
AI Overview
Above some results, only when Google decides it adds value.
AI Mode
Compare CRMs for a 5-person sales team
Its own conversation, for questions that need comparing and reasoning.
| AI Overviews | AI Mode | |
|---|---|---|
| Where it appears | Above some regular search results | Its own conversational experience |
| When | Only when Google’s systems decide it’s “additive to classic Search,” so they “often don’t trigger” | Questions that need “further exploration, reasoning, or complex comparisons” |
| Sources | Supporting links in the response | Supporting links in the response |
Google also notes that the two “may use different models and techniques, so the set of responses and links they show will vary.” Being cited in one doesn’t mean you’ll be cited in the other, so check both.
Step one: make sure you’re eligible
Before content, check that nothing is quietly keeping you out.
Indexed, with a snippet
No accidental noindex or nosnippet.
Included in Search Console
Settings › Search generative AI › Include. It’s the default.
Crawlable everywhere
Allowed in robots.txt and by your CDN or firewall.
Important content in text
Not only in images, video or scripts.
Google-Extended is a separate decision. Blocking it affects Gemini training and Gemini Apps, not your inclusion in Google Search.
- Indexed and snippet-eligible. To be a supporting link, Google says a page “must be indexed and eligible to be shown in Google Search with a snippet.” A
nosnippetdirective or anoindextag removes that eligibility. - Included in the Search generative AI setting. Since August 31, 2026, every site has a “Search generative AI” control in Search Console (Settings > Search generative AI). “Include” is the default. If someone switched it to “Exclude,” your links and content won’t appear in AI Overviews, AI Mode or Discover’s generative AI features, and Google adds that excluding is not used as a ranking signal elsewhere in Search.
- Crawlable everywhere. Google lists the usual basics: crawling allowed in robots.txt and “by any CDN or hosting infrastructure,” internal links to your important pages, a good page experience, and important content “available in textual form.” Check your firewall or bot protection too, not just robots.txt.
- Google-Extended is a separate decision. Blocking it controls whether your content trains Gemini models and grounds answers in Gemini Apps. Google says it “does not impact a site’s inclusion in Google Search.”
None of this guarantees anything. Google is explicit that meeting every requirement “doesn’t mean that Google will crawl, index, or serve its content.” It just means you’re in the running.
What earns the citation
Publish non-commodity content
This is the part of Google’s guide worth reading twice. It says that creating content people find “unique, compelling, and useful will likely influence your website’s presence in generative AI search in the long run more than any of the other suggestions in this guide.”
Google contrasts two kinds of content. “Commodity” content, like “7 Tips for First-Time Homebuyers,” is “often based on common knowledge” and “typically adds little unique insight.” Non-commodity content provides “unique expert or experienced takes that go beyond common knowledge,” like Google’s own example: “Why We Waived the Inspection & Saved Money: A Look Inside the Sewer Line.” Its advice: “Don’t just recycle what others on the internet have already said, or could easily be produced by a generative AI model.”
For a B2B brand, that usually means:
- Your own data. Benchmarks, survey results, aggregated (and anonymized) product data.
- First-hand experience. How you implemented something, what went wrong, what you’d do differently.
- Facts only you can state. Pricing model, specifications, integrations, policies and company details, in plain text.
- A clear point of view. Where you disagree with common advice, and why.
Write for people, organized for scanning
Google asks for content that’s “well written and easy to follow,” organized “by paragraphs and sections, along with headings that provide a clear structure.” It also says you don’t need to rewrite content in a special way for AI, because its systems “understand synonyms and general meanings.” Answer the question near the top of the section, use the questions buyers ask as headings, and keep the page current.
Add images and video where they help
Google’s AI features “can bring in relevant images and video,” which means more ways for your site to appear beyond a text link. If you already follow image and video SEO best practices, Google says “you’re already optimizing for generative AI search.”
Keep product and local data up to date
If you sell products or have locations, AI responses can include product listings and local business information. Google points to Merchant Center feeds and Business Profiles as the way to make that data available.
Be talked about, honestly
Google’s AI features “can show what’s being said about products and services across the web, including in blogs, videos, and forum discussions.” Independent data points the same way (see below). What doesn’t work is manufacturing it: Google says “seeking inauthentic ‘mentions’ across the web isn’t as helpful as it might seem.”
What independent data adds
Google tells you what it wants. Third-party studies show patterns from the outside. They are useful context, not rules, and the three below measure different things, so read each on its own terms.
Ranking in the top 10 helps, but it isn’t enough. In a March 2026 update, Ahrefs analyzed 863,000 keyword SERPs and 4 million AI Overview URLs, and found that 37.9% of the URLs cited in AI Overviews also appeared in the first 10 results for the same query, counting SERP features such as ads and featured snippets. In July 2025, its earlier study put that figure at about 76%. The two figures aren’t directly comparable: Ahrefs has since changed its parsing method and more than doubled the dataset. Counting only the regular blue links, 37.1% of cited pages ranked in the top 10 for the original query, 26.2% ranked 11 to 100, and 36.7% didn’t rank in the top 100 at all. Ahrefs’ reading: AI Overviews now lean more on the results of fan-out searches. Method note: Ahrefs’ data comes from its own tool, Brand Radar, and counts only overlap with the original query’s SERP.
Citations also in the top 10
AI Overview URLs in the first 10 results for the same query, counting SERP features such as ads and featured snippets
Where cited pages ranked
March 2026, blue links only, original query
Brand mentions correlate more than backlinks. In a May 2025 study of 75,000 brands, Ahrefs found that branded web mentions had the strongest correlation with brand visibility in AI Overviews (Spearman 0.664), compared with 0.218 for the number of backlinks. Method note: this is a correlation, as Ahrefs itself stresses (“correlation ≠ causation”), and it predates the 2026 changes to AI Overviews.
Correlation with AI Overview brand visibility
Spearman correlation, 75,000 brands
Adding schema didn’t move citations. In a May 2026 study, Ahrefs tracked 1,885 pages that added JSON-LD, against 4,000 matched control pages. Adding schema “produced no major uplift in citations on any platform”: +2.4% in AI Mode, statistically indistinguishable from zero, and a small, unexplained −4.6% in AI Overviews. Method note: every page studied was already heavily cited, so the study can’t say whether schema helps a page that isn’t cited yet.
Change in AI citations after adding schema
1,885 pages that added JSON-LD vs 4,000 matched controls
Gray on purpose: neither is a meaningful lift. Only pages already being cited were studied.
Taken together, these are consistent with Google’s guidance: rank well, be the source others talk about, and don’t expect a markup tweak to do the job.
Myths you can skip
Google’s guide has a whole “mythbusting” section. The short version:
Myth
“You need an llms.txt file.”
Reality: For Google Search, no. Google says it doesn’t use them: creating one “will neither harm nor help your site’s visibility or rankings in Google Search, as Google Search ignores them.” Other services may use them; that’s a separate decision.
Myth
“You need special schema.”
Reality: “There’s no special schema.org markup you need to add.” Structured data is still worth using for rich results, as long as it matches the visible text.
Myth
“Chunk your content into tiny pieces.”
Reality: Google says there’s “no requirement” to do this, and “there’s no ideal page length.”
Myth
“Create a page for every fan-out query.”
Reality: Doing this “primarily to manipulate rankings or generative AI responses” violates Google’s scaled content abuse spam policy. Google also calls it “an ineffective long-term strategy.” In May 2026, Google clarified that its spam policies apply to generative AI responses too.
Myth
“Plant mentions everywhere.”
Reality: See above: inauthentic mentions don’t help, and Google’s spam systems feed into its AI features.
Myth
“This tool shows Google’s AI ranking.”
Reality: “No third-party tool has access to our internal ranking or AI systems,” Google says. Good tools observe what the answers actually show.
A checklist to start this week
- Check eligibility. In Search Console, confirm Settings > Search generative AI is set to include your site, and use URL Inspection on your 10 most important pages to confirm they’re indexed.
- Audit blockers. Look for
noindex,nosnippetormax-snippetwhere you didn’t intend them, and confirm your CDN and firewall let Googlebot in. - Write your question set. List 20 to 50 questions your buyers ask, including comparisons (“X vs Y”), “best X for Y,” and problems your product solves.
- See who’s cited today. Run them in Google, note which have an AI Overview, try them in AI Mode, and record which pages each answer links to.
- Find your gaps. Pick the questions where competitors or third-party sites are cited and you aren’t. Look at what those cited pages offer that yours don’t.
- Upgrade one page at a time. Add what only you can say: your data, your experience, your specific facts. Improve the existing page rather than spinning up new ones for each variation.
- Earn real coverage. Reviews, comparisons, industry publications and customer stories, in every language you sell in.
- Measure, then repeat. Use the same questions on a fixed schedule and compare the trend.
How to measure your AI Overviews and AI Mode visibility
Search Console · Google only
ExampleGenerative AI performance report
Impressions in AI Overviews and AI Mode, by page, country, device and date.
Question set · every engine
ExampleYour buyers’ questions, over time
CitedNot cited
For each answer: mentioned, how you’re described, cited, which competitors appear.
In Search Console
AI Overviews and AI Mode are counted inside the regular Performance report (the “Web” search type). Google announced a dedicated Generative AI performance report in June 2026 and rolled it out to all sites on August 31, 2026. It shows how often links to your site appeared in AI Overviews and AI Mode (impressions), with breakdowns by page, country, device and date. If you don’t see it, Google says that’s usually because the site hasn’t received enough impressions yet, or because it was excluded in the generative AI setting.
Two things to keep in mind. The report counts impressions, so pair it with clicks and conversions from the Performance report and your analytics. Google has said that clicks from results pages with AI Overviews are “higher quality,” meaning people “are more likely to spend more time on the site,” so time on site and conversions are worth watching. And it only covers Google.
Across AI engines
Your buyers also ask ChatGPT, Gemini, Perplexity and others. To see the whole picture, track the same question set across engines and record four things for each answer: whether you’re mentioned, how you’re described, whether your pages are cited, and which competitors appear when you don’t. Answers vary from run to run, so look at the trend across many questions, not a single screenshot. Our AI visibility glossary explains how to calculate share of voice and citation rate, and how to rank in ChatGPT covers the engine Google’s features are most often compared with.
How Infinita helps
Infinita runs the questions your buyers ask on Google AI Overviews, Google AI Mode, ChatGPT and the other major AI engines on a schedule, saves every answer with its sources, and counts mentions, share of voice and citations from those saved answers. You see which questions trigger an AI answer, which pages it cites, how your brand is described, and which competitors are recommended when you’re not.
With Ask Vera, your AI marketing agent, you can ask questions about your own data, like “Which questions in AI Mode cite a competitor and not us?”, and get guidance on which page to improve first.
We’re learning alongside the teams we work with, and Google’s features keep changing. If you’d like to see how your brand shows up in AI Overviews and AI Mode today, book a demo.