To rank in Google AI Overviews and AI Mode, a page first has to be indexed and eligible to show a snippet in Google Search, because both features take their supporting links from Google’s own index. Beyond that, Google says there are no special requirements: the pages that get linked are the ones that answer a specific question, and the narrower questions around it, clearly and credibly. The work is ordinary SEO done with more discipline, plus a sharper focus on answer-first writing and on covering follow-up questions.
This guide covers what Google has said publicly, what remains unknown, and a step-by-step plan for the pages that matter most to you.
Key takeaways
- Eligibility comes first. A page must be indexed and allowed to show a snippet before it can appear as a link in an AI Overview or in AI Mode.
- There is no AI-only markup or file. Google says the same foundational SEO practices apply and that no special optimization is required.
- Fan-out is the opportunity. Both features can run several related searches for one question, so pages that answer the narrower sub-questions get more chances to be cited.
- Google-Extended does not decide this. Google says the token does not affect inclusion or ranking in Search, so blocking or allowing it has no bearing on AI Overviews.
- Measurement takes extra work. Search Console now shows impressions in AI Overviews and AI Mode, but not how an answer described you or who else it linked to, so you still need to sample answers yourself or use a tracker.
What AI Overviews and AI Mode are
AI Overviews are the AI-generated summaries that appear above or among the regular results for some Google searches. They give a short answer and show links to the pages that support it, in a panel beside the text and sometimes inline. Google decides per query whether to show one, and many queries never get an AI Overview at all.
AI Mode is a separate, conversational way to search. You open it from a tab or button on the results page, ask a longer question, get a fuller answer with links, then ask follow-ups in the same thread. It suits comparison, planning and how-to questions that used to take several searches.
Both are built on Gemini models working with Google’s search systems, and both cite web pages. The practical difference is depth: an AI Overview usually links a few sources for one question, while an AI Mode conversation can touch many sub-questions and cite a wider range of pages.
If you are new to the wider discipline, start with what answer engine optimization is, and keep the AI search glossary open for terms such as fan-out and citation.
What Google has said publicly
Google publishes guidance for site owners on its AI features. The main points are worth knowing precisely, because they rule out a lot of bad advice.
- No extra requirements: to be eligible as a supporting link, a page must be indexed and eligible to be shown in Google Search with a snippet. There is no separate application, file or markup.
- Same best practices: Google says foundational SEO still applies: let Googlebot crawl, make content findable through internal links, provide a good page experience, put important content in text, and keep structured data consistent with what is visible on the page.
- Query fan-out: AI Overviews and AI Mode may issue multiple related searches across subtopics and data sources to build a response, so they can surface a wider and more varied set of pages than a classic results page.
- Snippet controls apply: the existing controls for snippets, such as nosnippet, data-nosnippet, max-snippet and noindex, also govern how your content can appear in these features.
- Reporting has its own view: Search Console’s Generative AI performance report shows impressions in AI Overviews and AI Mode by page, country, device and date. The same data still counts in the main Performance report.
What is not known
Google has not published how it chooses which pages to link in a given AI Overview or AI Mode answer, how many sources it considers, or how much weight any single signal carries. Anyone who claims to know the formula is guessing.
A few things are reasonable to infer from Google’s own statements and from looking at answers, but treat them as working assumptions rather than rules:
- Because the links come from Google’s index, pages that are strong in organic search are natural candidates. Links do not always mirror the top organic results, though, which fits with fan-out searches pulling pages from related queries.
- Answers vary with wording, location, language and time, and the same question can produce different links on different days.
- Formatting can make a passage easier to quote, but no format guarantees selection. A page earns its place by answering the question better than the alternatives.
Tip: Be skeptical of anyone who promises a fixed position inside an AI Overview. Google has not published a ranking order for these links, and placement can change between runs of the same query.
Step 1: Confirm your pages are eligible
Start with the boring checks, because a page that fails them cannot appear no matter how good it is. The AI search optimization checklist covers the same ground for every engine.
- Check indexing. Use the URL Inspection tool in Search Console on your most important pages. Confirm they are indexed, that the Google-selected canonical is the URL you expect, and that the rendered HTML contains your main content.
- Check snippet controls. Look for nosnippet, max-snippet or data-nosnippet left behind by a theme, a plugin or an old experiment. These limit what Google can show, in AI features as well as in regular snippets.
- Check crawl access. Make sure robots.txt, your CDN and any security plugin let Googlebot through. Our guide to which AI crawlers to allow covers rules per bot, and the free AI robots.txt generator builds a clean file.
- Decide on Google-Extended separately. Google says this token does not affect inclusion or ranking in Google Search. Make that call for Gemini reasons, not for AI Overviews.
- Put the content in text. Key facts locked inside images, PDFs, late-loading tabs or client-side scripts are harder to use. Keep the answer in the HTML.
Step 2: Find the questions where AI answers appear
You cannot optimize for every query, so start with the questions that matter to revenue and check which of them trigger AI features.
Write a list of 25 to 50 questions your buyers ask across the journey: definitions, how-to questions, comparisons, “best X for Y” and pricing. The free AI prompt ideas generator gives you a first draft to trim. Then search each one in Google and note whether an AI Overview appears, which pages it links, and whether yours is among them. Ask the same questions in AI Mode and record the links there too.
Keep the method consistent: same country, same language, signed out where possible, and on more than one day. A single check is an anecdote. Patterns across several checks are what you can act on.
When you finish, sort the questions into three groups: you are linked; a competitor or third party is linked and you are not; no AI answer appears. The middle group is your priority list.
Step 3: Map the fan-out around each question
Because AI Overviews and AI Mode can run several related searches for one question, the pages they cite often answer a narrower sub-question than the one the person typed. A question like “how do I choose a CRM for a small agency” might fan out into pricing, integrations, migration effort, alternatives and reviews.
For each priority question, list the sub-questions a careful researcher would ask next. Then check whether you have a page, or a clearly headed section, that answers each one. Gaps on this list are often easier to win than the head question, because fewer strong pages target them.
- Pricing and cost questions, including what drives the price up or down.
- Comparison and alternative questions that name competitors.
- How-to and setup questions with concrete steps.
- Suitability questions: who it is for, and who it is not for.
Step 4: Write answer-first sections that stand alone
AI answers often quote or summarize a single section rather than a whole page. So each important section should open with a direct answer and make sense without the paragraph before it.
- Lead with the answer. Put a two or three sentence answer under each key heading, then add detail, exceptions and steps.
- Use descriptive headings. Phrase headings the way people ask, for example “How long does onboarding take?” rather than “Onboarding”.
- Be specific. Replace vague claims with numbers, ranges, names, dates and conditions that a reader could check.
- Keep paragraphs short. One idea per paragraph, with lists for steps or options.
- Show experience. Google’s guidance on helpful, people-first content stresses first-hand experience and expertise. Add what you have actually tested, measured or seen with customers.
Paste a draft into the free answer readiness checker to see how directly it answers its target question. It is a writing aid, not a model of Google’s systems, but it catches buried answers and vague openings quickly.
Step 5: Back the page with clear signals
Content does the heavy lifting, but a few supporting signals remove doubt about what the page is and who stands behind it.
Structured data that matches the page
Use schema.org markup that describes the page truthfully: Organization on your home or about page, Article on guides, Product or SoftwareApplication on product pages, LocalBusiness for locations, and FAQPage where the questions and answers are visible on the page. Google says structured data should match the visible content, and it is not required for AI features, so treat it as clarity rather than a ranking lever. The free FAQ schema generator produces valid FAQPage JSON-LD in a minute.
Authorship and dates
Show who wrote or reviewed the page and when it was last updated. For topics where facts change, such as pricing or regulations, a visible and honest update date helps readers and gives search systems a reason to prefer the current version.
Internal links
Link to your answer pages from related pages with descriptive anchor text, so they are easy to discover and clearly connected to the topic. Orphan pages are harder to crawl and harder to understand in context.
Images, video and business data
Google’s advice for its AI search experiences encourages supporting text with good images and video, and keeping Merchant Center and Business Profile information up to date. If you sell products or serve local customers, those feeds and profiles are part of how you show up, not an afterthought.
Step 6: Earn the reputation the answer relies on
AI answers often cite third-party pages: reviews, comparisons, forums and industry publications. If the pages linked for your priority questions are mostly third-party, the route to visibility runs through those sources as well as through your own site.
Look at the domains that appear again and again for your questions. Make sure your brand is described accurately on the review and directory sites among them, contribute useful expertise where your audience discusses the topic, and pursue coverage in publications that already rank. Our guide to how AI search engines choose sources explains why corroboration across independent sites matters.
To see how your link profile compares with the sites that keep winning, run your domain and theirs through the free domain authority checker. It measures link strength, not AI visibility, but a large gap is worth knowing about.
Step 7: Measure what changes
Measuring AI Overview visibility takes more effort than tracking a ranking, for two reasons: answers vary between runs, and impressions alone do not show how an answer described you or who else it linked to.
- Search Console: open the Generative AI performance report to see impressions in AI Overviews and AI Mode for the pages you are working on, and compare them with clicks in the main Performance report. Rising AI impressions with flat clicks is worth investigating.
- Answer sampling: re-run your question list on a schedule and record whether you are linked, who else is linked and how the answer describes you. Compare over weeks, not days. The method in how to track brand visibility in ChatGPT applies here too.
- Analytics: check landing pages and engagement for organic visits to the pages you improved. Treat AI Mode and AI Overview visits as part of Google organic unless your tools can separate them.
- Crawl data: confirm in your server logs that Googlebot is fetching the pages you updated. The AI crawler analytics docs explain how to read bot activity from logs.
When you lose a link you used to have, work through the diagnosis in how to win back lost citations before rewriting anything.
If you want to limit how Google uses your content
Some publishers want the opposite: less of their content shown in AI answers. Search Console has a Search generative AI setting that excludes a whole site from AI Overviews, AI Mode and generative AI features in Discover while it stays in regular results. For single pages or passages, the snippet controls still work: nosnippet or max-snippet on a page, data-nosnippet around specific passages, or noindex to remove the page from Search entirely. Snippet controls also change how the page appears in regular results, so test them on a few pages first.
Common mistakes
- Adding schema, llms.txt or “AI-friendly” summaries while the page itself still buries the answer.
- Judging success from one screenshot instead of repeated checks over several weeks.
- Chasing the head question only and ignoring the sub-questions that fan-out reaches.
- Publishing many thin pages for every variation of a question, which spreads your signals thin.
Frequently asked questions
Do I need special markup to appear in Google AI Overviews?
No. Google says there are no additional technical requirements beyond being indexed and eligible to show a snippet in Search. Structured data can make a page easier to interpret, but it is not required for AI features, and it must match what is visible on the page.
Does llms.txt help with AI Overviews?
There is no public indication that Google uses llms.txt for AI Overviews or AI Mode. It is a lightweight file that some other AI tools may read, and the free llms.txt generator makes one in minutes, but for Google your effort is better spent on indexing, content and snippet eligibility.
Will blocking Google-Extended remove me from AI Overviews?
No. Google states that Google-Extended does not affect a site’s inclusion or ranking in Google Search, and AI Overviews and AI Mode are part of Search. To keep content out of AI features, use snippet controls or noindex, and accept the effect on regular results.
Can I see AI Overview traffic in Search Console?
Yes, for impressions. Search Console’s Generative AI performance report shows how often links to your site appeared in AI Overviews and AI Mode, by page, country, device and date, and those impressions also count in the main Performance report. It does not show how the answer described you, so pair it with answer sampling to see which pages are linked and why.
How long does it take to appear after updating a page?
There is no fixed timeline. The page has to be recrawled and reprocessed, and AI answers can change from day to day. Request indexing for important updates, then judge the result over several weeks of repeated checks.
Do I have to rank on page one to be linked?
Not necessarily. Links come from Google’s index and strong organic pages are natural candidates, but fan-out searches mean a page that answers a related sub-question can be linked too. Ranking well remains a good foundation; it is not the only route.
Track your Google AI visibility with aeoora
aeoora asks your buyer questions every day in Google AI Overviews and AI Mode, alongside ChatGPT, Perplexity, Gemini, Claude and Copilot, and shows which pages each engine links for you and for your competitors. It also measures AI referral traffic and suggests content and site fixes ranked by priority. The two Google engines are included on the Signal plan at $69 a month and the Scale plan at $299 a month, while Shine at $29 a month covers the other main assistants. Every plan has a 2-day free trial and 25% off yearly.
For a quick outside-in look first, get the free AI visibility report, which asks 25 buyer questions in several AI engines and emails you a PDF. When you are ready to track your own prompts, start your 2-day free trial and follow the quickstart to get your first results the same day.


