How to Rank in Google AI Overviews and AI Mode

To rank in Google AI Overviews, a page has to be indexed, eligible to show a snippet, and good enough to be retrieved by Google's core ranking systems for the main query or its fan-out sub-queries. Google says there is no special markup or file for AI Overviews. The work is strong SEO, non-commodity content, and authentic third-party evidence.

This guide covers what Google officially says (including its May 2026 optimization guide and the new Search Console controls), what independent studies show about which pages get cited, how Google AI Mode differs, a step-by-step plan, and how to measure results in Search Console. Facts are current as of October 2026.

How do Google AI Overviews choose which sources to cite?

AI Overviews cite pages that Google's normal ranking systems retrieve from the Search index. Google's guide to optimizing for generative AI features says these features are "rooted in our core Search ranking and quality systems" and rely on two techniques: retrieval-augmented generation (grounding) and query fan-out.

Grounding means the model retrieves relevant, current pages from the index, reviews what they say, and writes a response with clickable links to the pages that support it. If your page is not retrieved, it cannot be cited.

Query fan-out means the model issues several related searches at once. Google's own example: for "how to fix a lawn that's full of weeds", fan-out queries might include "best herbicides for lawns", "remove weeds without chemicals" and "how to prevent weeds in lawn". A page that ranks for one of those sub-queries can be cited even if it does not rank for the original query. Our explainer on query fan-out covers the mechanics.

The model layer keeps changing. On 27 January 2026, Google made Gemini 3 the default model for AI Overviews globally and let people ask follow-up questions that continue in AI Mode, Search Engine Journal reported. The retrieval principle has stayed the same: ranking comes first.

What the studies show about cited pages

Independent data supports Google's description, with a twist. An Ahrefs analysis of 1.9 million AI Overview citations (July 2025) found that 76% of cited pages also ranked in Google's top 10. By its March 2026 re-analysis, that overlap had fallen to about 38%, which Ahrefs links partly to fan-out pulling sources from related searches.

Not every query triggers an overview. Ahrefs' study of 146 million SERPs (September 2025 data) found AI Overviews on 20.5% of results pages. They appeared on 57.9% of question queries versus 15.5% of non-questions, and on 46.4% of queries with seven or more words versus 9.5% of one-word queries.

These are correlations, not ranking factors Google has confirmed. They do tell you where to focus: informational, question-shaped queries where you can already reach page one.

What Google says you need, and what you can skip

Google's position is that "optimizing for generative AI search is optimizing for the search experience, and thus still SEO." Its AI features documentation states there are "no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary."

The May 2026 guide goes further and lists tactics you can ignore for Google Search:

• llms.txt and other "special" files. Google Search does not use them. They neither help nor hurt. (They may matter for other systems; see our llms.txt guide.)
• Chunking content into tiny pieces. Google says its systems can find the relevant part of a long page. Writing in clear, self-contained sections still helps readers and other AI engines, which is the case our semantic chunking guide makes.
• Rewriting content "for AI." AI systems understand synonyms, so you do not need every long-tail variant.
• Inauthentic mentions. Planted mentions are filtered by the same spam systems as the rest of Search.
• Overfocusing on structured data. No special schema is needed for AI features, though structured data still earns rich results.

Google also warns that creating separate pages for every fan-out variation "primarily to manipulate rankings or generative AI responses" violates its scaled content abuse policy.

What are the technical requirements for AI Overviews?

A page must be indexed, eligible to show a snippet, and meet Google's Search technical requirements. Since 2026 there is one more condition: your site must be included in the new Search generative AI setting in Search Console.

Google's Search generative AI control, rolled out to all sites worldwide as of 31 August 2026, lives under Settings > Search generative AI. "Include" is the default for top-level properties. "Exclude" removes your links and content from AI Overviews, AI Mode and generative AI features in Discover without affecting classic rankings. Child properties inherit from the parent unless overridden.

Use this checklist before any content work:

1. Confirm the Search generative AI setting is "Include" for every property you care about.
2. Check that Googlebot is not blocked in robots.txt, at your CDN, or by firewall rules.
3. Confirm key URLs return HTTP 200 and are indexed, using URL Inspection.
4. Remove accidental nosnippet, data-nosnippet or restrictive max-snippet values. Google says more restrictive preview controls limit how your content appears in AI features.
5. Keep important content in text, not only in images or scripts. If you rely on JavaScript, follow Google's JavaScript SEO practices.
6. Consolidate duplicates with canonicals so one URL collects the signals.

Google-Extended is not an AI Overviews switch: Blocking Google-Extended limits use of your content for training and grounding in Google's other AI products. It does not remove you from AI Overviews or AI Mode. Use the Search Console control or snippet controls for that.

For a wider site review, use our AI search readiness audit.

How to rank in AI Overviews: a step-by-step plan

The plan below turns Google's guidance and the citation data into an order of work. Each step builds on the one before it.

1. Choose queries where an AI Overview actually appears

Start with question-led, longer informational queries in your category, because those trigger overviews most often. Check the live results page for each target and note which domains are cited today. Drop queries with no overview from this project; they belong in your normal SEO plan.

2. Map the fan-out sub-questions

For each target query, list the sub-questions Google is likely to search. Our free Query Fanout tool generates related queries for a topic across research, comparisons, use cases, implementation and limitations. Cover the important sub-questions inside one strong page, or in a small number of genuinely distinct pages, rather than one thin page per variant.

3. Rank on page one for the head term and the sub-questions

Classic SEO is the entry ticket. Improve the page that already ranks closest to the top 10, strengthen internal links to it, and fix intent mismatches. Top-10 pages still earn a large share of citations in Ahrefs' data, so moving from position 14 to position 6 is usually worth more than any formatting change.

4. Make the content non-commodity

Google says unique, valuable content will influence your presence in AI search "more than any of the other suggestions" in its guide. Its example contrasts "7 Tips for First-Time Homebuyers" (commodity) with a first-hand account of a specific decision (non-commodity). Add what only you have: original data, test results, pricing you have seen, screenshots, expert quotes, and worked examples.

5. Put the answer where a reader expects it

Google does not require a special format, but readers and extraction both benefit from clarity. Open each section with a one- or two-sentence direct answer, use descriptive headings that match how people ask, and use tables for comparisons. This is good writing first; any AI benefit is a side effect.

6. Support text with images, video, feeds and listings

AI responses can include images, video, product listings and local business details. Google recommends high-quality images and video, current Merchant Center feeds for products, and an accurate Google Business Profile for local businesses.

7. Build authentic evidence beyond your website

Google AI answers mostly cite other people's pages. In Lantern's AI Recognition Gap research (four emerging brands, April to September 2026), third-party sites made up 99.5% of Google AI source links. In a related study of where new brands earn AI authority, independent sources accounted for 77.6% of Google AI citation activity, competitors 16.4%, and brand-owned pages 0.5%. Earn genuine reviews, analyst and press coverage, and inclusion in independent comparisons.

8. Strengthen the owned pages Google AI actually cites

When Google AI does cite a brand's own site, it leans on product pages. In Lantern's Beyond the Homepage report, product pages were 53.6% of owned Google AI citations, the homepage 36.8%, and editorial content 7.2% for the brand studied. Make product and service pages specific: who it is for, what it does, pricing where public, and proof.

9. Refresh with substance, then measure

Update facts, data and examples when they change, and show a real "last updated" date. Cosmetic date changes do not add value. Our post on why old content loses AI citations covers how to prioritize refreshes. Then measure, using the Search Console steps below.

Google AI Mode: how it differs and how to show up

AI Mode is Google's conversational search experience for questions that need reasoning, comparison or several steps. It uses the same eligibility rules and the same Search index as AI Overviews, but Google notes the two "may use different models and techniques", so their answers and links can differ.

AI Mode is now a mainstream surface. Google said at I/O on 19 May 2026 that AI Mode had passed 1 billion monthly users and AI Overviews had over 2.5 billion. The same week, Google began rolling out a redesigned search box with an AI Mode shortcut, and tapping "Show more" on an AI Overview now moves people into AI Mode, according to 9to5Google, which reported AI Mode runs on Gemini 3.5 Flash.

To show up in AI Mode, extend the AI Overviews plan in three ways:

• Cover the decision, not just the definition. AI Mode handles "which is better for X" and "what should I choose if" questions, so publish fair comparisons, selection criteria and use-case pages.
• Answer the follow-ups. Anticipate the second and third question a buyer asks and answer them on the same page or a linked one.
• Prepare for agents. Google's guide points to agent-friendly website practices, the Universal Commerce Protocol, and Business Agent, a conversational experience that lets customers chat with a brand on Search.

How to measure AI Overview performance in Search Console

Search Console now has a dedicated Generative AI performance report. Google announced it on 3 June 2026 and rolled it out to all websites worldwide as of 31 August 2026.

The report shows impressions from AI Overviews and AI Mode, broken down by pages, countries, devices and dates. A search type filter splits text queries from multimodal ones (Lens, Circle to Search and image uploads), added on 24 September 2026. Clicks from AI features remain inside the standard Performance report under the Web search type, and the AI report currently centers on impressions.

A simple monthly workflow:

1. Export Generative AI impressions by page and sort to find the URLs Google AI surfaces most.
2. Compare those pages with their Web clicks and average position in the main Performance report.
3. Flag pages with rising AI impressions but falling clicks; check whether the page confirms the answer quickly and offers depth worth clicking for.
4. Track engagement and conversions for those landing pages in Google Analytics. Google says clicks from results pages with AI Overviews tend to be higher quality.
5. Re-check citations on your target queries by hand or with a tracker, sampling repeatedly because answers vary.

Search Console tells you that you appeared, not which competitors were cited beside you or how you were described. That is the gap Lantern fills: it tracks mentions, citations and share of voice in Google AI Overviews alongside ChatGPT, Gemini and Perplexity, and connects that to your Search Console data. Google's own guide cautions that no third-party tool has access to its internal ranking systems, so treat any tool as observed data, not insider data.

Common mistakes that keep pages out of AI Overviews

Most missed citations come from a few avoidable errors:

• Leaving a property set to "Exclude" or inheriting an exclusion from a parent property.
• Using nosnippet sitewide for unrelated reasons, which also limits AI features.
• Publishing commodity summaries of what already ranks, with nothing new to cite.
• Spinning up dozens of near-duplicate pages for fan-out variants, which risks scaled content abuse.
• Measuring only clicks, and missing impressions and conversion quality.
• Assuming structured data or llms.txt alone will get you cited.

Frequently asked questions

How do I get my website listed in AI Overviews?

There is no submission form. Make sure the page is indexed and snippet-eligible, your Search generative AI setting is "Include", and the page ranks for the query or a related sub-question. Google then decides case by case whether to show an overview and which pages to link.

Can I opt my site out of AI Overviews and AI Mode?

Yes. Set Search generative AI to "Exclude" in Search Console to remove your links and content from AI Overviews, AI Mode and Discover AI features without changing classic rankings. Google says most content drops out within one to two days. Page-level controls like nosnippet also limit what can be shown.

Is AI Overview SEO different from regular SEO?

Google says no: optimizing for its generative AI features is still SEO. The emphasis shifts toward non-commodity content, coverage of related sub-questions, and evidence about your brand on other sites, but the foundations are the same crawl, index and ranking work.

Can Google rank AI-generated content in AI Overviews?

Google does not ban AI-assisted content. Its guidance says content made with generative AI tools must meet Search Essentials and spam policies, and mass-producing pages mainly to manipulate rankings violates the scaled content abuse policy. Content that adds original value is what gets retrieved and cited.

Does schema markup help you rank in AI Overviews?

Google says structured data is not required for generative AI features and no special schema exists for them. It still recommends structured data for rich results, and markup must match the visible page. Treat it as good hygiene, not a citation lever.

How long does it take to appear in AI Overviews?

There is no fixed timeline, and Google says meeting every requirement does not guarantee crawling, indexing or serving. Changes only count once Google recrawls the page, which can take days or longer. Track impressions in the Generative AI report over several weeks rather than judging one day's results.

Lantern helps marketing teams see where they stand in Google AI Overviews and ChatGPT, which sources are cited instead of them, and which pages to fix first, then puts agents to work on review-ready updates. Start with a free AI Visibility Report, and see our companion guide on how to rank in ChatGPT for the other half of AI search.