How to Rank in Google AI Mode: What Google’s Own Guide Says

By Ace Zhuo | October 20, 2026

Illustration of one search query branching out into many web pages, representing how Google AI Mode splits a question into related searches

In short: There’s no secret AI Mode algorithm to game. Google says AI Mode runs on its core Search ranking and quality systems, and that optimizing for it “is optimizing for the search experience, and thus still SEO” [1]. In practice: check that your site hasn’t been excluded in Search Console, make sure your pages can be indexed and shown with a snippet, rank well for the question and the sub-questions AI Mode searches for, and publish content nobody else has. Then measure it with Search Console’s Generative AI report, because a single check of AI Mode tells you very little. (Whether you call it GEO, AEO or AI SEO, Google’s position is that for Google Search it’s still SEO [1].)

Most advice on AI Mode is guesswork from people looking at a handful of answers. In July 2026 Google published its own guide to generative AI features in Search [1]. This post walks through what it says, adds what independent studies have measured, and turns it into steps you can do this week.

What AI Mode is, and why it’s different from AI Overviews

AI Mode is Google’s conversational search mode. Google says it’s built for questions that need “further exploration, reasoning, or complex comparisons” [2]. To answer, it uses a technique Google calls query fan-out: it splits your question into subtopics and searches for each one at the same time, then writes an answer with links [6].

AI Mode and AI Overviews look similar, but Google says they “may use different models and techniques, so the set of responses and links they show will vary” [2]. Ahrefs measured this: across the same searches, AI Mode and AI Overviews cited the same pages only 13.7% of the time, even though their answers were 86% similar in meaning [7]. So being cited in one doesn’t mean you’re cited in the other. (If AI Overviews are your main concern, see our step-by-step guide: How to Optimize for AI Overviews.)

Step 1: Check that your site is included in Search Console

This is the step almost every AI Mode guide misses. Google’s 2026 guide says that, in addition to the normal technical requirements, “a site must be included in Search generative AI features in Search Console to be eligible” [1].

Inclusion is the default for every property [3], so most sites are already in. But it’s worth two minutes to confirm nobody switched it off:

  1. Open Search Console and go to Settings > Search generative AI [3].
  2. Check the setting is Include (or Inherit control from parent, and the parent property is set to include).
  3. If it says Exclude, your site gets no traffic or impressions from AI Overviews, AI Mode or Google Discover’s AI features [3].

Google says this control isn’t used as a ranking signal for the rest of Search [3], so including your site doesn’t hurt anything else.

Step 2: Make sure Google can index your page and show a snippet

To appear in AI Mode at all, “a page must be indexed and eligible to be shown in Google Search with a snippet” [2]. That means:

  • The page is indexed. Check it with URL Inspection in Search Console.
  • Crawling is allowed in robots.txt and by your CDN or hosting firewall. Google calls out both [2]; a security setting that blocks bots can quietly block Google too.
  • No snippet restrictions you didn’t mean to set. A nosnippet tag stops a page being used as a direct input for AI Overviews and AI Mode, and max-snippet limits how much of it can be used [2].
  • The main text is in the page’s HTML, not only loaded later by scripts.

None of this is new. That’s the point: if a page can’t win a normal snippet, it can’t be used in AI Mode.

Step 3: Rank for the question and its sub-questions

Because of query fan-out, AI Mode doesn’t only look at results for the exact question someone typed. It runs several related searches and draws on what it finds [6]. A page that ranks for the sub-questions has more chances to be used.

How much does ranking matter? Studies disagree, partly because they measure different things:

StudyWhat they measuredResult
SE Ranking, 10,000 US keywords (Jun 2025) [8]AI Mode links that also rank in Google’s organic top 1014% of URLs, 21.9% of domains
Semrush, AI Mode study (Jul 2025) [9]Same, different sample and method~35% of URLs, ~54% of domains

Either way, the pattern is clear: ranking helps, but most AI Mode links aren’t top-10 pages for the original query. Many come from the related searches AI Mode runs in the background.

What to do: write down the main question your customer asks, then the 4–6 follow-up questions they’d ask next (price, comparison, “is it worth it”, how long it takes, what can go wrong). Check where you rank for each in Google; the follow-ups are often less competitive than the main query. Then answer them on the page that owns the topic.

Step 4: Build one strong page, not one page per variation

It’s tempting to create a separate page for every version of a question. Google warns against exactly this: creating content for every variation “primarily to manipulate rankings or generative AI responses” breaks its scaled content abuse spam policy [1]. Google also says you don’t need to chase every long-tail variation [1].

Instead, cover the main question and its sub-questions on one page, with clear headings, and link to a separate page only when a sub-topic genuinely needs its own depth.

Step 5: Publish something only you can say

Google’s guide is blunt about what matters most: content that people find “unique, compelling, and useful will likely influence your success in generative AI search in the long run more than any of the other suggestions” in it [1]. Its examples include a unique point of view and first-hand experience.

AI Mode is built for comparisons and reasoning [2], so the content that stands out is specific: real prices and what changes them, your own data or tests with dates, clear criteria for choosing between options (including when you’re not the right choice), and photos or worked examples from your own work. A page that says what every other page says gives AI Mode no reason to link to you.

Step 6: Keep your Business Profile and product data accurate

If you’re a local business or you sell products, this step matters as much as your website. Google says AI responses can include local business information and product listings, and that using Google Business Profiles and Merchant Center feeds “can help your products and services to be visible in both AI responses and other Google Search results” [1].

  • Local businesses: make sure your Google Business Profile uses the same business name as your website, and that your services, hours, area, phone number and photos are current. Reply to reviews too.
  • Online stores: keep your Merchant Center feed accurate, especially prices and availability, so AI Mode isn’t working from out-of-date product information.

What you can ignore

Google’s guide names several popular “AI SEO” tactics that don’t help in Google Search [1]:

  • llms.txt and other AI text files. Google Search ignores them: they “will neither harm nor help” [1].
  • “Chunking” content into tiny pieces for AI. There’s no requirement to [1]. Write for people.
  • Special schema for AI. Structured data “isn’t required for generative AI search” and there’s no special markup to add [1]. It’s still useful for rich results.
  • Chasing inauthentic mentions across the web [1].
  • Tools that claim to know Google’s internal AI metrics. Google says no third-party tool has access to its internal ranking or AI systems [1].

How to measure AI Mode visibility

Search Console’s Generative AI performance report shows how many times links to your site appeared in AI Overviews and AI Mode [4]. Since August 31, 2026 it’s available to all sites [4]. Things to know:

  • It reports impressions, grouped by page, country, device and date [4]. Clicks from AI Mode are counted in the main Performance report, under the Web search type [2][5].
  • Two results from your site in the same AI answer count as one impression [4].
  • If you don’t see the report, your site may not have had enough AI impressions yet [4].

Check the answers by hand too, and more than once. AI Mode answers change: in SE Ranking’s test, only 9.2% of linked URLs were the same across three runs of the same query on the same day [8]. So test a short list of real customer questions, repeat each one, and track how often you appear, not whether you appeared once.

FAQ

Is Google AI Mode the same as Gemini? No. AI Mode is part of Google Search and draws on Search’s index and ranking systems [1]; the Gemini app is a separate product.

Do I need schema markup to rank in AI Mode? No. Google says structured data isn’t required for generative AI search [1]. Use it where it earns rich results, and keep it matching what’s visible on the page.

Can I stop my content appearing in AI Mode? Yes. Set the Search generative AI control to Exclude in Search Console [3], or use nosnippet, data-nosnippet, max-snippet or noindex on specific pages [2]. Excluding means no traffic from these features.

Where is AI Mode available? In many countries and languages. Google keeps the current list on its AI Mode help page [6]. Answers can also be personalized for signed-in users, which is another reason two people may see different results.

Check where you stand

Our free tools don’t check Google AI Mode itself, but they cover the two things that usually come first:

  • The free AI SEO audit checks whether search engines and AI search crawlers can reach your homepage, whether anything blocks indexing, and whether your key business details are clear.
  • The free AI visibility checker asks ChatGPT, Claude and Gemini a question your customers would ask, with live web search, and shows whether you’re mentioned and which sites they cite instead.

Sources