To show up in AI Overviews and Gemini's search-grounded answers, you do the same work that gets a page into Google Search: let Googlebot crawl it, get it indexed, and make it the most useful answer to a question your buyers actually ask. Google says these features rely on its existing Search index rather than a separate AI index, and the Google-Extended setting in robots.txt does not decide whether you appear in AI Overviews.
That sounds like an anticlimax, but it changes where you should spend your effort. Below is what Google has published, what it means in practice, and what we see when we put buyer questions to Gemini and the other assistants every week.
What does Google actually say about AI Overviews and Gemini?
Google's guidance is unusually direct. In Optimizing your website for generative AI features on Google Search, it states that SEO best practices remain relevant because AI Overviews and AI Mode rely on the same core Search ranking and quality systems. It also explains that grounding works by pulling relevant, up-to-date pages from Google's existing Search index, not from a separate AI-specific index.
Its page on AI features and your website adds that AI Overviews and AI Mode surface links to help people find information, and that these features open opportunities for a wider range of site types, not only informational content. A service page, a product page or a local practice's page can be one of those links.
The Wikipedia entry on generative engine optimization notes that in 2026 Google's official guidance described 'AEO' and 'GEO' as terms people may use while asserting that optimizing for generative AI search is fundamentally still SEO, the first time Google put that position in formal documentation rather than only saying it at conferences. If you want the broader picture of how the two disciplines overlap, see GEO vs SEO.
So the honest summary: for Google's AI surfaces, there is no back door. If a page cannot rank and be retrieved in Search, it has nothing to feed the AI answer.
Does Google-Extended control whether you appear in AI Overviews?
No. Google-Extended is a robots.txt token that is easy to mistake for the switch for 'Google AI'. It is not the switch for AI Overviews or AI Mode. Those features are part of Search, and Search is crawled by Googlebot.
In practice that means two things:
- Blocking Google-Extended does not keep you out of AI Overviews. If Googlebot can crawl and index the page, it can be retrieved for an AI Overview.
- Allowing Google-Extended does not get you in. If the page is not indexed or does not rank for the question, opening that token changes nothing for AI Overviews.
The lever that matters is Googlebot access plus indexing. If you block Googlebot or mark a page noindex, you lose the page from Search and therefore from the AI features built on it.
| What you change | Effect on AI Overviews and AI Mode | What to do |
|---|---|---|
| Block Google-Extended in robots.txt | Does not remove you; these features use the Search index | Decide on it for its own reasons, not for AI Overviews |
| Allow Google-Extended | Does not add you; inclusion still depends on indexing and ranking | Do not expect visibility from this alone |
| Block Googlebot or a path prefix by mistake | Pages drop out of Search, so they cannot be grounded on | Test every Disallow rule against your real URLs |
| Get a page indexed that answers a buyer's question well | Makes it a candidate for retrieval | This is where the work is |
What does 'it is still SEO' mean in practice?
It means the boring basics decide most of the outcome. Three of them matter more than anything else we have seen.
Can Googlebot reach your pages at all?
We made this mistake ourselves. On itgrows.ai, a robots.txt rule meant to hide a single redirect page (/b) was written as a prefix. Because robots.txt rules match prefixes, from 11 August to 21 October 2026 it also closed /blog, /business and every article in our business blog to all crawlers. Bing Webmaster Tools listed our business page as 'discovered but not crawled' since 11 August 2026, and Google had none of the business pages indexed. We changed the rule to /b$ on 21 October 2026.
No amount of good writing would have put those pages into an AI Overview during that period. Check your own rules line by line; our guide on whether your website is blocking AI crawlers walks through it.
Is there enough of you to find?
Markup alone does not carry a site. We have seen a site with complete markup (server rendering, Article, FAQPage and Organization JSON-LD, a sitemap) but only 75 URLs that was still absent from answers on its own article's exact topic, while a site with about 190,000 crawlable URLs was crawled constantly. Surface beats markup: the number of distinct, useful pages answering distinct questions matters more than how perfectly a handful of pages is tagged.
Googlebot is also not a heavy visitor compared with AI crawlers. Across the full retained nginx logs of an edtech platform we work with (pickaclass.com), from 28 July to 11 August 2026, ClaudeBot made 19,272,125 requests, GPTBot 2,744,985 and Googlebot 52,292. Google is selective, which is another reason every crawlable page should earn its place.
Is being indexed the same as being found?
Not quite. On 19 October 2026 we asked assistants for the exact titles of 18 articles, each older than three weeks, published on three client sites. ChatGPT found 1 of 18 and Claude found 7 of 18, while Google had indexed 29 of 30 articles of one of those sites. For Google's own AI features, indexing is the entry ticket. Ranking for the actual question is what gets you retrieved.
Which questions can actually name your business?
An AI answer only names businesses when the question calls for one. In a 23 September 2026 test of 20 German business questions, the model answered 10 without searching at all and cited nobody. Where it searched, it cited primary sources such as laws, ministries, the OECD and large accounting firms on 9 of 10. Firms were named only on 'who should I hire' or 'how to choose' questions, and two runs of the same question named two different sets of firms.
A local example shows the same pattern. On 20 October 2026, for an orthopaedic practice in Bonn, the four assistants named the practice in 12 of 80 answers to 20 patient questions: Claude 5 of 20, Gemini 4, ChatGPT 2, Perplexity 1. The directory jameda.de was cited in 12 answers, competing practices' own websites were the most cited sources, and 'who to choose' questions named practices while questions about costs, risks and treatments named none.
What to take from this:
- Write for selection questions. Pages that help someone choose a provider in your category, with honest criteria, are where businesses get named.
- Still answer the informational questions. They will rarely name you, but they build the surface and can be cited as a source.
- Mind directories and competitors' sites. If those are what Gemini cites, your own site needs pages that answer the same question at least as clearly.
The original research on generative engines, the GEO paper by Aggarwal et al., describes these systems as synthesizing information from multiple sources. Your page is one candidate among several; it has to contribute something the answer needs.
How do you know whether it is working?
Do not judge by referral traffic alone. In about two weeks of logs from three of our course websites (roughly 1.2 million requests in the two busiest logs), human visitors arriving from AI assistants were: Copilot 1; ChatGPT, Perplexity, Gemini, Claude, Meta AI and others 0. Assistants mostly answer in place. Being named or cited can matter even when nobody clicks.
Measure the answers instead, and measure them more than once. A 2026 survey of generative engine optimization concludes that different AI engines cite different ecosystems of sources, citation behaviour varies across repeated runs of the same query, and visibility must be measured across several engines and time periods rather than as one score. That matches what we see: asked the same question three times, an assistant often names different businesses. We count a business as named for a question if any of three asks names it, and compare weeks only on the same questions and the same number of asks.
It is cheap to do. With API prices of October 2026, one question asked once with web search costs us about $0.009 with Gemini, and a report of 20 questions across four assistants with three asks each costs about $5.40. For a step-by-step method, see how to check whether AI assistants mention your business.
Expect to start low. On 21 October 2026 we put 20 English and 20 German questions about getting a business into AI answers to ChatGPT, Claude, Gemini and Perplexity, three times each, and ItGrows was named in 0 of 160 answers (what we learned). The most cited sources were the blogs of GEO agencies and tool vendors. Movement does happen: in our weekly measurement of 20 October 2026, Perplexity cited an article we had published on a client's site, a German company-formation advisory. And among clients, as they report it, MagiScan, an iOS 3D-scanning app, gets hundreds of installs a month from ChatGPT, Claude and Gemini.
What should you do next?
- Audit robots.txt for Googlebot, not just for AI bots. Test each Disallow rule against real URLs; a prefix can close a whole section.
- Confirm indexing in Google Search Console for the pages you want cited. Not indexed means not eligible.
- List the questions your buyers ask, especially 'who should I choose' and 'how do I compare' questions, and publish one clear page per question on your own domain.
- Grow the surface steadily. A few dozen good pages beat five perfectly tagged ones.
- Measure weekly across engines, with repeated asks, on a fixed question set.
If you want a quick baseline first, the free check at itgrows.ai generates five buyer questions from your homepage and shows in 25 to 40 seconds whether ChatGPT and Perplexity name you. ItGrows publishes the articles, fixes crawler access and indexing, and measures four assistants weekly for $499 a month (plus a one-time $499 setup, free with 3 months or a year). Nobody can guarantee that Gemini or any other assistant will name a business, and we do not.
Frequently Asked Questions
Is there a special way to submit my site to AI Overviews?
No. Google says AI Overviews and AI Mode use the same core Search ranking and quality systems and pull from the existing Search index. Getting crawled and indexed by Googlebot is the only entry point.
Should I block or allow Google-Extended?
Decide that on its own merits, because it does not control AI Overviews. Inclusion in AI Overviews depends on Googlebot access and indexing, not on that token.
Do I need special schema markup to appear in Gemini answers?
Google's guidance says its AI features rely on the same core Search ranking and quality systems as regular Search, so the Search basics apply. We have seen a site with complete JSON-LD but only 75 URLs stay absent from answers on its own topic, so useful pages answering real questions matter more than tags.
Why does Gemini name me one day and not the next?
Answers vary between runs of the same question, and different engines cite different sources. Ask each question several times and compare the same questions week to week before drawing conclusions.