How to Get Your Website Cited in Google AI Overviews and ChatGPT

To get your website cited in Google AI Overviews and ChatGPT, you need three things: pages that search engines can crawl and index, content that answers questions better and more specifically than other sources, and access settings that let AI search crawlers in. There is no secret markup or special file. Google says AI Overviews and AI Mode pick supporting links from its normal search index, and ChatGPT search relies on its own crawler, OAI-SearchBot, which you must not block.

Softwix has worked on search visibility for businesses in more than 20 countries since 2017, and the pattern since AI answers arrived is clear: sites that already did solid SEO are the ones being cited. This guide turns that into a practical checklist you can apply page by page. If you would rather have a specialist review your site, our SEO services team covers AI search visibility as part of every engagement.

AI Overview and ChatGPT citations at a glance

Area	What to do	Why it matters
Eligibility	Make sure pages are indexed and can show a snippet	Google only uses indexed, snippet-eligible pages as supporting links
Crawler access	Allow Googlebot and OAI-SearchBot in robots.txt	Blocking OAI-SearchBot removes you from ChatGPT search answers
Content	Publish first-hand, specific, non-commodity content	AI systems compare many sources and favour ones that add something new
Structure	Answer first, clear headings, tables, short FAQ sections	Makes the right passage easy to find and quote accurately
Local and shop data	Keep Google Business Profile and Merchant Center feeds accurate	AI answers can include business and product listings
Measurement	Use Search Console’s Generative AI report and Bing’s AI Performance report	Shows which pages are actually being cited
What to skip	llms.txt for Google, “chunking”, fake mentions	Google says these do not help in its search

How do AI Overviews choose which websites to cite?

AI Overviews pull their supporting links from Google’s normal search index, using Google’s core ranking systems. In other words, the pages Google already trusts for a topic are the pages most likely to be cited.

Google’s guide to optimising for generative AI features, published in May 2026, explains two techniques behind this. The first is retrieval-augmented generation, where the AI retrieves relevant pages from the index and grounds its answer in them. The second is “query fan-out”, where the model runs several related searches at once to gather more information.

Fan-out matters for planning. A question such as “how to choose an accounting firm” may trigger related searches about fees, qualifications and switching providers. A page that covers those follow-up angles well has more chances to be retrieved.

What Google says you do not need

Google’s own AI features documentation (updated December 2025) 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 things you can ignore for Google Search:

• llms.txt and other AI text files: Google Search does not use them. They neither help nor harm your Google visibility.
• “Chunking” content into tiny pieces: Google can understand multiple topics on one page.
• Rewriting for AI: AI systems understand synonyms, so you do not need every keyword variation.
• Buying or faking mentions: inauthentic mentions are not as useful as they look, and spam systems filter them.
• Special schema: structured data is not required for AI features, though it still helps with rich results.

Google also says AEO and GEO work is, from its perspective, still SEO. That is useful when you assess agencies selling “AI optimisation” as a separate product.

What technical setup do you need before AI tools can cite you?

A page must be crawlable, indexed and allowed to show a snippet in Google. If it fails any of those, it cannot be a supporting link, however good the content is.

Work through these checks first:

1. Indexing: use the URL Inspection tool in Search Console to confirm your key pages are indexed.
2. Snippet controls: check you have not added nosnippet, a very low max-snippet value or noindex to pages you want cited. Google confirms these controls also limit how content appears in AI features.
3. robots.txt: make sure you are not blocking Googlebot from important sections, CSS or JavaScript.
4. Search Console AI setting: Google now has a “Search generative AI” control under Settings in Search Console. Sites are included by default, but if someone changed it to “Exclude”, your content will not appear in AI Overviews, AI Mode or Discover’s AI features. Check it.
5. Text, not images: keep important information in HTML text. Prices or key facts locked inside images or sliders are harder to extract.
6. Page experience: fast, mobile-friendly pages where the main content is easy to tell apart from ads and widgets.

Our technical SEO checklist covers 25 of these checks in more detail, including canonical tags, sitemaps and JavaScript rendering.

How do you let ChatGPT find your pages?

Allow OpenAI’s search crawler, OAI-SearchBot, in your robots.txt file. OpenAI’s crawler documentation states that sites opted out of OAI-SearchBot “will not be shown in ChatGPT search answers.”

OpenAI runs separate agents for separate jobs, and each robots.txt rule is independent:

User agent	Purpose	Block it?
OAI-SearchBot	Finds pages to show and link in ChatGPT search	No, if you want ChatGPT citations
GPTBot	Collects content for training OpenAI models	Your choice; blocking it does not remove you from search
ChatGPT-User	Visits a page when a user asks ChatGPT to open it	Not an automatic crawler; robots.txt may not apply

A common mistake is a security plugin or CDN rule that blocks all “AI bots” in one click. That can quietly remove you from ChatGPT search. OpenAI says robots.txt changes take about 24 hours to take effect on its side.

What kind of content gets cited in AI answers?

Content that adds something the AI cannot get from ten other pages gets cited. Google calls this “non-commodity” content: first-hand experience, original data, expert judgement and specific detail.

Google’s May 2026 guide contrasts a generic “7 Tips for First-Time Homebuyers” style article with a piece built on real experience. The first restates common knowledge. The second gives the AI, and the reader, a reason to choose that source.

Ways to make a page less “commodity”

• Use your own numbers. Average delivery times, typical project costs, response rates from your own campaigns. Only publish figures you can stand behind.
• Show the process. Screenshots, step-by-step workflows and real settings, not just a description.
• Give a clear opinion. Say which option you recommend and why, including when you would not recommend it.
• Add local or market detail. Prices in local currency, local regulations, payment methods your customers actually use.
• Name the author. A byline and a short author bio help readers see who is behind the advice. Google’s helpful content guidance encourages accurate bylines where readers expect them.
• Keep it current. Update dates, prices and screenshots. Out-of-date pages lose trust with readers and AI systems alike.

If producing this level of content in-house is the bottleneck, our content writing service builds articles around your team’s own expertise rather than rewriting what already ranks.

How should you structure a page so AI can quote it?

Put the direct answer first, then the detail. Use headings that match real questions, and use tables and lists for anything people compare.

Google says you do not need special formatting for AI. Microsoft’s Bing team is more explicit: its February 2026 announcement of the AI Performance report in Bing Webmaster Tools recommends “clear headings, tables, and FAQ sections” and “examples, data, and cited sources” to help AI systems reuse content accurately. These are also good for human readers, so there is no conflict.

A simple page template that works

1. Answer-first opening: two or three sentences that fully answer the main question.
2. Summary table: the key facts, options or steps at a glance.
3. Question-based H2s: one section per follow-up question, each opening with a one or two sentence answer.
4. Evidence: examples, figures with named sources, screenshots.
5. FAQ section: four to six short answers to the questions customers ask most.
6. Clear next step: what the reader should do now.

This is the same structure Softwix uses on its own blog, including this article.

Do you need FAQ schema or other structured data?

No, not for AI features. Google states that structured data is not required and there is no special schema for AI Overviews. Keep using structured data where it earns rich results, such as Product, Article, LocalBusiness or Breadcrumb markup, and make sure it matches what is visible on the page.

Does local and ecommerce data affect AI answers?

Yes. Google’s guide notes that AI responses can include product listings and local business information, and it points to Merchant Center feeds and Google Business Profile as the sources.

For local businesses, that means a complete, accurate Business Profile with correct categories, hours, services and photos. Our guide to setting up a Google Business Profile walks through each field. For online shops, keep product titles, prices, stock and attributes accurate in your Merchant Center feed, because AI shopping answers draw on that data.

How much traffic do AI Overview citations actually send?

Less than a traditional top ranking, in most studies so far. Being cited builds visibility and trust, but you should not expect every citation to turn into a visit.

A 2025 Pew Research Center study of 900 US adults’ Google searches in March 2025 found:

• About 18% of searches produced an AI summary.
• Users clicked a traditional result in 8% of visits when an AI summary appeared, compared with 15% when it did not.
• Users clicked a link inside the AI summary in just 1% of visits.

That is one study from one market, and behaviour is still changing. The practical lesson is to aim for citations on questions that lead to a decision, such as pricing, comparisons and “how to choose” topics, and to make sure the cited page gives readers a strong reason to click through. For a wider view of where search is heading, see our SEO trends for 2027.

How do you track whether your site is being cited?

Use the Generative AI performance report in Google Search Console and the AI Performance report in Bing Webmaster Tools. Both are free and show which of your pages appear in AI answers.

Google Search Console

Google announced the Search Generative AI performance reports in June 2026 and says they had rolled out to all websites worldwide by 31 August 2026. The reports show impressions in AI features in Search and Discover, which pages appeared, countries, devices and trends over time. AI feature traffic also remains part of your overall Performance report.

Bing Webmaster Tools

Bing’s AI Performance report shows total citations, average cited pages, the “grounding queries” the AI used when retrieving your content, and page-level citation counts. It covers Microsoft Copilot, AI summaries in Bing and selected partner integrations.

Simple monthly routine

1. Export your top cited pages from both tools.
2. Compare them with your most important commercial pages. Are your service and product pages being cited, or only blog posts?
3. Check the grounding queries in Bing to see what questions the AI was answering.
4. Improve pages that get impressions but few clicks: sharpen the opening answer and give readers a reason to visit.
5. Test a few key questions manually in Google, ChatGPT and Copilot and note which sources appear.

Be wary of tools claiming to know Google’s “internal” AI metrics. Google’s guide states plainly that no third-party tool has access to its internal ranking or AI systems.

What mistakes stop websites from being cited?

The most common problems are technical blocks and thin, generic content. Both are fixable.

• Blocking OAI-SearchBot or Googlebot through a blanket “block AI bots” setting.
• Pages set to noindex or nosnippet by accident, often on staging-to-live migrations.
• Key information hidden in tabs, images or PDFs only.
• Publishing dozens of near-identical pages targeting every query variation. Google warns this can breach its scaled content abuse policy.
• Out-of-date prices, years and screenshots.
• No clear author or business behind the content.

Frequently asked questions

Can I pay to be cited in Google AI Overviews?

No. Supporting links in AI Overviews are organic and chosen by Google’s systems from its search index. Any paid placements on the results page are labelled as ads and are separate from these citations, so the only route to being cited is the work described in this guide.

Should I create an llms.txt file?

For Google Search, it makes no difference. Google says it ignores llms.txt and similar files. Other AI services may use them, so adding one does no harm, but do not expect it to improve your Google or ChatGPT visibility on its own.

If I block GPTBot, will I disappear from ChatGPT?

Not from ChatGPT search. GPTBot is for model training, while OAI-SearchBot handles search results. OpenAI treats them independently, so you can block GPTBot and still allow OAI-SearchBot.

How long does it take to start appearing in AI Overviews?

There is no fixed timeline. Because AI Overviews use the normal search index, a page usually needs to be indexed and performing reasonably well for related searches first. Improvements to existing, already-ranking pages tend to show up faster than brand-new pages.

Is “AEO” or “GEO” different from SEO?

Google’s position is that optimising for generative AI search is still SEO. The labels describe a focus on AI visibility, but the work is the same: crawlable pages, helpful specific content, clear structure and accurate business data.

Get help with AI search visibility

If you want to know whether your site is eligible, crawlable and structured well enough to be cited by Google and ChatGPT, start with a free SEO audit. Use it to find out whether indexing, crawler access or snippet controls are holding your most important pages back, and what to fix first.