How ChatGPT, Gemini, Perplexity and Claude cite websites

In short

Each AI assistant searches before it answers, but each uses a different source: Copilot uses Bing, Gemini and Google AI Overviews use Google Search, and ChatGPT, Perplexity and Claude run their own search crawlers. To be cited, your pages must be reachable by those crawlers and easy to quote.

AI assistants find websites by searching first and answering second. Copilot searches Bing, Gemini and Google's AI Overviews use Google Search, and ChatGPT, Perplexity and Claude each run search crawlers of their own.

So "getting cited by AI" is really five separate questions. Can each assistant's crawler reach your page? Is your page in the index it searches? And once found, is your page the clearest answer? This guide takes each assistant in turn, using what the companies themselves publish, as of November 2026.

Five rows linking each assistant to where it looks things up: ChatGPT to ChatGPT search, Copilot to Bing, Gemini to Google Search, Perplexity to its own index, Claude to Claude web search. Each row ends in a box reading your page, cited.
Every assistant searches before it answers. Each one searches somewhere different.

What do all these assistants have in common?

They all follow roughly the same three steps:

  1. Turn the question into one or more search queries.
  2. Read a few of the results.
  3. Write an answer and show links to the sources it used.

The details differ: how many searches they run, how many pages they read, how they show links. But a page that is blocked or missing from the index is out of the running in all of them. See what GEO is for the bigger picture.

How does ChatGPT find and cite websites?

OpenAI runs three main crawlers, each with its own job:

  • OAI-SearchBot finds pages for ChatGPT's search features. OpenAI states that "sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers." They can still appear as plain navigation links.
  • GPTBot collects content that may be used to train OpenAI's models. Blocking it does not affect search.
  • ChatGPT-User visits a page when a person asks ChatGPT to look at it. OpenAI says robots.txt rules may not apply to these visits because a user started them.

The practical point: if you want to appear in ChatGPT answers, OAI-SearchBot must be allowed. OpenAI says changes to robots.txt take about 24 hours to reach its search systems.

How does Microsoft Copilot find websites?

Copilot uses Bing. Microsoft's help page for Copilot Chat says it "generates a short query based on your prompt and sends it to Bing." A Sources button under the answer shows the exact query and the pages used.

That makes Bing indexing the gate. If your site is not in Bing, Copilot cannot cite it, however well you rank on Google. It is worth checking Bing on its own.

Microsoft also offers a rare view of results. Bing Webmaster Tools has an AI Performance report, and in June 2026 Microsoft added reports on where your content is cited in AI answers across Copilot and Bing. If you only set up one free tool after reading this, make it that one.

How do Gemini and Google AI Overviews choose sources?

Google's AI features draw on Google Search. For AI Overviews and AI Mode, Google says a page "must be indexed and eligible to be shown in Google Search with a snippet." Both features may run many related searches at once, which Google calls query fan-out, so they can surface a wider set of pages than one classic search.

Google also says there is no special trick. You do not need new AI files, special markup or special schema. Googlebot's normal access rules are the control.

The Gemini app works a little differently. It can show a Sources button with related links under a response. Google lists "Grounding in Gemini Apps" among the uses that the Google-Extended token controls, along with training future Gemini models. Blocking Google-Extended does not affect Google Search.

How does Perplexity pick its sources?

Perplexity searches the web in real time for each question. Its help center says "each answer includes numbered citations linking to the original sources."

It runs two crawlers:

  • PerplexityBot finds and links websites in Perplexity's search results. Perplexity says it is not used to crawl content for AI foundation models. Blocking it in robots.txt keeps you out of those results.
  • Perplexity-User fetches a page when a person's question needs it. Perplexity says it "generally ignores robots.txt rules", because a user asked for the page.

How does Claude search the web?

Claude can search the web and, according to Anthropic's developer documentation, gives "a final response with cited sources." In the API version, citations are always switched on for web search.

Anthropic describes three bots:

  • Claude-SearchBot crawls to improve Claude's search results. Blocking it can reduce your visibility in those results.
  • Claude-User fetches pages when people ask Claude questions.
  • ClaudeBot collects content that may be used to train Anthropic's models.

Anthropic says all three follow robots.txt. It also warns that blocking by IP address may not work reliably, because it stops the bots from reading your robots.txt file.

How do the five compare?

AssistantWhere it searchesSearch crawler to allowTraining crawler (separate)
ChatGPTChatGPT searchOAI-SearchBotGPTBot
CopilotBingbingbotNot covered here
Gemini appGoogle SearchGooglebotGoogle-Extended (also covers Gemini app grounding)
Google AI Overviews and AI ModeGoogle SearchGooglebotNot affected by Google-Extended
PerplexityIts own indexPerplexityBotPerplexity says PerplexityBot is not used for model training
ClaudeClaude web searchClaude-SearchBotClaudeBot

All names are as of November 2026. Companies add and rename bots, so check the linked pages before editing your files. Our AI crawlers and robots.txt guide has a ready-made snippet.

What makes one page more likely to be cited than another?

None of these companies publish their ranking rules, so be wary of anyone who claims to know them. What we can say from the companies' own guidance and from the GEO research paper is modest:

  • Be findable. Allowed crawlers, indexed pages, text in the HTML.
  • Be quotable. A direct answer near the top, headings that match real questions, facts in text and tables.
  • Be believable. A named business or author, contact details, dates and links to sources.

Our view is that the third point matters more than most guides admit. An assistant has many pages to choose from. A page that shows who stands behind it gives it one more reason to pick you.

How can I check where I stand?

Ask each assistant a question your customers would ask, and note which sites it cites. Then check the technical side: our 15-minute checklist covers it by hand. Our free scanner checks crawler access for the bots above and how quotable your pages are. It measures readiness and cannot see actual citations; the methodology page lists every check.

Frequently asked questions

Does ranking on Google help me in ChatGPT?

Indirectly at best. ChatGPT search relies on OAI-SearchBot, and Copilot relies on Bing. Good pages tend to do well everywhere, but each assistant needs its own crawler to be allowed.

Why does the same question give different sources each time?

These tools rewrite questions into searches, and the searches and answers can vary between attempts. One answer tells you little. Ask several times and look for patterns.

Can I pay to be cited by an AI assistant?

None of the five companies' crawler or help pages we cite describe a paid way into their answers' sources. Be careful with anyone selling guaranteed AI citations.

Should I submit my site to Bing?

Yes, if you want to appear in Copilot. Bing Webmaster Tools is free, and its AI Performance report is one of the few official sources of citation data.

Sources

  1. OpenAI: Overview of OpenAI crawlers
  2. Microsoft Support: How web search works in Microsoft Copilot Chat and agents
  3. Bing Blogs: New AI Visibility Insights in Bing Webmaster Tools (June 2026)
  4. Google Search Central: AI features and your website
  5. Google: Google's common crawlers (Google-Extended)
  6. Gemini Apps Help: View related sources from Gemini Apps
  7. Perplexity: Perplexity crawlers
  8. Perplexity Help Center: How does Perplexity work?
  9. Anthropic: Does Anthropic crawl data from the web, and how can site owners block the crawler?
  10. GEO: Generative Engine Optimization (Aggarwal et al., arXiv 2311.09735)
  11. Claude Platform Docs: Web search tool