To get cited by ChatGPT, Perplexity, Gemini and the other AI engines, do three things that apply to all of them, then check each engine on its own terms. First, let each engine’s search crawler reach your pages, because each one builds or queries its own index: OAI-SearchBot for ChatGPT, PerplexityBot for Perplexity, Googlebot for Google’s AI features and the Gemini app. Second, publish pages whose passages answer real questions with specific, checkable facts. Third, make sure other sites describe you accurately, because engines often answer questions about you from reviews, directories and comparison articles. None of this guarantees a citation: no engine publishes its ranking factors, and answers change from day to day. It removes the reasons an engine would pass you over.

This page is the overview. Each engine has its own guide with the detail: ChatGPT, Perplexity, Google AI Overviews and AI Mode, Claude, Grok and Meta AI. For the research behind GEO, see What Is Generative Engine Optimization?

Where each AI engine gets its sources

Every row comes from the vendor’s own documentation, checked November 10, 2026.

EngineWhere web answers come fromWhat to allowHow sources appearReporting for site owners
ChatGPTIts own search, which “sometimes partners with other search providers”OAI-SearchBot and its published IP rangesInline citations and a Sources buttonNone
Google AI Overviews and AI ModeGoogle’s Search index, with “query fan-out” (several related searches per question)Googlebot; page indexed and snippet-eligible; Search Console’s Search generative AI setting left on Include (the default)Supporting linksSearch Console Generative AI performance report (impressions)
Gemini app”Some Gemini responses are grounded on Search results”Googlebot, and don’t block Google-ExtendedSources button or inline links, when availableNone specific
PerplexityIts own index of over 200 billion URLsPerplexityBot and its published IP rangesNumbered citations on every answerNone
ClaudeA web search tool, plus web fetch for links users shareClaude-SearchBot and Claude-UserCitations on every search-based answerNone
Microsoft CopilotBing’s indexBingbot; avoid NOARCHIVE and NOCACHE (being replaced by AI Training and AI Grounding preferences from February 2027)Cited sources in the answerBing Webmaster Tools AI Performance report (citations)
GrokPublic X posts, a real-time web search, or both, decided per questionA public X account, plus crawlable pagesWeb URLs and x.com postsNone
Meta AIMeta’s own index, among other inputsMeta-WebIndexerLinks to cited contentNone

Three things stand out. Each engine has its own crawler or index, so access has to be checked engine by engine. Only Google and Microsoft give site owners first-party reports, and only Bing’s counts citations. And the Gemini app answers to a different control from AI Overviews, which is where most cross-engine advice goes wrong.

One robots.txt for every engine

Most AI companies separate three jobs: crawling for search (which earns citations), crawling for model training, and fetching a page because a user asked. Each is controlled separately, so you can be citable everywhere and still opt out of training.

CompanySearch crawler (needed for citations)Training crawler (optional)User-requested fetches
OpenAIOAI-SearchBotGPTBotChatGPT-User
PerplexityPerplexityBotNone (PerplexityBot isn’t used for foundation models)Perplexity-User
AnthropicClaude-SearchBotClaudeBotClaude-User
GoogleGooglebotGoogle-Extended (a robots.txt token, not a separate crawler; also controls Gemini app grounding)n/a
MicrosoftBingbotn/an/a
MetaMeta-WebIndexerMeta-ExternalAgent (training and product indexing)Meta-ExternalFetcher

A file that allows every search crawler and opts out of training looks like this:

User-agent: OAI-SearchBot
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: Claude-SearchBot
Allow: /

User-agent: Claude-User
Allow: /

User-agent: meta-webindexer
Allow: /

User-agent: GPTBot
Disallow: /

User-agent: ClaudeBot
Disallow: /

User-agent: meta-externalagent
Disallow: /

Meta doesn’t say whether blocking meta-externalagent affects Meta AI answers; see How to Get Cited by Meta AI.

If nothing blocks a crawler, it’s already allowed; the explicit lines just document the decision. Google-Extended is left out on purpose, because blocking it is a trade-off (see the Gemini section). Two details matter:

  • Your CDN can overrule your robots.txt. OpenAI says that to be eligible, you should allow OAI-SearchBot “and confirm that the website host or content delivery network allows traffic from OpenAI’s published searchbot IP addresses.” Perplexity publishes IP ranges and Cloudflare and AWS WAF rules for the same reason. Bot protection that challenges a crawler blocks it as surely as a Disallow line.
  • Changes aren’t instant. OpenAI says “it can take ~24 hours from a site’s robots.txt update for our systems to adjust.”

A shared training opt-out is coming. On October 30, 2026, Google said it plans to support a single robots.txt signal, based on an IETF draft, that opts a site out of AI training across providers, aiming for 2027 (Google Search Central). Anthropic says it began recognizing the same signal that day, and on the same date Microsoft said it will honor it from February 1, 2027 and OpenAI said it will do so “in the coming months.” Google will read that opt-out as disallowing Google-Extended, which also removes you from Gemini app grounding. If you use it and still want Gemini to cite you, allow Google-Extended explicitly: Google says sites that do “are eligible to appear in responses and receive clicks via grounding in Gemini Apps.”

ChatGPT

Older guides say ChatGPT search runs on Bing. OpenAI no longer says that. Its help article says ChatGPT search “sometimes partners with other search providers,” rewriting your question into targeted queries it sends them, and links to Microsoft’s and Shopify’s privacy policies for how those providers handle queries. OpenAI also runs its own search crawler, and its crawler page says sites opted out of OAI-SearchBot “will not be shown in ChatGPT search answers, though can still appear as navigational links.” Blocking GPTBot only opts you out of training. Optimizing for Bing alone isn’t enough; ChatGPT needs to read your pages itself. Ranking overlap with Google, shopping, ads and referral tracking are in How to Get Cited by ChatGPT.

Perplexity

Perplexity runs its own search engine. Its crawler documentation says PerplexityBot exists “to surface and link websites in search results on Perplexity,” and its help center says each answer includes numbered citations to the original sources. It also ranks pieces of pages: in a September 2025 research post, its search team described splitting documents into “self-contained spans, each of which can be individually retrieved and ranked at query time.” Self-contained, answer-first paragraphs matter most here. See How to Rank in Perplexity AI.

Gemini: the app is not AI Overviews

Google runs two kinds of AI surface that people loosely call “Gemini,” and they follow different rules.

AI Overviews and AI Mode live inside Google Search. To be a supporting link, Google says a page must be “indexed and eligible to be shown in Google Search with a snippet,” with “no additional technical requirements” and no special schema or AI text files (AI features and your website). The site must also stay included in Search Console’s Search generative AI setting, which is the default; excluding it removes you from AI Overviews and AI Mode. Google-Extended has no effect on them. Details are in How to Optimize for Google AI Overviews and AI Mode.

The Gemini app (gemini.google.com and the mobile apps) is a separate product with more than a billion monthly active users, per Alphabet’s Q3 2026 earnings remarks on October 28, 2026. Google’s Gemini documentation says three things site owners need to know:

  • It sometimes grounds answers in Search. The Gemini Apps Privacy Hub (updated October 26, 2026) says “some Gemini responses are grounded on Search results,” and Deep Research includes Google Search as a source by default (Gemini Apps Help). So Googlebot access, indexing and clear, snippet-eligible text matter here too.
  • Sources aren’t always shown. Gemini “may provide links to sources and related content,” under a Sources button or inline, but “not all responses include related links or sources” (Gemini Apps Help). If Gemini quotes a large amount of text from a page, that page appears in the sources list.
  • Google-Extended controls grounding as well as training. Google’s crawler documentation says the token manages whether your content may be used for “training future generations of Gemini models” and for “grounding (providing content from the Google Search index to the model at prompt time to improve factuality and relevancy)” in Gemini Apps and Vertex AI. It “does not impact a site’s inclusion in Google Search nor is it used as a ranking signal.” Google’s Gemini app overview confirms that for opted-out URLs, “Gemini will also not use such content for grounding.”

So Google-Extended is a real trade-off. Disallow it and you opt out of Gemini training, but you also drop out of the Gemini app’s grounded answers, while AI Overviews and AI Mode are unaffected. There’s no way to block one and keep the other. If Gemini app citations matter to you, leave it allowed.

Expect Gemini to link less consistently than Google Search. In Ahrefs’ October 2026 stability study, which asked five AI tools the same questions daily for 20 days, Gemini’s citations churned far more than AI Overviews’ (figures below), partly because more than half of Gemini’s answers cited no sources at all. Of the brands Gemini both cited and named, 40% lost the link the next day but were still named. On Gemini, track whether you’re named, not just linked.

Claude, Copilot, Grok and Meta AI

Claude. Anthropic’s help center says that when Claude searches, “every response includes citations,” and its crawler page says disabling Claude-SearchBot “may reduce your site’s visibility and accuracy in user search results.” See How to Get Cited by Claude.

Microsoft Copilot. Bing’s webmaster guidelines say Bing and Copilot “rely on the same core crawling, indexing, and ranking foundation as traditional search,” that NOARCHIVE “prevents content from being used in Copilot responses,” and that NOCACHE “limits Copilot to using only the URL, title, and snippet.” On October 30, 2026, Microsoft said separate AI Training and AI Grounding preferences will replace both for these uses: it honors the new preferences from February 1, 2027 and stops reading NOARCHIVE and NOCACHE as training or grounding preferences after March 1, 2027. Bing Webmaster Tools’ AI Performance report, in public preview since February 10, 2026, counts how often your pages are cited in Copilot and Bing’s AI summaries, the only vendor report that counts citations.

Grok. X’s help center says Grok decides for each question “whether or not to search X public posts and conduct a real-time web search,” and that a private account’s posts aren’t surfaced in answers. xAI’s citations documentation shows X posts cited by x.com address alongside web pages. See How to Get Cited by Grok.

Meta AI. Meta’s crawler documentation says allowing Meta-WebIndexer “helps us cite and link to your content in Meta AI’s responses.” Its other inputs are covered in How to Get Cited by Meta AI.

What to publish: the same for every engine

Access gets you considered; content gets you cited. Here the engines’ guidance agrees more than it differs.

  • Answer first. Put the direct answer in the first sentence or two under each heading, so a passage can be quoted on its own. You don’t need to go further: Google’s optimization guide says there’s “no requirement to break your content into tiny pieces” and no ideal page length.
  • Be specific and sourced. Numbers, dates, prices and named sources give an engine something checkable to quote. The original GEO paper (Aggarwal et al., Princeton and IIT Delhi, KDD 2024) found that adding statistics, quotations or citations to a source raised its visibility in generated answers by up to 40% on its benchmark, while keyword stuffing didn’t help. It was a controlled test, not a measurement of live engines; its limits are in the GEO guide.
  • Say something others don’t. Google’s guide contrasts commodity content (“7 Tips for First-Time Homebuyers”) with first-hand expert content, and says unique, useful content “will likely influence your website’s presence in generative AI search in the long run more than any of the other suggestions in this guide.” Original data, real prices and first-hand testing are what a summary has to cite you for.
  • Keep your facts identical everywhere: the same name, description and key facts on your site, LinkedIn, directories and review sites. See Entity SEO for AI Knowledge Graphs.
  • Own the questions about you. In the Ahrefs stability study, when a question named a brand, that brand’s site stayed cited all 20 days in 28–33% of cases on AI Overviews, AI Mode and ChatGPT, against 4–7% for the average cited site. Publish clear pages on your pricing, features, policies and alternatives.
  • Be described well elsewhere. In the same study’s 22 questions about one brand (Ahrefs itself), Google’s AI tools cited the brand’s site for factual questions but leaned on third-party sites for reviews and comparisons. Earn accurate listings, genuine reviews and coverage; don’t plant them. See Review Management for AI Visibility.

What you can skip. Schema markup isn’t a citation lever. Google says no special schema is needed for its AI features, and in a May 2026 study of 1,885 pages that added JSON-LD against 4,000 matched controls, Ahrefs found no citation gain in AI Mode or ChatGPT and a small AI Overviews decline it couldn’t clearly attribute to schema (the pages were already heavily cited, so it says less about pages that aren’t). FAQ rich results stopped showing on May 7, 2026, and HowTo rich results were retired in 2023. Keep markup accurate for the rich results that remain (structured data guide). Google Search also ignores llms.txt. The AI Citation Readiness Checklist turns all this into a page-by-page audit.

Why engines cite different sources

Being cited by one engine says little about the others. Using June 2025 data from about 76.7 million AI Overviews, 957,000 ChatGPT prompts and 953,500 Perplexity prompts, Ahrefs found only 7 websites in the top 50 most-mentioned sites on all three (Ahrefs, June 12, 2025). The data is old now, but the reasons hold: different indexes, query rewriting and source preferences.

Citations also move within one engine. In the Ahrefs stability study (October 29, 2026; about 37,500 citations tracked daily for 20 days), the share of day-one cited sites gone the next day was 50% on AI Overviews, 58% on AI Mode, 59% on ChatGPT, 81% on Gemini and 83% on Copilot, and about 75% of dropped citations came back within the 20 days. One missing citation isn’t a reason to rewrite a page; two or three weeks without it is worth a look.

How to measure citations across engines

Use the same three numbers on every engine, defined as in How to Measure AI Citations: appearance rate (the share of answers that name you), citation rate (the share that link to your site) and share of AI voice (your share of brand mentions across the competitors you track). Write 20–50 questions your buyers ask, run each more than once per engine, and report each engine separately, because the gaps point to different causes: a blocked crawler on one, missing third-party coverage on another, a Google-Extended block that only touches Gemini.

Then add the first-party data:

  • Google Search Console: the Generative AI performance report, open to every site since August 31, 2026, shows impressions from AI Overviews and AI Mode. It doesn’t cover the Gemini app.
  • Bing Webmaster Tools: the AI Performance report counts Copilot citations.
  • Google Analytics 4: the AI Assistant channel groups visits “from sources like ChatGPT, Gemini, Deepseek, Copilot, or Grok” and excludes AI Overviews and AI Mode, which count as Organic Search. Check Referral for perplexity.ai. OpenAI’s publisher FAQ also said ChatGPT adds utm_source=chatgpt.com to referral URLs (archived August 27, 2026; page since retired).
  • Server logs: visits from OAI-SearchBot, PerplexityBot, Claude-SearchBot, Googlebot, Bingbot and Meta-WebIndexer show which engines crawl which pages.

What to do this week

  • Check robots.txt and your CDN or firewall rules for every search crawler above, and decide deliberately about Google-Extended, and confirm Search Console’s Search generative AI setting is on Include.
  • Confirm in your server logs that each crawler gets 200 responses on your key pages.
  • Verify your site in Bing Webmaster Tools and open Search Console’s Generative AI performance report.
  • Run 20 buyer questions three times each on ChatGPT, Perplexity, the Gemini app and Google’s AI Mode, logging who is named and who is linked.
  • Rewrite the opening of your three most important pages so the first two sentences answer the main question with a specific, sourced fact.

Prefer to self-assess first? Download the 25-item AI Citation Readiness Scorecard (PDF), a printable self-assessment across schema, content architecture, authority, measurement and conversion. Free, instant download, no email required.

Want it done for you? The free AI Visibility Snapshot tests 10 prompts on ChatGPT, Google AI Overviews and Perplexity, with a short note on where you’re cited and where you’re not. The Citation Audit tests 30–50 locked prompts across six surfaces (ChatGPT, Google AI Overviews, Perplexity, Gemini, Claude and the OpenAI Responses API) and delivers competitor gaps, schema and content fixes and a prioritized 60–90 day roadmap in 5 business days. It’s $1,500, fully credited toward an Implementation Sprint or retainer signed within 30 days. We don’t guarantee citations or rankings. See pricing.

Request a free Snapshot or a Citation Audit →


FAQ

How do I get my brand cited by ChatGPT, Perplexity and Gemini?

Let each engine’s search crawler reach your site (OAI-SearchBot for ChatGPT, PerplexityBot for Perplexity, Googlebot for Gemini and Google’s AI features) and don’t block Google-Extended if you want Gemini app citations. Publish pages that answer your buyers’ questions with specific, sourced facts, make sure other sites describe you accurately, and measure each engine separately. No engine guarantees citations.

Does ChatGPT use Bing?

OpenAI no longer says so. Its help article says ChatGPT search “sometimes partners with other search providers” and links to Microsoft’s and Shopify’s privacy policies. OpenAI also runs its own search crawler, OAI-SearchBot, and says sites that block it won’t be shown in ChatGPT search answers, so optimizing for Bing alone isn’t enough.

Is the Gemini app the same as Google AI Overviews?

No. AI Overviews and AI Mode are part of Google Search and need only a page that’s indexed, eligible for a snippet, and not excluded in Search Console’s generative AI setting. The Gemini app is a separate product that grounds some answers in Google Search results. Blocking Google-Extended removes your content from Gemini app grounding and Gemini training, but doesn’t affect AI Overviews, AI Mode or Search rankings.

Should I block Google-Extended?

Only if opting out of Gemini model training matters more to you than being cited in the Gemini app, because Google uses the same token for both. It has no effect on AI Overviews, AI Mode or Google Search. If you adopt the shared “no AI training” signal Google plans to support, Google will treat it as disallowing Google-Extended unless you allow Google-Extended explicitly.

Does schema markup help me get cited by AI engines?

Not measurably, on current evidence. Google says no special schema is needed for its AI features, and Ahrefs’ May 2026 study of 1,885 pages that added JSON-LD found no citation gain in AI Mode or ChatGPT against matched controls. Keep structured data accurate for the rich results that remain.

Why did my AI citation disappear?

Often for no lasting reason. In Ahrefs’ October 2026 study, half or more of day-one citations were gone the next day on every engine tracked, and about 75% of those came back within 20 days. Judge visibility over many runs, and look at a page only if its citation stays gone for two or three weeks.

Sources

Related reading: How to Get Cited by ChatGPT, How to Rank in Perplexity AI, How to Optimize for Google AI Overviews and AI Mode, How to Get Cited by Claude, How to Get Cited by Grok, How to Get Cited by Meta AI, How to Measure AI Citations.