Home

›

›

How to Get Cited by ChatGPT and Other AI Assistants

How to Get Cited by ChatGPT and Other AI Assistants

A document card with a green check being picked up by a mechanical hand

Short answer

To get cited by ChatGPT, your pages first have to be reachable by OpenAI’s search crawler, OAI-SearchBot, and then they have to be the clearest, most specific answer to the questions people ask. OpenAI says any public website can appear in ChatGPT search if it does not block that crawler. Beyond access, the pages that get cited tend to answer directly, carry facts others do not have, stay current and are mentioned on other sites ChatGPT already trusts.

Key takeaways

  • OpenAI says sites that block OAI-SearchBot are not shown in ChatGPT search answers. Blocking GPTBot, its training crawler, is a separate choice.
  • OpenAI has said ChatGPT search uses third-party search providers, so being well indexed in the wider web, Bing included, helps.
  • In Ahrefs’ study of ChatGPT’s 1,000 most-cited pages, blog posts made up only 1.9%. Homepages, Wikipedia and educational pages dominated.
  • The same study found 28.3% of cited pages had no organic keywords, and most pages with an update date had been updated that year.
  • ChatGPT adds utm_source=chatgpt.com to the links it sends, so its referrals can be measured. We track them on UpRank as part of a public experiment.

Step 1: Let OpenAI’s search crawler in

OpenAI runs several crawlers, and they do different jobs:

CrawlerWhat it doesWhat blocking it means
OAI-SearchBotFinds pages for ChatGPT’s search featuresYour site is not shown in ChatGPT search answers
GPTBotCollects content that may be used to train modelsYour content should not be used for training
ChatGPT-UserVisits a page when a user asks ChatGPT toNot used to decide what appears in search; robots.txt may not apply

OpenAI’s documentation is direct: “Sites that are opted out of OAI-SearchBot will not be shown in ChatGPT search answers.” It recommends allowing OAI-SearchBot in robots.txt and notes that changes can take about 24 hours to reach its systems. Perplexity works the same way with PerplexityBot, and Google needs Googlebot to crawl and index a page before it can appear in AI Overviews.

Check your robots.txt for each of these crawlers, and check any firewall or bot protection too, since some block crawlers before robots.txt is even read. If you are new to the file itself, where it lives and what happens when it is missing, our guide to SEO for startups explains it step by step.

Step 2: Be indexed in the wider web

ChatGPT does not rely only on its own crawl. OpenAI has confirmed that its search model uses third-party search providers, without saying how much each contributes, and reporting since 2024 has found its results closely match Bing’s. In practice, a page that is well indexed in Google and Bing has more ways to be found.

Submit a sitemap to Google Search Console and Bing Webmaster Tools, fix pages that are excluded from the index, and keep one clean URL for each page.

Step 3: Answer the question in the first lines

Answer engines quote passages, not whole pages. A page that buries its answer under an introduction is harder to quote than one that states the answer at the top.

Write the opening for the question you want to be cited for. Two or three sentences that answer it completely, with the key number or fact included, and that still make sense if they are the only part a reader sees. Then give the detail, the reasoning and the exceptions.

Step 4: Publish facts nobody else has

The Ahrefs study of the 1,000 pages ChatGPT cited most in September 2025 shows where citations go. Wikipedia made up 29.7% of them, homepages and landing pages 23.8% and educational pages 19.4%. Blog posts and articles were only 1.9%.

That does not mean articles cannot be cited. It means the articles that are cited have to be worth citing over a reference page. The most reliable way is to publish something that exists nowhere else: your own data, a test you ran, a price you checked at its source. A page that repeats what ten others say gives an answer engine no reason to choose it.

Step 5: Get mentioned where ChatGPT already looks

The same Ahrefs study concluded that roughly two-thirds of ChatGPT’s top citations were “effectively off-limits” to outreach: Wikipedia, homepages of large brands and app store pages. The route it recommends is to be mentioned in the content that can be influenced, such as publications, review sites and blogs that answer engines already cite.

For a small site, that means being part of the conversation in your field: listed accurately in relevant directories, quoted in articles on your subject, and described consistently wherever your name appears.

Step 6: Keep pages current

In the Ahrefs study, update dates were found on just over half of the cited pages, and 89.7% of those had been updated in 2025. The study’s own reading was that ChatGPT appears to lean toward fresh content.

Show a visible update date, and update it only when you actually change something: a price, a rule, a figure. For questions about costs and limits, an old page is a wrong page.

Step 7: Make it clear who you are

An answer engine has to decide whether a source is credible. Help it. State who publishes the site, what it does and how it works, on an about page and in structured data. Name your sources and date them. Explain how you use AI and where you have an interest, as our editorial policy does.

Step 8: Measure what happens

You cannot improve what you do not measure, and AI citations are harder to see than rankings:

  • Referrals. OpenAI says ChatGPT “automatically includes the UTM parameter utm_source=chatgpt.com in referral URLs.” Any analytics tool can count those visits. Perplexity and other assistants show up as referring domains.
  • Google. Traffic from AI Overviews and AI Mode is counted inside Search Console’s normal Performance report, not separately.
  • Copilot. Bing has been testing an AI Performance report that shows how often Copilot cites your pages, though not the clicks.
  • A question log. Keep a fixed list of the questions you want to be cited for, ask them on a regular schedule, and record which sources appear.

How we are doing it on UpRank

This blog is part of the UpRank experiment: a test of whether useful content and answer engine optimization can bring a pay-to-rank board the attention that hype once did. Here is how each step applies:

  1. Crawlers. OAI-SearchBot and PerplexityBot are allowed in alongside Googlebot and Bingbot. The blog lives under the board’s domain, so the board’s robots.txt governs it, as our guide to SEO for startups explains.
  2. Index. Each page has one URL, category pages redirect to their guide, and a blog map links every article.
  3. Answer first. Every article opens with a short answer box written to be quoted on its own.
  4. Facts nobody else has. Our UpRank data page publishes what listings on the board paid and the clicks they received: $966 paid and 18,583 recorded clicks across 57 listings on October 12, 2026.
  5. Mentions. Getting UpRank described accurately in the directories and publications of its field is part of the plan, using the approach in our guide to directory submission. Each new mention goes into the experiment log.
  6. Freshness. Every source carries the date it was checked, and figures are updated when they change.
  7. Identity. The site describes UpRank in structured data, and our editorial policy declares our interest in our own board.
  8. Measurement. We track ChatGPT referrals through utm_source=chatgpt.com, AI and search traffic in Search Console and Bing, and a fixed set of questions such as “how much does a click cost on a pay-to-rank board” and “what does it cost to promote a startup”. Results are added to the experiment log with their dates.

For the background, read what AEO is, how AEO differs from SEO and our guide to organic visibility.

Frequently asked questions

Can I pay to be cited by ChatGPT?

OpenAI’s publisher guidance describes no paid placement in ChatGPT search answers. It says any public website can appear if it does not block OAI-SearchBot.

Does blocking GPTBot remove my site from ChatGPT search?

No. GPTBot collects content for model training. ChatGPT search depends on OAI-SearchBot, which is controlled separately in robots.txt.

How do I know if ChatGPT is sending me traffic?

OpenAI says ChatGPT adds utm_source=chatgpt.com to the links it sends. Filter your analytics by that source to count the visits.

Why does ChatGPT cite some small sites and not others?

OpenAI does not publish a ranking formula. Studies show cited pages tend to be current, specific and often unique, and 28.3% of the most-cited pages in one study had no organic keywords, so search rankings are not the only route.

How long does it take to get cited?

There is no published timeline. OpenAI says robots.txt changes take about 24 hours to reach its search systems, but being chosen as a source depends on the page and the question. We log our own timeline on the UpRank experiment page.

The board

Put your product on UpRank

Your rank is your lifetime total, the cheapest spot is $1, and every listing shows the clicks it receives. Check what a spot costs today before you decide.

Felix Archer Avatar