How to get cited by ChatGPT
Published by Krister Ross · Updated May 2026
When ChatGPT answers a question, it names a handful of sources. This guide explains how it chooses them and the concrete work that puts your brand among the few it cites.
The 30-second answer
ChatGPT cites sources when it browses the web, and its browsing runs on the Bing search index. To get cited, give it a clear, extractable answer near the top of a page, mark it up with structured data, keep your brand identity consistent everywhere, earn mentions on sites Bing already trusts, and keep the page fresh. You cannot pay ChatGPT for a citation. Influence is indirect, through authority, structure and trust.
Read the complete AI visibility guide for the broader picture.
How ChatGPT chooses what it cites
ChatGPT cites sources when it uses web browsing to answer a question. Its browsing retrieves results from the Bing search index, so the pages it can reach and cite are largely the pages Bing surfaces well. Strong traditional search visibility for a query is the entry ticket to being available as a source.
From the candidate pages, the model favours a few things. Recency matters for anything time-sensitive, so a recently updated page beats a stale one. Trust signals matter, so a clear author, a real organisation behind the page and mentions from established sources all help. Structured data helps the model read the page reliably. And clear, extractable answers matter most, because the model is looking for a passage it can lift and attribute, not a page it has to interpret.
So the practical question is simple. Does Bing surface your page for the query, and once the model lands there, can it find a clean answer it trusts enough to quote?
ChatGPT is not the only answer engine
Each assistant retrieves and cites a little differently. Getting cited by ChatGPT is the start, not the finish, because your customers ask more than one of them.
ChatGPT
Browses via the Bing index
Favours pages Bing surfaces well, with a clear extractable answer and visible authorship.
Google AI Overview
Built on Google's own index
Pulls from pages that already rank in Google, so classic SEO and clean structure feed it directly.
Perplexity
Crawls and cites the open web
Leans on fresh, source-rich pages and shows its citations openly, rewarding genuinely referenceable content.
Gemini
Uses Google Search grounding
Grounds answers in Google results, so entity clarity and search visibility decide whether it reaches for you.
What a citable answer looks like
The single biggest lever is giving the model a clean passage it can lift and attribute. The difference is rarely the facts, it is the shape.
Hard to cite
There are many factors involved when it comes to how AI tools decide what to show, and over the years we have seen a lot of change in this space, which is why it pays to think holistically about your whole presence before drawing any conclusions.
Easy to cite
ChatGPT cites a source when three things are true: the page ranks in Bing for the question, the answer sits in the first paragraph in plain language, and the page shows a named author and organisation it can verify.
Seven concrete steps to get cited
A practical order of work. Each step makes your page easier to find, read and trust.
- 1
Write a clear, short answer to the question high on the page, in plain language the model can lift directly.
- 2
Add structured data so the model reads the page reliably, starting with Organization, Article and FAQPage.
- 3
Keep your entity consistent everywhere, the same name, category and location on your site, in Wikidata and in industry directories.
- 4
Earn mentions and links on sources Bing already trusts, since the model reaches you through pages it can find.
- 5
Keep pages fresh by updating facts and dates, because recency is a real ranking factor for time-sensitive questions.
- 6
Add FAQ and HowTo schema where the content fits a question-and-answer or step-by-step shape.
- 7
Measure citations and sentiment over time, then adjust the pages and topics that underperform.
Schema examples that help
Structured data tells the model what a page is in a language it reads directly. Two of the most useful types for being cited, as JSON-LD.
FAQPage
{
"@context": "https://schema.org",
"@type": "FAQPage",
"mainEntity": [{
"@type": "Question",
"name": "Does ChatGPT cite sources?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Yes, ChatGPT cites sources when it browses the web to answer a question."
}
}]
}Article
{
"@context": "https://schema.org",
"@type": "Article",
"headline": "How to get cited by ChatGPT",
"author": { "@type": "Person", "name": "Krister Ross" },
"publisher": { "@type": "Organization", "name": "CitationLab" },
"dateModified": "2026-05-27"
}How to measure it
You cannot improve what you cannot see. The honest measure of progress is whether ChatGPT actually names you, and for which questions, tracked over time rather than checked once.
CitationLab AI Monitor tracks your citations and the sentiment around them across the major models, so you can see which pages get picked up, which questions you still lose, and whether a change moved anything.
Common mistakes
The patterns that quietly keep brands out of the answer.
Burying the answer under introductions, so the model never reaches a clean passage to quote.
Writing for clicks instead of extraction, with vague copy the model cannot lift with confidence.
Inconsistent brand identity across the site, Wikidata and directories, which weakens the entity the model relies on.
Letting pages go stale, then wondering why a competitor with a freshly updated page gets cited instead.
Trying to pay or trick your way in, when influence only comes from authority, structure and trust.
Frequently asked questions about getting cited by ChatGPT
The questions we hear most often.
Yes. When ChatGPT browses the web to answer a question, it cites the sources it used. Its browsing retrieves results from the Bing search index, so the pages it can cite are largely the ones Bing surfaces well.
Usually because the model cannot find or trust your page for the question. Either Bing does not surface it well, the answer is buried or vague, your entity is inconsistent across the web, or the page is out of date compared to competitors.
For browsing-based answers, improvements can show within weeks once Bing indexes an updated, well-structured page. Building enough authority for the model to reach for you consistently usually takes a few months.
No. There is no paid placement for citations. Influence is indirect and comes from authority, clear structure and trust signals, the same things that make a page worth quoting.
It builds on SEO. Strong search visibility makes your page reachable, since browsing runs on a search index. The added work is making the answer extractable, the entity consistent and the trust signals clear so the model is willing to quote and attribute you.