---
title: "How to Get Cited by ChatGPT: A Practical ChatGPT SEO Guide"
date: 2026-07-09T06:00:00Z
modified: 2026-09-15T17:32:21Z
permalink: "https://suparank.io/get-cited-by-chatgpt/"
type: post
status: publish
excerpt: How ChatGPT search picks the sources it cites, the seven page traits citation studies keep finding, the robots.txt rules that matter, and a repeatable test to check your citations.
wpid: 26
categories:
  - Get Cited by AI
tags:
  - Get Cited by AI
  - AEO
  - AI Citations
  - ChatGPT
  - Schema Markup
author: Chase Allen
timestamp: 2026-09-15T17:32:21Z
---

To get cited by ChatGPT, a page has to be found by ChatGPT search, picked from the results it retrieves, and quoted. In practice that means letting OAI-SearchBot crawl you, answering the question in the first sentence under a matching heading, adding facts only you publish, and being mentioned on other sites that ChatGPT already trusts.

This guide covers how that pipeline works, the seven page traits that published citation studies keep finding, the robots.txt rules that decide whether you are eligible at all, and a repeatable test to check whether ChatGPT cites you today.

## How ChatGPT search finds the sources it cites

ChatGPT answers from two different places. Its training data is a fixed snapshot collected in the past, which you cannot change after the fact. Its search feature is live: for prompts that need current or specific facts, it runs web searches and cites the pages it uses. Citations come from the second path, so that is the one to optimize.

The search path runs in roughly four steps:

1. **Fan-out.** ChatGPT rewrites the prompt into several narrower sub-queries, each hunting for one fact.
2. **Retrieval.** Each sub-query returns a list of results carrying a title, a URL and sometimes a snippet. Ahrefs found that 88% of the URLs ChatGPT ends up citing come from this general search channel [1](#sprk-src-1).
3. **Selection.** ChatGPT reads the title, URL and snippet first, then opens some pages. Across 1.4 million prompts, it cited only about 50% of the URLs it retrieved [2](#sprk-src-1).
4. **Synthesis.** The model writes one answer and attaches citations to the passages it relied on.

Which index feeds retrieval matters for where you do technical SEO. In a 2025 analysis, Seer Interactive found that 87% of SearchGPT citations matched Bing’s top results [3](#sprk-src-2). Submitting your sitemap to Bing Webmaster Tools is a cheap way to make sure that path can see your pages.

## Can you rank in ChatGPT?

Not in the way you rank in Google. There is no stable list of ten positions: every answer is generated fresh, and the same prompt can cite different pages from one run to the next. What you can measure is whether you are included, and how.



| Outcome | What it looks like | What it tells you |
| --- | --- | --- |
| Mention | Your brand is named in the answer text | ChatGPT associates you with the topic |
| Citation | Your URL is linked as a source | A specific page of yours was retrieved and used |
| Position | Does not exist | Order within one answer changes run to run, so it is not a durable metric |

Mentions and citations are the two things you can track in ChatGPT; positions are not.So “how to rank in ChatGPT” really means raising the share of relevant prompts where you are mentioned or cited. That is the same idea behind [AEO vs GEO](/aeo-vs-geo-vs-seo/): the unit of success is inclusion in an answer, not a slot on a page.

## 7 traits of pages ChatGPT quotes

These come from two public studies: a Search Engine Land audit of 15 domains with nearly 2 million monthly organic sessions and 7,500 ChatGPT referral sessions, and Ahrefs’ analysis of 1.4 million ChatGPT prompts. Both are correlational, so treat them as strong patterns, not guarantees.

### 1. An answer capsule under a question heading

Search Engine Land defines an answer capsule as a self-contained explanation of roughly 120 to 150 characters (about 20 to 25 words) placed directly after a title or a question-style H2. In its audit, 72.4% of blog posts cited by ChatGPT had one, making it the single most consistent trait [4](#sprk-src-3).

Write the capsule so it still makes sense if it is the only sentence quoted. Put the context and nuance in the paragraphs that follow. An illustrative before and after for a question heading “Does blocking GPTBot remove my site from ChatGPT search?”:



| Text under the heading |  |
| --- | --- |
| Before | Great question! OpenAI runs several crawlers, and robots.txt can get complicated, so let’s walk through the options one by one. |
| After | No. GPTBot collects content for model training, while OAI-SearchBot decides whether your pages can appear in ChatGPT search answers. |

Illustrative example, not taken from the study. The after version is a quotable capsule; the before version gives ChatGPT nothing to cite.### 2. Facts that only you publish

Original or owned data was the second-strongest differentiator: 52.2% of cited posts had original data or a branded insight, and 34.3% combined it with an answer capsule, the best-performing mix in the audit [5](#sprk-src-3). A survey result, a benchmark from your own product, or a clearly labelled recommendation gives ChatGPT a reason to attribute the fact to you rather than to the tenth page repeating it.

### 3. A title and URL that match the sub-questions

Because ChatGPT screens results on title and URL before it opens a page, those two fields do early work. Ahrefs found that cited titles were more semantically similar to ChatGPT’s fan-out queries, and that results with natural-language URL slugs had an 89.78% citation rate versus 81.11% for those without [6](#sprk-src-1). Name the entity and the question plainly: “How to connect GA4 and Search Console” beats a clever headline.

### 4. Recent, dated content

In an analysis of 17 million citations, Ahrefs found ChatGPT cited URLs that were on average 458 days newer than the pages in Google’s organic results, the strongest freshness preference of the platforms it tested [7](#sprk-src-4). Show a visible “last updated” date and actually update the facts behind it; changing the date alone does not add anything worth citing.

### 5. Content that is readable in the raw HTML

OpenAI’s documentation describes its crawlers but does not promise they execute JavaScript, so do not rely on it. If the answer only appears after client-side rendering, a fetcher may see an empty shell. View the page source (not the inspector) and confirm the answer text is in it.

### 6. No links inside the answer itself

More than nine in ten answer capsules in the Search Engine Land dataset contained no links at all [8](#sprk-src-3). The likely reason: a link implies the real answer is on another page. Keep the capsule clean and put internal and external links in the supporting paragraphs.

### 7. Corroboration from other sites

![A two-column comparison contrasting what you control on your own site versus why third-party entity presence across the rest of the web matters more for ChatGPT citations.](https://media.suparank.io/uploads/2026/07/sprk-get-cited-by-chatgpt-3.webp)

Your own page is only part of the picture. In a study of 75,000 brands, Ahrefs found branded web mentions correlated at 0.66 to 0.71 with brand visibility across ChatGPT, AI Mode and AI Overviews, and YouTube mentions correlated even more strongly (about 0.737). The number of pages on a brand’s own site barely correlated at all (about 0.194) [9](#sprk-src-5). Correlation is not causation, but the direction is consistent: being discussed elsewhere matters more than publishing more pages.

## Let OAI-SearchBot crawl your site

![Three icon cards showing the pre-flight checks for ChatGPT discoverability: Bing Webmaster Tools sitemap, robots.txt crawler access, and raw HTML answer rendering.](https://media.suparank.io/uploads/2026/07/sprk-get-cited-by-chatgpt-1.webp)

OpenAI runs separate user agents for separate jobs, and blocking the wrong one is the most common way sites remove themselves from ChatGPT search. Sites opted out of OAI-SearchBot are not shown in ChatGPT search answers, although they can still appear as navigational links [10](#sprk-src-6).



| User agent | Job | Effect of blocking it in robots.txt |
| --- | --- | --- |
| OAI-SearchBot | Surfaces websites in ChatGPT search results | Your pages stop appearing in ChatGPT search answers |
| GPTBot | Crawls content that may be used to train OpenAI’s foundation models | Signals your content should not be used for training; OpenAI documents it separately from search |
| ChatGPT-User | Visits a page when a user’s request in ChatGPT or a custom GPT needs it | OpenAI says robots.txt rules may not apply, and it is not used to decide search inclusion |

OpenAI’s user agents as documented by OpenAI, checked September 2026.A common setup for sites that want citations but not training use:


```
# Allow ChatGPT search to surface and cite this site
User-agent: OAI-SearchBot
Allow: /

# Opt out of training data collection (optional)
User-agent: GPTBot
Disallow: /
```

OpenAI notes it can take about 24 hours for a robots.txt change to affect search results, and it publishes the IP ranges for each agent so you can allow them in a firewall [11](#sprk-src-6). If a CDN or WAF blocks unknown bots by default, robots.txt alone will not let OAI-SearchBot in. Our [robots.txt for AI crawlers](/tools/robots-for-ai-generator/) generator writes these rules, and the [agent-readiness checker](/tools/agent-readiness/) tests whether AI agents can actually fetch and read your pages.

## Earn mentions off-site

Given the correlation data above, the off-site work is not optional. Focus on places that already show up in answers for your category:

- **Comparison and “best of” lists** written by publishers in your space. Ask ChatGPT your category question and note which list pages it cites; those are your outreach targets.
- **Review sites and marketplaces** where your category is listed. Keep your name, description and category consistent everywhere.
- **Communities** such as Reddit threads and industry forums. Answer real questions with specifics; promotional posts get removed and do not help.
- **YouTube.** Demos, talks and interviews that say your brand name in the title or transcript count as mentions in the study above.

Two things to avoid. Paying for placements on low-quality list sites rarely helps, because those pages seldom get retrieved for competitive prompts. And inconsistent naming (a product called three different things across your site, your docs and your listings) makes it harder for any engine to connect the mentions to one entity.

## Why ChatGPT cites a competitor instead of you

When you run the test below and a competitor keeps getting the citation, the cause is usually one of these. Work down the table in order, because the first rows block everything after them.



| Symptom | Likely cause | What to check or change |
| --- | --- | --- |
| Your site never appears, even for your brand name | OAI-SearchBot is blocked in robots.txt or by a firewall | Test robots.txt and firewall rules; allow OAI-SearchBot and OpenAI’s published IP ranges |
| Your page exists but a thinner page is cited | Title and URL do not match the question ChatGPT is asking | Rename the title and heading to the question in plain words |
| You are cited for some prompts but not the ones that matter | No direct answer for those questions on the page | Add a question-style H2 with a link-free answer capsule for each |
| A list article or Reddit thread is cited instead | Others discuss the topic; you only describe yourself | Get included in the lists and threads ChatGPT already cites |
| An older competitor page is cited over your newer one | Your page has no specific facts to attribute | Add original data, dated figures or a clearly labelled recommendation |

A diagnostic order based on the traits above. Fix access first; content changes cannot help a page that cannot be fetched.## How to check whether ChatGPT cites you

Because answers vary between runs, one screenshot proves little. Use a fixed prompt set and repeat it.

1. Write 10 to 20 prompts a buyer would actually ask, mixing category questions (“best tools for X”) with problem questions (“how do I fix Y”).
2. Run each prompt in ChatGPT with search on, in a logged-out or clean session, and record the result in a table like the one below.
3. Repeat the same set at least three times on different days before drawing conclusions.
4. Calculate a mention rate (prompts that name you, divided by prompts run) and a citation rate (prompts that link a page of yours).
5. In GA4, check referral traffic from chatgpt.com to see which cited pages send visits.



| Prompt | Run date | Brand mentioned? (Y/N) | Your URL cited? (Y/N) | Which page | Competitors cited |
| --- | --- | --- | --- | --- | --- |
| best \[category\] tools for \[audience\] |  |  |  |  |  |
| how to \[problem your product solves\] |  |  |  |  |  |
| \[your brand\] vs \[competitor\] |  |  |  |  |  |

A template to fill in. These rows are placeholders, not results.The manual method works for a handful of prompts. Once the set grows, a tool saves the repetition: Suparank runs your prompt set across ChatGPT, Google AI Overviews and AI Mode (Growth and Enterprise add Gemini and Claude), and paid plans repeat it on a weekly schedule, so you can [check whether ChatGPT cites you](/product/ai-visibility/) without logging runs by hand.

## ChatGPT vs Perplexity vs Google AI Overviews

The fundamentals overlap, but each engine retrieves differently, so check the details for each one you care about.



| Engine | Where sources come from | One documented difference | Guide |
| --- | --- | --- | --- |
| ChatGPT search | Web search results gathered by fan-out queries; OAI-SearchBot controls inclusion | Cites about half of the URLs it retrieves (Ahrefs, 2026) | This page |
| Perplexity | Its own retrieval, crawled by PerplexityBot | Reddit earned 21.6% of its citations across 3.1 million US queries (Ahrefs, 2026) | [Rank in Perplexity](/rank-in-perplexity/) |
| Google AI Overviews | Google’s search index | Cited pages averaged slightly older than organic results, unlike the chat assistants (Ahrefs) | [Google AI Overviews](/google-ai-overviews-optimization/) |

Differences drawn from the Ahrefs studies cited in this guide.Reddit’s share of Perplexity citations comes from Ahrefs’ tracking of 3.1 million US queries [12](#sprk-src-7).

## Sources

1. [Ahrefs ChatGPT citation study, 2026](https://ahrefs.com/blog/why-chatgpt-cites-pages/?utm_source=suparank.io) [1](#sprk-fnref-1) [2](#sprk-fnref-2) [6](#sprk-fnref-6)
2. [Seer Interactive, 2025](https://www.seerinteractive.com/insights/87-percent-of-searchgpt-citations-match-bings-top-results?utm_source=suparank.io) [3](#sprk-fnref-3)
3. [Search Engine Land, 2025](https://searchengineland.com/how-to-get-cited-by-chatgpt-the-content-traits-llms-quote-most-464868?utm_source=suparank.io) [4](#sprk-fnref-4) [5](#sprk-fnref-5) [8](#sprk-fnref-8)
4. [Ahrefs freshness study, updated 2026](https://ahrefs.com/blog/do-ai-assistants-prefer-to-cite-fresh-content/?utm_source=suparank.io) [7](#sprk-fnref-7)
5. [Ahrefs brand visibility study, 2026](https://ahrefs.com/blog/ai-brand-visibility-correlations/?utm_source=suparank.io) [9](#sprk-fnref-9)
6. [OpenAI crawler documentation, 2026](https://developers.openai.com/api/docs/bots?utm_source=suparank.io) [10](#sprk-fnref-10) [11](#sprk-fnref-11)
7. [Ahrefs Perplexity citation study, 2026](https://ahrefs.com/blog/most-cited-domains-perplexity/?utm_source=suparank.io) [12](#sprk-fnref-12)

## Topics

**Categories:** [Get Cited by AI](https://media.suparank.io/uploads/wp-mfa-exports/taxonomy/category/get-cited-by-ai.md)

**Tags:** [AEO](https://media.suparank.io/uploads/wp-mfa-exports/taxonomy/post_tag/aeo.md), [AI Citations](https://media.suparank.io/uploads/wp-mfa-exports/taxonomy/post_tag/ai-citations.md), [ChatGPT](https://media.suparank.io/uploads/wp-mfa-exports/taxonomy/post_tag/chatgpt.md), [Schema Markup](https://media.suparank.io/uploads/wp-mfa-exports/taxonomy/post_tag/schema-markup.md)