How to Rank in ChatGPT: Get Your Brand Mentioned and Cited

To rank in ChatGPT, let OpenAI's OAI-SearchBot crawl your site and get your brand named on the pages ChatGPT already cites. Then publish pages that answer buyer prompts in their opening lines. Ranking in ChatGPT means 2 things: your brand is named in the answer (a mention), or your page is linked as a source (a citation). There is no fixed first place. Repeat a prompt and the list changes.

This ChatGPT SEO guide combines 3 kinds of evidence. The first is OpenAI's crawler and publisher documentation. The second is SparkToro's consistency study and 4 Ahrefs studies of AI visibility and citations. The third is our own audit of the robots.txt files of the 50 websites ChatGPT cites most. We checked every third party figure on its source page on October 30 or 31, 2026, and we label our own calculations.

What does it mean to rank in ChatGPT?

Ranking in ChatGPT means appearing in a high share of answers to your buyers' prompts, as a named brand or a cited source. A Google result holds a position for a keyword. A ChatGPT answer is generated fresh on every run.

SparkToro measured how much answers move in a study published on January 27, 2026. 600 volunteers ran 12 prompts through ChatGPT, Claude and Google's AI answers a combined 2,961 times. There was less than a 1 in 100 chance that ChatGPT returned the same list of brands in any 2 responses. The same list in the same order came closer to 1 in 1,000 runs. SparkToro concluded that visibility percentage is a reasonable metric across dozens to hundreds of repeated prompts (SparkToro AI consistency study).

How does ChatGPT search choose which sources to show?

ChatGPT answers from 2 places: what the model learned in training, and live web results it retrieves when a question benefits from a search. OpenAI states that ChatGPT searches automatically in those cases. It may share disassociated queries with third party search providers such as Bing. Answers that use search carry inline citations, and a Sources button lists cited sources plus other relevant links (OpenAI ChatGPT search help article).

OpenAI's publisher FAQ says any public website can appear in ChatGPT search. Its main technical instruction is to avoid blocking OAI-SearchBot (OpenAI publishers and developers FAQ). None of the 3 OpenAI pages cited in this guide lists ranking factors or weights.

In our view the live search path is the faster one to influence. You control crawl access and your pages today. In our understanding, training memory changes only when OpenAI ships a new model.

Can OpenAI's crawlers reach your site?

Check robots.txt for OAI-SearchBot first. OpenAI states that sites opted out of it are not shown in ChatGPT search answers, apart from plain navigational links. OpenAI lists 4 user agents. The OAI-SearchBot and GPTBot settings are independent of each other (OpenAI crawler overview).

A robots.txt that keeps you eligible for ChatGPT search while opting out of training looks like this:

# ChatGPT search: allowed
User-agent: OAI-SearchBot
Allow: /

# Model training: optional opt out
User-agent: GPTBot
Disallow: /

Then run 5 checks:

1. Check for a catch all group. A robots.txt with no OAI-SearchBot group applies its User-agent: * rules to OpenAI's crawler, so a catch all Disallow: / blocks it.
2. Allow OpenAI's published IP ranges (openai.com/searchbot.json) in your firewall or CDN.
3. Review Cloudflare under Security Settings, then Configure AI bot policies. Cloudflare's documentation sorts bots into Search, Agent and Training presets. It states that from September 15, 2026, new domains block Training and Agent bots on pages with ads by default, while Search stays allowed.
4. Wait about 24 hours, the time OpenAI gives after a robots.txt update.
5. Use a noindex meta tag, not a robots.txt block, on pages you want kept out. OpenAI notes its crawler reads the tag only if it can crawl the page.

Bing indexing is a sensible extra check, because OpenAI names Bing as an example search provider. Add your site to Bing Webmaster Tools, submit your sitemap and inspect your 10 key URLs. Afterwards, search your server logs for the string OAI-SearchBot. Requests to your service pages confirm the crawler gets through.

What do the most cited sites do with OpenAI's crawlers?

We read the robots.txt files of the 50 most cited websites in ChatGPT on October 31, 2026, using the Ahrefs list in the next section. 41 files were readable, and those 41 sites hold 69.6% of citations to the top 50. The other 9 returned a security check, an access denied page, an error page or a browser block. They include Reddit and BBB. We tested each file against a sample article path and applied the most specific matching group, as robots.txt rules require. PMC keeps articles under /articles/, which its file allows, so we tested that path there.

The 8 GPTBot blockers are Forbes, Healthline, Edmunds, Tripadvisor, ScienceDirect, Dictionary.com, LinkedIn and Facebook, and together they hold 14.0% of citations to the top 50. 7 of the 8 still allow OAI-SearchBot, which matches OpenAI's statement that the 2 settings are independent. Opting out of training did not keep them out of answers. The Ahrefs data is labelled September 2026, and we read the files on October 31. Facebook is the 1 site that blocks OAI-SearchBot, through a catch all rule, yet it still holds 1.4% of citations. OpenAI's crawler page explains part of this: opted out sites can still appear as navigational links. Its publisher FAQ ties summaries and snippets to OAI-SearchBot access, so a blocked page loses the quoted text.

Which websites does ChatGPT cite most often?

Reddit is the most cited website in ChatGPT, with 16.8% of citations among the top 50 sources, ahead of Wikipedia at 7.0%. The data comes from Ahrefs Brand Radar for broad US queries across all topics, labelled September 2026 (Ahrefs most cited websites in ChatGPT). The 50 shares on that page add up to 100%, so each figure is a share of citations to those 50 domains.

We grouped all 50 domains by page type and added their shares. The grouping is ours, and a borderline site could fit 2 groups.

By this count, 49.7% of citations to these 50 domains go to review sites, communities and customer review profiles. In our view, communities and customer review profiles are open to any brand through helpful participation and real customers. Review sites split: comparison publishers accept pitches, while testing labs such as Consumer Reports and RTINGS choose products independently. News publishers (14.0%) open up through digital PR, and retailers (8.7%) through complete product listings. In our view, dictionaries, health and government sources sit outside a brand's reach.

Wikipedia, at 7.0%, rewards notability rather than outreach. An article about a company rests on independent coverage in reliable publications, so press coverage comes first. Keep your company's Wikidata entry accurate too.

Your niche has its own mix. Build a niche list from the Sources panel of your own prompts, or from the cited sources report in the AI visibility tools we compared.

Do best of lists decide who ChatGPT recommends?

"Best X" blog lists were the most prominent page type in ChatGPT sources, at 43.8% of cited page types across 750 top of funnel prompts. Glen Allsopp at Ahrefs ran prompts in 3 categories: software, products and agencies (Ahrefs best lists study). The study reported 4 more findings:

• Position: brands placed higher in the associated lists were recommended more often.
• Freshness: 79.1% of 1,100 cited lists were last updated in 2025, and 26% in the 2 months before the check.
• Authority: 35% of cited lists sat on low authority domains, so DR alone does not decide which list counts.
• Own lists: a brand's self published list appeared in more than a third of software responses that recommended that brand.

To find your target lists, run your top 10 buyer prompts and open Sources on each answer. Record every best of page, your position on it, and its last updated date. Lists that name 3 of your competitors but not you come first. Treat Reddit threads in Sources the same way: note each thread, its subreddit and its age. In our experience, a pitch with a fresh fact gives an editor a reason to update. Good examples are a new feature, a price change or a new market.

Our brand mentions service places brands in relevant listicles and reviews for $286 per placement. Sites are DR 30 to 60 with 1,000+ monthly traffic. Clients can request a top position in a list, and we prioritize publishers that offer one. No provider can guarantee that ChatGPT cites a placement, and we do not claim to.

Which pages on your own site can ChatGPT cite?

General landing pages are cited more often than homepages when ChatGPT recommends a software brand or an agency. In the same Ahrefs study, first party landing pages appeared in 37.2% of software prompts and 30.4% of agency prompts. Homepages appeared in 15.6% and 14%. In product prompts, the brand's product page appeared in 87.2%.

Give every service, use case and product its own page, built from facts a model can lift into an answer. This is our checklist for those pages:

The examples are illustrative. Schema markup is optional here: none of the 3 OpenAI pages cited in this guide mentions it. We still use Organization schema with sameAs links to our social profiles, so machines read one consistent set of brand facts. Keep your brand name, founding year, headquarters and core services identical on your site, LinkedIn, review profiles and directories. Our guide to on page SEO covers the titles, headings and internal links behind the same pages in Google.

Does ChatGPT prefer to cite fresh content?

Yes. ChatGPT showed the strongest preference for recent pages of any AI assistant Ahrefs measured. Ahrefs published the study on July 28, 2025. It analyzed 16.975 million cited URLs from ChatGPT, Perplexity, Gemini, Copilot, AI Overviews and Google results (Ahrefs study of 17 million AI citations).

ChatGPT citations averaged 458 days newer than Google's organic results. Ahrefs measured age by the date its crawler first saw each page, so the data shows a preference, not proof that updating causes citations.

We treat freshness as a reason to keep facts current, not to change dates. Refresh the pages that carry numbers on a fixed cycle: prices, plan limits, statistics and screenshots. Change the visible updated date only when a fact on the page changed.

Do YouTube, Reddit and backlinks help you show up in ChatGPT?

Brand mentions, led by YouTube mentions, correlate with AI visibility far more than backlinks or page counts do. Ahrefs studied 75,000 brands across ChatGPT, AI Mode and AI Overviews with the Spearman correlation coefficient. It published the results on December 12, 2025 (Ahrefs 75,000 brand correlation study). Ahrefs states that correlation is not causation.

Ahrefs reported very weak correlations for backlink counts and URL Rating. It also found ChatGPT has the weakest correlations with traditional brand authority signals of the 3 assistants. In Ahrefs' words, ChatGPT seems to be the most accessible entry point.

YouTube

Ahrefs counts a YouTube mention whenever a brand name appears in a video title, transcript or description. Ahrefs found the volume of mentions seems to matter slightly more than their reach. Get named in 3 kinds of video: reviews by niche creators, comparison videos, and tutorials that use your product. Ahrefs' count includes transcripts, so in our view it pays to say your brand name aloud in your own videos too.

Reddit

Start with the threads ChatGPT already cites for your prompts, found in the Sources panel. Answer the thread's question fully first, and name your brand only where it helps the reader. Our community mentions campaigns follow that rule: 10 thread mentions on Reddit, Quora and niche forums for $510.

How do you track your ChatGPT visibility?

Track ChatGPT visibility by running 30 fixed prompts 5 times each every month and recording the share of answers that name or cite you. Our method has 6 steps:

1. Write 30 prompts: 10 "best X for Y" prompts, 10 "X vs Y" or alternatives prompts, and 10 problem questions your buyers ask.
2. Run each prompt 5 times in a new chat with memory off, on the same model and from the same location, for 150 answers.
3. Log 4 fields per answer: brand named, page cited, cited URL, and every best of list in Sources.
4. Calculate visibility rate as answers that name you divided by total answers. Calculate citation rate as answers that cite your URL divided by total answers.
5. Repeat on the same day each month, and track 3 named competitors with the same prompts.
6. Filter Google Analytics 4 sessions by the source chatgpt.com. OpenAI adds utm_source=chatgpt.com to referral links for sites that allow OAI-SearchBot.

The worksheet uses sample numbers to show the formulas. It is not client data.

This sample brand wins comparison prompts and misses problem questions, so its next pages answer those problems. Visits without the utm tag, such as copied links, are not attributed to ChatGPT in GA4. Dedicated trackers automate the runs; we compare 12 of them in our AI visibility tools guide.

Which ChatGPT visibility problem do you fix first?

Fix problems in this order: crawl access, missing mentions, missing pages, then wrong facts. Each symptom in your audit points to a different cause:

We correct a wrong fact in 4 places, in this order. First comes your own page, then your Google Business Profile and LinkedIn page. Next come review profiles, then directories and Wikidata. Then pitch a correction to any cited article that repeats the old fact. Rerun the affected prompts a month later.

We advise against 3 tactics. Hidden text written to instruct AI models is a manipulation attempt that undermines trust. A new date on a page with no new facts misleads readers. Sites with more pages showed almost no relationship with AI visibility in the Ahrefs data. For the same work across Perplexity, Gemini and Google, read our answer engine optimization guide.

Frequently asked questions