How do I get listed on ChatGPT?

People ask this question in a very specific tone. They have already tried ChatGPT, asked it to recommend a business like theirs, watched it name three competitors, and now they want to know where the form is. Understandable — every previous generation of discovery had one. Google had Add URL. Yelp had a claim page. The App Store had a review queue.

ChatGPT has nothing like that, and it is not going to. What it has instead is a set of crawlers, a retrieval layer, and a model that decides — per answer, per question — which pages are worth quoting. Once you understand those three pieces, the question stops being "how do I submit" and becomes "what makes a page quotable." That is a question you can actually work on.

First, the part everyone gets backwards: GPTBot vs OAI-SearchBot

OpenAI runs several crawlers and they do completely different jobs. Confusing them is the single most common self-inflicted visibility wound we see on audits — a site blocks everything with OpenAI's name on it, then wonders why it stopped appearing in ChatGPT answers.

The distinction matters because a lot of sites made a reasonable-sounding decision in 2023 and 2024 — "we don't want AI training on our content" — and implemented it with a blanket block. That decision was about training. The cost landed on search.

The robots.txt that opts into visibility without opting into training

This is the configuration most businesses actually want: be findable and citable in AI answers, but keep your content out of the training corpora. Google and Apple both ship a dedicated second token for exactly this, and OpenAI's split does the same job.

# Answer engines: allowed to crawl so we can be cited
User-agent: OAI-SearchBot
Allow: /

User-agent: Googlebot
Allow: /

User-agent: bingbot
Allow: /

User-agent: PerplexityBot
Allow: /

User-agent: Claude-SearchBot
Allow: /

# Model training: excluded
User-agent: GPTBot
Disallow: /

User-agent: Google-Extended
Disallow: /

User-agent: Applebot-Extended
Disallow: /

User-agent: CCBot
Disallow: /

Sitemap: https://example.com/sitemap.xml

Two honest caveats. First, blocking training crawlers is a values decision, not a visibility decision — it will not help you get cited, and for some publishers the training exposure is worth having. Second, robots.txt is not a wall. OpenAI's own publisher documentation notes that even a blocked page can still surface as a bare link and title if ChatGPT obtained the URL from a third-party search provider. To be excluded entirely you need a noindex meta tag, which means the crawler has to be allowed to fetch the page in order to read the instruction telling it to go away.

Second, get indexed by Bing — it is the closest thing to a submit button

ChatGPT does not maintain an independent web index at Google's scale. It leans heavily on an existing search pipeline, and multiple independent analyses have put Bing at the center of it — one widely-cited study by Seer Interactive found roughly 87% of ChatGPT citations matched pages already ranking in Bing's top results.

That has a very practical consequence. Most businesses have spent a decade optimizing for Google and have never once opened Bing Webmaster Tools. If your pages are thin or missing in Bing's index, you are invisible to the retrieval layer that ChatGPT reaches for — no amount of on-page polish fixes that.

1. Create a Bing Webmaster Tools account and import your Google Search Console properties — it takes about four minutes and pre-populates everything.
2. Submit your sitemap and confirm your key pages are actually indexed, not just submitted.
3. Turn on IndexNow so new and updated URLs get pushed to Bing immediately instead of waiting for a recrawl.
4. Check the crawl-errors report. Bing is fussier than Google about redirect chains, soft 404s, and heavy client-side rendering.

Third, write pages a model can quote without editing

This is where the real work is, and where most advice gets vague. Fortunately there is measured evidence about what cited pages have in common. Ahrefs analyzed 1.4 million prompts and compared pages ChatGPT cited against pages it did not.

Strip out the statistics and the finding is almost embarrassingly simple: the page that gets cited is the page whose title is closest in meaning to the question that was asked. Not the page with the most backlinks. Not the page with the most words. The page that is visibly, semantically about that exact question.

The same study found cited pages skewed older than you might expect — a median age around 500 days for search-derived citations. Freshness is not the lever people assume it is outside of news. A well-aged page that squarely answers a question beats a new page that circles it.

What that looks like on the page

• One question per page, stated in the title. "How do I get listed on ChatGPT?" not "AI Visibility Solutions." If a reader would not type your H1 into a chat box, it is not an answer page.
• A self-contained answer in the first 60 words. Assume the model lifts that paragraph and nothing else. If it only makes sense after the two paragraphs above it, it will not survive extraction.
• Readable URL slugs. /blog/how-do-i-get-listed-on-chatgpt carries meaning that /blog/post-4471 does not.
• Specifics that can be quoted. Named crawlers, exact directives, real numbers, dated studies. Generic advice is unquotable — there is nothing in it to cite.
• Structure that survives being chopped up. Headings that are questions, short paragraphs, tables for comparisons, lists for sequences.

Why ChatGPT recommends your competitor

Usually one of four reasons, in descending order of how often we find them on audits.

1. You are blocked. OAI-SearchBot cannot reach you. Fixable in one line.
2. You are not in the index it reaches for. Strong in Google, absent or thin in Bing.
3. You have no page for the question. Your competitor wrote "best CRM for a two-person marketing team" and you wrote "Solutions." The model needed a page about that question and yours was not one.
4. They are talked about elsewhere and you are not. Analyses of AI citations consistently find the majority come from third-party pages — directories, roundups, forum threads, press — rather than a brand's own site. Your homepage is not the main lever here; your presence in other people's content is.

Do not chase the citation mix

You will find posts telling you to get on Reddit because Reddit dominates AI citations. Be careful with that advice. Semrush tracked more than 100 million citations across 230,000 prompts over 13 weeks in 2025 and watched ChatGPT's Reddit citation share fall from roughly 60% to about 10%, with Wikipedia dropping from around 55% to under 20% in the same window. Perplexity and Google's AI Mode stayed comparatively stable over the same period; ChatGPT did not.

The lesson is not "Reddit is worthless." It is that the source-type mix is a policy decision made by a vendor, it can change in a week, and building your strategy on top of it is building on someone else's product roadmap. The stable strategy is the boring one: be crawlable, be indexed, and be the clearest answer to a real question.

A checklist you can run this afternoon

1. Read your robots.txt. Confirm OAI-SearchBot is allowed. Decide GPTBot deliberately rather than by accident.
2. Verify your site in Bing Webmaster Tools, submit the sitemap, and enable IndexNow.
3. List the ten questions your customers ask before they buy. Check whether you have a page whose title is one of those questions.
4. For every page you do have, move the direct answer into the first paragraph.
5. Check your analytics for referrals carrying utm_source=chatgpt.com — that is how ChatGPT tags its outbound clicks, and it is the only first-party proof you are being cited.
6. Re-run in 30 days. Citation is not a one-time state; it is a position you hold or lose.