Can I pay or submit to be included in ChatGPT's training data?

No, so Prefer works on the part you can influence instead: being retrieved and cited when ChatGPT searches. OpenAI does not offer a way to submit your site to its training set or to guarantee inclusion. GPTBot may crawl public pages that could be used in future training, but you cannot control whether, or how, your brand ends up in the model.

If training data is out of reach, how does ChatGPT name brands?

Mostly through live web search, the layer Prefer tracks. When you ask about a current product or company, ChatGPT searches the web and cites what it finds. That retrieval layer, not the frozen training set, is where you can actually influence whether your brand appears.

Should I block GPTBot then?

That is a separate decision about training, not search. Prefer's free crawler access checker reads your robots.txt and shows whether GPTBot and OAI-SearchBot are allowed or blocked. Blocking GPTBot keeps your content out of future training but does not affect ChatGPT search, which uses OAI-SearchBot. If you want to appear in ChatGPT's answers today, keep OAI-SearchBot allowed.

How do I get my brand into ChatGPT's training data?

You mostly cannot, and it is the wrong goal: ChatGPT names brands from live search. Prefer tracks whether it cites you and writes the content that earns it.

Javed Khatri Co-founder, Prefer

5 September 2026 Updated 30 September 2026 2 min read Answer

The short answer

Can I add my company to ChatGPT's knowledge?

You mostly cannot, and it is the wrong goal: OpenAI takes no submissions to its training data, and ChatGPT answers current questions from live web search. Prefer works on that search layer, tracking whether ChatGPT cites you and writing the content that earns the citation.

Key takeaways

OpenAI documents separate crawlers: GPTBot crawls content that may be used to train its foundation models, while OAI-SearchBot is used to surface websites in ChatGPT's search features. Training and search are different systems. (OpenAI, overview of OpenAI crawlers (bots))
In Prefer's 2026-09-02 measured run, ChatGPT answered 'What is Prefer' by retrieving and citing tryprefer.com and its pricing page in a live search, whereas an earlier gpt-4o run answered the same question from memory with fabrication. Retrieval, not training, surfaced the brand. (Prefer AEO loop run, 2026-09-02 (ChatGPT))

Training data is the wrong lever#

You cannot submit your brand to ChatGPT’s training data, and even if you could, it would not do what you want. Training data is a fixed snapshot gathered up to a cutoff date and baked into the model. OpenAI offers no submission process and no inclusion guarantee. Its GPTBot crawler collects public content that “may be used” in training, but “may” is doing real work there: crawling is not inclusion, and inclusion is not something you steer.

There is a deeper reason not to chase it. Even when a brand is in the training set, the model’s memory of it is fuzzy, out of date, and prone to invention. A frozen snapshot cannot know your current pricing, your latest launch, or a company founded after the cutoff.

Retrieval is what actually names your brand#

When you ask ChatGPT about a current company or product, it answers from a live web search, not from memory. OpenAI runs two different systems with two different crawlers: GPTBot for training, and OAI-SearchBot for ChatGPT’s search feature. The search system is the one that decides whether your brand shows up in answers about you today.

We watched this happen in our own tracking. On an early run, asking an older ChatGPT model “What is Prefer” produced a fabricated answer from memory. In our September 2026 run, the same question was answered by ChatGPT retrieving and citing our actual site and pricing page in a live search. Nothing changed in any training set between those runs. What changed was that the content became retrievable and citable at answer time. (Prefer is our own product, noted for disclosure.)

So the goal is not “get into the training data.” The goal is to be the source ChatGPT retrieves and cites when the question is about you. In practice that means:

Keep OAI-SearchBot allowed in robots.txt so ChatGPT search can reach you. Blocking it, or leaving your key pages un-crawlable, is the most common self-inflicted wound.
Publish clear, current, answer-shaped pages on the questions buyers ask, so there is something worth retrieving and quoting.
Build off-page presence on the sources engines already trust, so your brand is corroborated beyond your own domain.

That is answer engine optimization, and it works on the retrieval layer you can actually influence rather than the training snapshot you cannot. Our guide to appearing in ChatGPT search results walks through the retrieval work step by step. To see which questions ChatGPT already answers about you, and whether it cites you or guesses, run a free AI visibility audit.

Sources