In short
- To appear in ChatGPT Search, your site must be accessible to the OAI-SearchBot crawler: a site that blocks it is not cited in answers.
- OAI-SearchBot (search) and GPTBot (model training) are independent: you can allow one and block the other.
- Cited content is content that answers precisely, with verifiable facts: the same qualities that serve your SEO.
- This is a growing visibility channel that complements Google without replacing it: your category pages remain the main asset.
With the rise of ChatGPT Search, a new question has joined the list for e-commerce teams: how do you get your brand and products to appear in ChatGPT answers? The good news is that part of the answer is technical and simple. The rest comes down to what you are already doing for Google.
The prerequisite: let OAI-SearchBot through
ChatGPT Search relies on a crawler, OAI-SearchBot, to find and cite web pages. OpenAI is explicit: a site that blocks OAI-SearchBot does not appear in ChatGPT search answers (it can only appear as a plain navigation link).
So check your robots.txt file. Some sites blocked all AI crawlers outright in 2023 and 2024, without distinguishing between their uses.
Search and training: two crawlers, two decisions
OpenAI uses several crawlers, which can be configured independently:
| Crawler | Purpose | If you block it |
|---|---|---|
| OAI-SearchBot | Finding and citing pages in ChatGPT Search | Your site is not cited in search answers |
| GPTBot | Collecting content to train models | Your content is not used to train future models |
You can therefore allow OAI-SearchBot to stay visible, while blocking GPTBot if you do not want your content to be used for training. That choice is yours; what matters is that it is deliberate.
What gets a page cited
Once access is open, the same qualities as for Google's AI Overview apply, as we detail in our GEO guide:
- a precise answer to the question, stated clearly from the start;
- verifiable facts: features, dimensions, compatibility, sourced figures;
- up-to-date information: prices, availability, product ranges;
- an accessible, fast page, with its content present in the HTML.
OpenAI also publishes recommendations dedicated to product discovery in ChatGPT: complete and accurate product data, pages accessible to the crawler.
What is pointless
- Miracle files that are supposed to guarantee a citation: no file replaces accessible, useful pages.
- Keyword stuffing: the reference study on GEO (Princeton, 2024) shows that it does not improve visibility in generated answers.
- Generic content: an interchangeable text has no reason to be cited rather than another.
How much room should this channel get?
ChatGPT is a real and growing visibility channel, particularly for advice and comparison searches. It complements Google without replacing it: for an e-commerce site, your category and facet pages remain the main asset on purchase queries. In our study of 599 French queries, Google shows a product module on 94% of transactional queries.
The good news is that the two converge: factual, precise and well-structured category pages serve both your SEO and your visibility in generative engines. That is the standard unblind holds every text to: claims about your offer confirmed by the page, no information that will be outdated next season, no generic formula.
Frequently asked questions
How do you appear in ChatGPT?
First check that your robots.txt does not block OAI-SearchBot, the crawler OpenAI uses for ChatGPT Search. Then the fundamentals apply: accessible, fast pages that answer questions precisely with factual, up-to-date information.
Does blocking GPTBot prevent you from appearing in ChatGPT?
No. According to OpenAI, GPTBot is used to train models and OAI-SearchBot for search. The two settings are independent: you can block GPTBot and allow OAI-SearchBot to keep appearing in search answers.