Assistant
ChatGPT
The largest audience, and the one most people mean when they say a customer asked AI about their business.
Last reviewed 09 August 2026. This subject moves monthly — products ship, crawlers change and documentation is rewritten without announcement. For anything newer than this page, OpenAI’s own site is the source of truth.
- Crawlers
- GPTBot (training) · OAI-SearchBot (search) · ChatGPT-User (live fetch)
- Retrieval
- Its own search index, with a history of Bing partnership — treat both as relevant
- You control
- Crawler access, structured data, whether your content renders without JavaScript
- You do not control
- Whether it names you, or how it summarises you
- Pillar weight
- Find you and Describe you
ChatGPT is where most people first notice that AI is answering questions about their business. It is also the platform where the least is publicly documented about how sources get chosen.
The three crawlers, and why the difference matters
OpenAI operates several agents and they do different jobs. Blocking one is not blocking the others.
GPTBot gathers content that may be used for training. Blocking it does not remove you from models already trained, and it does not affect whether ChatGPT can fetch your page when answering a live question.
OAI-SearchBot builds the search index behind ChatGPT's browsing. This is the one that matters for being found today.
ChatGPT-User fetches a page in the moment, because a user asked about it or pasted a link.
Most businesses want all three allowed. If you block, block deliberately and know which behaviour you are affecting — our crawler guide has the exact robots.txt lines.
What actually gets you named
Nothing exotic. The same structural work that serves every other assistant: an unambiguous identity, content that exists in the HTML before JavaScript runs, and answers written in the shape of the questions people ask.
The thing that surprises people is how often the failure is mechanical rather than editorial. A restaurant with a PDF menu is not being outranked; it is being skipped, because there is nothing to read.
What you cannot control
Whether it mentions you. Whether the summary is flattering. Whether it prefers a directory listing about you over your own website — which it often will, because directories are structured and consistent, and many business sites are neither.
That last point is the argument for entity clarity. If your own site is the most machine-readable source about you, it becomes the one worth quoting.
On accuracy. Retrieval architectures change without announcement, and none of these companies publish full details. This page states what is documented, flags what is inferred, and avoids claiming internals nobody outside the company can verify. Spotted something out of date? Tell us.
Take this to your assistant
Paste it into ChatGPT, Copilot, Claude or Gemini and apply it to your own website.
Nothing is sent anywhere. The text is copied to your clipboard.