To understand how your brand appears in ChatGPT, start by measuring the answers buyers receive. When ChatGPT uses web search, inspect its cited sources. Some answers do not use search, and appearing in one answer does not mean your brand will be recommended in the next.
This guide explains crawler controls, how to inspect cited sources, and why a repeatable measurement should inform the work you choose. It does not offer a formula for guaranteed recommendations.
How ChatGPT decides what to recommend
ChatGPT can search the web when a question benefits from current information. Responses using search may include citations, as OpenAI explains in its search guide. For a brand study, inspect the sources actually cited in the saved responses rather than assuming that every answer used search or that your own website was its main source.
OpenAI documents separate crawler controls: OAI-SearchBot is used to surface websites in ChatGPT search, while GPTBot crawls content that may be used for training. The settings are independent. See OpenAI’s crawler documentation for the distinction and the relevant robots.txt controls. For the broader robots.txt and noindex discussion, see our full guide: /seo-para-chatgpt/.
Keep your business information current in Business Hub
Business Hub gives eligible businesses another way to supply information to ChatGPT. Use it alongside your website and the other sources buyers rely on. Our Business Hub guide for Mexican businesses explains what you can set up today, how to prepare your information, and what to organize for future Business Agents.
What does not work
Google’s guide describes its own generative AI features in Google Search. Its advice does not establish what ChatGPT or other assistants use or require. Read Google’s guide to generative AI features in Search:
- llms.txt. Google says it does not use this file for Google Search and that it neither helps nor harms visibility there. Its guide notes that other services may use such files. We do not sell it as a Google Search visibility lever.
- Special schema for Google’s AI features. Google says no special markup is required for its generative AI Search features. That does not establish what other assistants require.
- Breaking content into tiny pieces. Google says there is no requirement to do this for its generative AI Search features. Write clearly for people and structure content to suit the subject.
- Guarantees. Nobody can guarantee an appearance in ChatGPT; the answers vary by user, by phrasing and by run. What a serious provider does commit to is the instrument and the run: a prompt bank complete and useful for your industry, and a study that produces findings you can act on. What nobody commits to, us included, is that the actions you take afterwards will move the numbers in the next measurement. A guaranteed position is the one promise to walk away from.
For Google Search, Google says foundational SEO best practices remain relevant to its generative AI features. That is Google’s guidance for its own products, not proof that every assistant works the same way. Clear, useful content and sound technical basics remain worthwhile.
What actually moves it
If the answer is assembled from the open web, the open web is where the work is:
- Be present in the sources the models cite. For consumer products that usually means retailer product pages, review platforms, comparison guides and earned media. Which ones, in your category, is a measurement question, because they differ.
- Keep content recoverable. Do not block OAI-SearchBot by accident, and check what your robots.txt and noindex settings actually do before you assume.
- Entity consistency. The same name, the same category description, the same facts about who you are, everywhere. Models reconcile sources; contradictions make you harder to recommend.
- Earned mentions. Third parties saying your name in the right context beats you saying it about yourself. PR and reviews are not decoration here; they are inputs.
Notice that almost none of these moves live on your own domain. Working through the sources the models cite — retailer pages, commissioned editorial, earned media — is its own job: /trabajo-sobre-fuentes/.
Why measure before you optimize
Here is the trap: you ask ChatGPT once, you are absent, and you conclude you are invisible. Then you ask again tomorrow and you are there. A single ask proves nothing in either direction; the same question returns different lists across runs.
So before spending on optimization, spend on knowing two things: which sources actually carry the answers in your category, and where you stand today against the brands that get named instead of you. That is a measurement problem, and we built a tool for it.
The full guide, including how a study works: /seo-para-chatgpt/. When you want the measurement instead of another opinion (the sources that carry the answers in your category, and where you stand against the brands named instead of you), the study is here: /auditoria-visibilidad-ia/.