ChatGPT is one of the assistants your question list runs against each day. The agents record whether you are named, how you are described and which sources are shown with the answer. Responses can differ from run to run, so we read patterns across many runs rather than treating a single reply as the truth.
Because customers phrase things differently, we vary the wording of important questions and keep each version in the record. That shows whether a change in your visibility is real or just a quirk of one phrasing. Every version is stored, so you can read the exact answers behind any figure in a report.
ChatGPT results feed the same reports as every other assistant, so you can compare them side by side. A gap that shows up only here may point to a specific source it favours. A gap everywhere points to something broader about how your business is described online.
Some questions are too personal or too current for an assistant to answer reliably, so tracking them tells you little. We steer the list toward the recommendation and comparison questions where buyers ask and assistants do name businesses. The list is reviewed with you, not decided for you. If ChatGPT is not an assistant your customers use, we say so and weight the report elsewhere. Visibility is about where your audience asks, and your team knows that better than any tool.