It is used mostly when the problem is representation (AI describes you wrongly or confuses you with another company) rather than presence.
The risk of misusing it: talking as if a model could be “trained” or “edited” from outside. It cannot. What you work on is the consistency of your brand entity across every public source (name, description, address, services) and the accuracy of the data the model retrieves.