Brand mentions in generated answers respond to factors the team controls. Name instead of refer, provide the data nobody has, and exist outside your own domain.

The page that ranks and the page that gets cited are not always the same. The five-step cycle I use to find out which passage backed each mention.

If you do not measure presence in generated answers, you are reporting on a universe that shrank. What stays invisible without this measurement, and the argument that gets budget approved.