Thirty prompts · unlocks on this page
Find out what AI says about you when you are not in the room.
Five categories of prompt: baseline perception, category capture, head to head, objection surfacing, and source tracing. Run them in ChatGPT and Perplexity, screenshot what comes back, and you have a visibility baseline nobody sold you.
How to use them
- 1Replace the bracketed placeholders with your brand, category and top competitor.
- 2Run each one in a fresh chat with no memory or personalisation, otherwise you are measuring your own history rather than the model.
- 3Record whether you were named, where you ranked in the list, and which sources were cited.
- 4Repeat monthly. The delta is the only number that matters.
The source-tracing set at the end is the most useful and the least used. When a model tells you which threads it read, it has just handed you your content brief.
You ran the prompts. You did not like the answers.
That is the normal outcome, and it is fixable. The Reddit and community engine exists precisely for this.