Most owners who use ChatGPT use it to write things. Emails, posts, the About page. That's fine. But it's just as useful doing the opposite: picking apart what you already wrote, the way a skeptical customer would, before a real one ever sees it.
The prompt
Paste in your price sheet, your homepage, a quote you're about to send, or a new offer. Then add this:
"Pretend you're a skeptical customer comparing me to two other businesses. Ask me the eight hardest questions you'd want answered before you paid me. Be blunt. Don't be nice about it."
Then actually answer the questions. Out loud or typed, it doesn't matter.
What it catches
Missing information
"What's included?" "How long does it take?" "What if I'm not happy?" If you can't answer in a sentence, your page can't either.
Price pushback, before it happens
It'll ask why you cost more than the other guy. Better to write that answer now than make it up on the spot at the counter.
Words you stopped noticing
If it asks what a word means, that's a word your customers don't know either.
Where people go wrong
Don't let it rewrite the whole thing afterward. The point is the questions. Your answers, in your own words, are what customers are buying.
And don't paste in anything private. Customer names, account numbers, anything you'd hate to see outside the building. A plain rule like that is most of an AI policy, by the way.
Your next move
Pick the one thing a customer reads right before paying you. Run it through the prompt tonight. Fix the top two questions it asks and leave the rest for next week.