A good product prompt isn't a magic trick. It's a checklist you don't have to remember, run by something that never gets bored.
Name the failure modes
Treat the first output as a draft, not a verdict. The prompt does the typing; you keep the judgment about what ships.
- Compare two outputs instead of trusting one
- Lower the reading level, then check what broke
- Keep one ask per prompt — chaining beats cramming
Make the output reviewable
Write down what you never want as explicitly as what you do. The exclusions often do more work than the instructions.
Ask for the reasoning, then the result. When the model shows its steps you can see exactly where to push back.