You don't need a hundred data work prompts. You need five you trust and actually reuse.
Reuse the recipe
Ask for the reasoning, then the result. When the model shows its steps you can see exactly where to push back.
Keep a negative list
Write down what you never want as explicitly as what you do. The exclusions often do more work than the instructions.
- Compare two outputs instead of trusting one
- Have the model flag what it could not cover
- Lower the reading level, then check what broke
Start with the smallest version
Treat the first output as a draft, not a verdict. The prompt does the typing; you keep the judgment about what ships.
- Name the audience and the single goal before you generate
- Save the prompt that worked so the next person reuses it
- Compare two outputs instead of trusting one
Ask for the reasoning, then the result. When the model shows its steps you can see exactly where to push back.