Turn what you need into an AI work assignment.
Answer a few plain-English questions. We will turn them into a clear assignment you can use with Claude, ChatGPT, Gemini, or an AI agent. You can skip any question you do not need.
What would a good result look like?
Start with the finish line, not the command. Describe the result you want.
What should the AI know first?
Share the facts, audience, resources, prior attempts, or background a capable helper would need.
How will you know it worked?
Tell the AI what matters most. This helps it make good tradeoffs instead of just producing a long answer.
How much freedom should the AI have?
Choose how the AI should act. You remain in control.
What must it protect or avoid?
Add boundaries, risks, tone, budget, privacy needs, or anything that would make an answer unacceptable.
How should it check its work?
A good assistant does not simply answer. It checks the answer against your standards.
When should it stop and ask you?
This is especially useful for money, health, safety, customers, or decisions with real consequences.
What should the finished answer include?
Specify the format and when the job is done.
Ready to copy and use
For high-stakes health, legal, or financial decisions, use AI for questions and organization—not as your only professional advice.
Outcome
Name the result that matters.
Decision rights
Say what AI may decide and what remains yours.
Verification
Require the AI to check its work.
Human judgment
Keep meaningful decisions with people.