A good AI prompt is not a magic formula. It's a short briefing: the goal, what matters, what you already know, and what a good answer must demonstrate.

  1. Define the decision.

    Replace “Give me business ideas” with “Compare three small, testable products I could create with the skills and time I have.”

  2. Share constraints.

    Include your budget, available time, technical comfort, non-negotiables and the main risk you want to avoid. Those facts should change the recommendation.

  3. Separate evidence from assumptions.

    Ask for four labels: Verified fact, reasonable inference, recommendation, and unknown. If a fact would change your decision, verify it with a primary source.

  4. Ask for options and failure conditions.

    Don't request only a favorite. Ask what could make each approach fail and what would convince you to stop.

  5. Reduce it to one next test.

    Prefer a low-cost, reversible first step that produces useful evidence before you make a major commitment.

  6. Review what actually happened.

    Return with the results. Adjust the plan based on observed outcomes rather than repeating the original assumption.

AI limitation: A confident answer is not the same as an accurate one. Verify factual claims, current rules, prices, technical details and high-stakes advice before relying on them.

A short prompt you can use

Help me evaluate this decision. My goal is [goal]. My constraints are [constraints]. Separate verified facts, assumptions, recommendations and unknowns. Compare realistic options and identify the smallest next test. Do not invent prices, statistics, or evidence.

Or use our free guided prompt builder to assemble one from your own situation.