Reasoning Prompts
Devil's Advocate
Generate the strongest possible counterarguments to your plan or idea to stress-test it before committing.
Prompt
Play devil's advocate against the following plan or idea. Your goal is to construct the strongest, most credible case against it — not to be contrarian, but to surface real risks and weaknesses before I commit. **Plan / Idea**: [PLAN] Argue against it from these angles: 1. **The fundamental assumption that's wrong**: What core premise does this plan rely on that could be false? 2. **The overlooked alternative**: What better option exists that this plan ignores or dismisses too quickly? 3. **The known failure mode**: What well-documented pattern (from history, similar industries, or prior attempts) does this plan repeat? 4. **The unintended consequences**: What second- and third-order effects could make success worse than failure? 5. **The fatal edge case**: What specific scenario — even if unlikely — would cause this plan to fail catastrophically? After making the case, rate the overall strength of your counterargument: **Weak / Moderate / Strong**. A strong rating means you've identified a likely, significant problem. A weak rating means the concerns are real but manageable.
How to Use
Paste your plan, proposal, or decision into [PLAN]. This works best when you've already thought through the idea and feel reasonably confident about it — the AI's job is to find what you've missed. Use it before pitching to stakeholders, committing budget, or writing a spec. The five fixed angles ensure coverage across assumptions, alternatives, precedents, side effects, and edge cases rather than only attacking the most obvious weakness.
Variables
| Variable | Description |
|---|---|
| [PLAN] | The plan, idea, or proposal to argue against — include key details like goals, approach, resources, and timeline for more targeted criticism |
Tips
- After receiving the counterarguments, respond with your rebuttals and ask: "Given these rebuttals, where does the plan remain weakest?" for a second-pass critique.
- If the AI's counterargument rating is "Weak", that's a signal your plan is robust — but still address the fatal edge case before proceeding.
- You can run this prompt on a competitor's strategy to understand their vulnerabilities before they realize them.
- Add "Focus especially on [specific concern]" if you have a nagging doubt you want the AI to develop into a full argument.