Compare options for a decision
Stress-test a decision instead of outsourcing it: your options against your constraints, trade-offs made explicit, and a recommendation that argues for itself.
Last updated
Fill in the blanks
Assembled in your browser — nothing you type is stored or sent anywhere.
Your prompt
Model-agnostic — works in Claude, ChatGPT, Gemini or any AI chat.
Why this prompt works
- →Ordered criteria turn "pros and cons" into an actual evaluation — without weights, every comparison ends in a diplomatic tie.
- →Demanding the strongest case against the recommendation is the debiasing step: it breaks the model's habit of arguing only for its pick.
- →"What should change my mind" converts the answer from a verdict into a decision you can revisit on evidence.
Part by part
- My constraints and what matters most, in order
- Ordered criteria, so the evaluation has a weighting rather than treating every factor as equal. Most decisions are decided by the top two.
- not generic pros and cons
- Rules out the default output, which is a table of general characteristics that would be identical for anyone considering the same options.
- the strongest argument AGAINST each option, including the one you end up recommending
- The clause that makes this useful. A recommendation with no stated cost is advocacy, and the cost is what you need to weigh.
- Flag anything I seem to be missing
- Invites an option or constraint you did not list, which is frequently where the actual answer is.
- what evidence or event should change my mind
- Converts a recommendation into a testable position, and tells you what to watch for after deciding.
Common mistakes
What are the pros and cons of these options?Produces a generic table that ignores your situation entirely. The same output would serve anyone comparing the same two things, which means it decides nothing.
Which of these should I pick?Without stated criteria the model guesses at what you value, usually optimising for cost or popularity. The recommendation is confident and built on assumptions you never made.
Give me an unbiased comparison.Unbiased is read as refusing to recommend, so you get a balanced summary and the decision is handed back untouched. Stating a position with its strongest objection is more useful than neutrality.
Questions
Why ask for the argument against the recommendation?
Because it is the part you will otherwise discover after committing. A recommendation that names its own weakness lets you check whether that weakness is one you can live with.
How should I order the criteria?
By what would actually make you reject an option, not by what sounds important. Cost usually ranks lower than people write down, and reversibility usually ranks higher than they expect.
What if it flags an option I dismissed?
Worth re-examining, and it may be dismissed for a good reason you did not state. The value is being asked, since the reason for excluding something is often the assumption most worth checking.
Should I follow the recommendation?
Treat it as a structured argument to test rather than an answer. It cannot see the political constraints, the team's appetite, or the history that makes one option harder than it looks.