Prompts for thinking through a decision

What it is good at

• Good at surfacing unstated assumptions behind a decision you think is already reasoned through.
• Useful for genuinely arguing the opposite case rather than a token counterpoint.
• Helps separate a decision into its component parts, facts, values, risk tolerance, so you see what's actually driving it.
• Fast at listing what information is still missing before a decision should be made.

Where it falls short

• It has no stake in the outcome and no access to your gut instinct, treat it as a structuring aid, not the decision-maker.
• It will construct a plausible-sounding case for almost any position if asked, so a persuasive answer isn't proof of a good one.
• It cannot know your risk tolerance or values unless you state them explicitly.

How to use it for creative work

1. 01

   State the decision precisely

   Write the actual decision, not the general situation, 'should I raise prices' is different from 'should I raise prices before or after the new collection launches'.

2. 02

   Ask it to find your unstated assumptions first

   Before any options or recommendations, get it to list what you seem to be assuming that hasn't been checked.

3. 03

   Get the strongest case for each option, not a pros/cons list

   A real argument for each path forces sharper thinking than a bullet-point comparison.

4. 04

   Ask what would change your mind

   Naming the piece of evidence that would flip the decision either way clarifies what actually matters here.

5. 05

   Separate the decision from the feeling

   Ask it to distinguish which parts of your hesitation are about facts you don't have versus values or risk tolerance, these need different fixes.

Ready-to-use prompts

Here's a decision I'm facing: [describe decision and context]. Before suggesting anything, list the assumptions I seem to be making that I haven't actually verified.

I'm deciding between [option A] and [option B] for [context]. Make the single strongest case for A, then the single strongest case for B, as if you had to convince a skeptical version of me of each.

My current leaning is [describe leaning] on this decision: [describe decision]. What single piece of information, if I learned it, would be most likely to change that leaning?

Here's why I'm hesitating on this decision: [describe hesitation]. Split my hesitation into: things I don't know (facts I could find out) versus things I do know but feel uncertain about (values/risk tolerance). Treat them separately.

Here's a decision: [describe]. Argue how this will look in hindsight in 1 month, 1 year and 10 years if I choose [option], versus if I choose [other option].

Assume it's [timeframe] from now and [decision] has failed badly. Write the most plausible explanation for why it failed, based on what you know of the situation: [context].

Play the role of [specific type of advisor, e.g. a blunt operations mentor / a risk-averse family member / an ambitious peer founder] reacting to this decision: [describe]. Give me their honest, in-character reaction.

Before I look at any specific options, help me define what actually matters for this decision: [describe context]. List 4-5 criteria I should be weighing, and ask me to rank them before we go further.

For this decision, [describe], how reversible is each option realistically, and does that change how much time I should spend deciding versus just acting and adjusting?

Here's my current thinking on this decision, written out messily: [paste rough notes/thoughts]. Play it back to me as a clean, structured argument, and tell me where the logic has a gap.

Questions people ask

Should I just ask AI what to do?

Asking directly tends to produce a confident-sounding but shallow answer, better results come from asking it to structure the decision, argue both sides and surface assumptions, then deciding yourself.

Can it help with decisions under real time pressure?

Yes, the pre-mortem and reversibility prompts work well fast, they take under a minute to run and often clarify whether a decision even needs lengthy deliberation.

Does it help to tell it my gut feeling upfront?

Yes, but ask it to stress-test that feeling rather than agree with it, state your leaning, then explicitly ask what would change your mind.