Most people use AI for decisions the wrong way: they describe the situation and ask what to do. The AI obliges with a balanced, well-structured non-answer, and nothing moves.
These five prompts work differently. Each one targets a specific cognitive failure mode in major decision-making — the blind spots, rationalizations, and miscalibrations that keep people stuck.
Copy them. Fill in the brackets. Run them in a single conversation or spread across sessions.
Prompt 1: The Clarifying Interview
What it’s for: Getting AI to surface your actual priorities before any analysis begins.
The problem it solves: Most people describe a decision but haven’t clarified what they actually care most about. Analysis built on unclarified priorities produces outputs that miss the point.
I'm facing a major decision and I want to think it through carefully. Before we analyze anything, interview me — ask questions that will help you understand my actual priorities, values, and constraints around this choice. Focus on things I might be taking for granted or haven't explicitly stated. Here's the situation: [describe your decision in plain terms, including what you're afraid of and what you want].
Run this first, always. The questions AI generates are often more useful than the answers.
Prompt 2: The Adversarial Pass
What it’s for: Stress-testing your current lean with the strongest possible opposing argument.
The problem it solves: Confirmation bias. By the time you’re articulating a decision, System 1 has usually already formed a preference. You need the best case against it, not a reassuring list of “considerations on both sides.”
I'm currently leaning toward [your option or direction]. I want you to argue the strongest possible case against this choice — not generic caution, but the most substantive, uncomfortable argument against doing this specifically, given what I've told you. Don't soften it.
Pay attention to which arguments you dismiss immediately — those often deserve more examination, not less.
Prompt 3: The Reversibility Map
What it’s for: Distinguishing what’s genuinely irreversible from what just feels that way.
The problem it solves: Most decisions feel more permanent than they are. We deliberate with the urgency appropriate for genuinely irreversible choices on decisions that are actually quite recoverable — and occasionally treat real irreversibilities casually.
Help me map the reversibility of this decision. If I make [choice] and it turns out to be wrong in two years, walk me through: what could I easily undo? What would be hard but recoverable? What would be genuinely foreclosed — paths that close permanently? Be specific. And do the same analysis for the alternative option.
Run this for both options. The alternative to your preferred choice also has a reversibility profile that often gets overlooked.
Prompt 4: The Long-View
What it’s for: Accessing the regret minimization perspective — which choice would you grieve not having tried?
The problem it solves: Present-state bias. The risks that feel enormous at 35 often look small from 72. Research on long-horizon regret shows that omissions (paths not taken) tend to dominate over commissions (missteps along paths taken) over time.
Imagine I'm 75 years old looking back on this decision. From that vantage point, help me think through: which choice would I more likely regret not having tried? What would I grieve if I never did it? And from that long-range perspective, which risks that feel significant right now would look quite small?
This prompt doesn’t override practical constraints. It adds a temporal lens that short-term anxiety naturally excludes.
Prompt 5: The Hidden Assumption Check
What it’s for: Surfacing the assumptions your entire analysis is resting on — including ones you don’t know you’re making.
The problem it solves: Every decision analysis rests on assumptions. Some of them are wrong. The most dangerous ones are the assumptions you’ve never articulated — you’re just operating as if they’re true.
Based on everything I've told you, what are the key assumptions my reasoning seems to be resting on? Include the ones I've stated explicitly and the ones I appear to be taking for granted without stating. For each one, flag whether it seems well-grounded, uncertain, or potentially worth verifying before I decide.
Run this near the end of your session, after the other prompts. It catches what the earlier analysis missed.
How to Use These Prompts Together
For a major decision, run all five in sequence in a single conversation — or across two sessions separated by a night’s sleep.
Start with the clarifying interview. Move to the adversarial pass. Then reversibility. Then long-view. Then assumption check at the end.
The session takes 60–90 minutes. The value isn’t the AI’s conclusions — it’s the quality of your own thinking by the time you finish. These prompts create conditions where you’re engaging with the decision seriously rather than circling it.
Then close the laptop. Sleep on it. Come back with fresh eyes tomorrow.
Related:
- How to Use AI for Major Life Decisions
- The Complete Guide to AI for Major Life Decisions
- 5 AI Decision-Making Approaches Compared
Tags: AI prompts, major life decisions, decision-making, thinking tools, life design
Frequently Asked Questions
-
Do these prompts work with any AI?
Yes — Claude, ChatGPT, Gemini, or any capable LLM. The prompts are platform-agnostic. Paste them in, fill in your specifics, and the structure does the work. -
Should I use all five prompts for every decision?
Not necessarily. For quick decisions, one or two prompts is sufficient. For high-stakes, long-horizon decisions, working through all five in sequence (with time between) produces the most rigorous result. -
How specific should I be when filling in the bracketed sections?
As specific as possible. The more context you give AI about your actual situation, values, and constraints, the more relevant the output will be. Generic inputs produce generic outputs.