Decisions prompt

Decision Matrix Prompt — Pick Between Options Without Endless Debate

When you’ve been debating two or three options for too long, the problem usually isn’t lack of information — it’s lack of structure for comparing what you already know. A weighted decision matrix forces you to name the criteria, weight them, and score each option. This prompt builds the matrix for you and gives you a pick at the end.

✓ Works in ChatGPT, Claude, Gemini ✓ Tested template ✓ Copy + paste ready

The template

Replace the [bracketed] fields with your specifics. Send to any AI chat.

I'm choosing between:
- [option A]
- [option B]
- [option C]

My constraints: [budget, time, risk tolerance, anything that limits the choice]
My goals: [what success looks like]

Build a weighted decision matrix:
- 5-7 criteria that actually matter to MY goals (not generic)
- Weights summing to 100 (justify each weight in one sentence)
- Score each option 1-10 per criterion (be willing to score harshly)
- Total weighted score per option

Then tell me:
- Which one to pick
- What would have to be different for the runner-up to win
- The strongest argument someone could make AGAINST your pick

How to use it

List options that are actually comparable. ‘Apple vs orange vs banana’ is fine; ‘Apple vs orange vs go to school’ isn’t a useful matrix.
Be specific about constraints. ‘Budget is $500/mo total’ is useful; ‘limited budget’ isn’t.
Push back on the weights. If the AI weights criteria in a way that doesn’t match your priorities, say so and rerun — the weights are doing 70% of the work.

Example output

Here’s an example of what a quality AI returns when you run this template:

**Criteria + weights (summing to 100):** 1. **Total monthly cost (25 pts):** You said $500/mo cap — this is hard constraint, deserves heaviest weight. 2. **Time to deployment (20 pts):** You said “need it live in 3 weeks”, and missing this means option doesn’t count. 3. **Integration with your existing Postgres setup (20 pts):** You already have it, switching adds risk + cost not in budget. 4. **Vendor lock-in risk (15 pts):** You mentioned wanting flexibility for future moves — significant but not dominant. 5. **Performance under your load (10 pts):** All three should handle your stated load, so this is a tiebreaker, not a driver. 6. **Documentation quality (10 pts):** Affects your team’s ramp time more than top-line capability. **Scores:** | Criterion | Weight | PlanetScale | Neon | Supabase | |—|—|—|—|—| | Cost | 25 | 7 (175) | 9 (225) | 8 (200) | | Time to deploy | 20 | 7 (140) | 8 (160) | 6 (120) | | Postgres compat | 20 | 6 (120) | 10 (200) | 9 (180) | | Lock-in risk | 15 | 5 (75) | 8 (120) | 7 (105) | | Performance | 10 | 8 (80) | 7 (70) | 7 (70) | | Docs | 10 | 9 (90) | 7 (70) | 8 (80) | | **TOTAL** | **100** | **680** | **845** | **755** | **Pick: Neon.** Cheapest option that meets all the hard constraints, native Postgres so no migration cost, lowest lock-in risk. **What would change the answer:** if Postgres compatibility stopped mattering (you decided to migrate the rest of the stack to MySQL anyway), PlanetScale’s scaling story would pull ahead. **Strongest argument against Neon:** their cold-start time is currently 3-5x slower than competitors. If your load pattern is bursty (long idle then spike), you’ll feel this. For steady load like yours, it’s a non-issue.

Why this prompt works

Specific weights expose hidden priorities.

Without a forcing function, people debate options in the abstract and never realize they’re disagreeing about weights, not options. Making weights explicit reveals the actual disagreement.

‘Be willing to score harshly’ fights the equal-scores trap.

Default AI behavior is to score everything 6-8 (‘it depends on your needs’). Permission to score 3s and 9s produces actual differentiation.

‘What would have to be different for runner-up to win’ is the highest-information question.

It tells you what new evidence would change your mind, which is the real test of a decision’s robustness.

The strongest counter-argument keeps you honest.

Always asking for the strongest case against your pick prevents AI’s pattern of just confirming whatever you seemed to want.

Which AI to use

**ChatGPT** for math-heavy weight calculations and large option sets. **Claude** for the qualitative reasoning — clearer counter-arguments and weight justifications. **Gemini** for decisions involving data you have in Sheets.

Read the full comparison in ChatGPT vs Claude vs Gemini in 2026 →

Frequently asked questions

Isn’t this just over-engineering simple decisions?

For decisions you’ve decided in 10 minutes, yes — skip the matrix. For decisions you’ve been ruminating on for days, the matrix often reveals you’ve been weighing one criterion 5x more than you’d consciously endorse, and the ‘obvious’ option isn’t actually.

Should I trust an AI to pick for me?

Trust the structure, not the conclusion. The value is the matrix itself — by the time you’ve reviewed and disagreed with the AI’s scores, you’ve usually clarified your own priorities enough to decide.

What if my goals are hard to articulate?

Use the [smart friend prompt](/ai-prompt-templates/#21-talk-to-me-like-a-smart-friend) first to clarify what you actually want, then come back to the matrix once you can state your goals in one paragraph.

Can I use this for personal decisions, not just business?

Yes, often better. Personal decisions (relationships, where to live, career moves) suffer the same trap of debating without naming the criteria. The matrix doesn’t make the choice for you, but it surfaces what you’re actually weighing.

What if I have 10+ options to compare?

Filter to 3-5 first. Matrices for 10+ options become unreadable and the AI loses scoring discipline. Pick your top 3-5 by gut feel, then matrix those.

Build any prompt in 30 seconds

Free tool, no signup, runs in your browser. Pick a category, fill in the blanks, copy your prompt.

Try the prompt generator → Read the full guide