No Coding Required โ€ข 100% Free Setup 6 min read โ€ข November 15, 2026

The 'Pre-Mortem' Decision Prompt: Force AI to Find Your Blind Spots Before You Launch

By default, AI chatbots are trained to be agreeable sycophants โ€” if you pitch a flawed business idea or risky career move, they will cheerlead you right off a cliff. Here is the exact 'Pre-Mortem' prompt that forces Claude or ChatGPT to act like a ruthless, constructive sceptic and save you months of regret.

1. The Real-Life Problem (And Why Default Habits Fail)

โš ๏ธ The Daily Friction

You are about to make a high-stakes decision โ€” pitching a new project to leadership, changing jobs, or launching a side business. You ask AI 'What do you think of my plan?' and it replies with 5 paragraphs of generic praise that completely misses the fatal flaw.

๐Ÿ›‘ The Common Trap

LLMs are RLHF-tuned for helpfulness and agreement. Unless you explicitly invert the frame ('Assume this failed 6 months from now'), the model mirrors your optimism instead of stress-testing your assumptions.

2. Three Simple Setup Clicks

1
Write Your Plan + Your Unspoken Worry in 4 Sentences
State what you plan to do, your timeline, your constraints, and the 1 thing you are secretly nervous about.
2
Run the '6-Month Pre-Mortem Red-Team' Prompt
Paste the prompt below into Claude or ChatGPT. It forces the model to fast-forward 6 months, assume the plan failed, and work backward to diagnose the exact root causes.
3
Add the 3 Safeguards Before You Commit
Use the AI's 'Pre-Emptive Fix' table to add concrete circuit breakers or scope cuts before presenting your plan.

3. Real-World Transformation (Before vs. After)

โŒ Default 'What Do You Think?' Prompt Output
"That sounds like a fantastic and ambitious Q4 migration plan! Your timeline is proactive and your team will surely appreciate the modern stack. Good luck!"
โœ… Pre-Mortem Prompt Output (Spots the Real Trap)
"๐Ÿ’€ Failure Cause #1 (75% probability): Dual-running both billing pipelines during November peak traffic without automated shadow-diffing will freeze 2 engineer-weeks on manual reconciliation. ๐Ÿ›ก๏ธ Fix Today: Ship read-only shadow comparison for 1% of traffic before cutting over."

4. Copy-Paste Tested Prompts

โœจ The 6-Month Pre-Mortem & Blind-Spot Detector
Do NOT compliment my idea or be polite. Act as a battle-tested Chief Operating Officer and skeptical advisor. Here is the decision / plan I am about to execute: """ [DESCRIBE YOUR PLAN, GOAL, TIMELINE, AND RESOURCES] """ Fast-forward 6 months into the future and assume this plan ended in a painful failure. Give me a ruthless, constructive breakdown: 1. ๐Ÿ’€ THE 3 MOST LIKELY CAUSES OF FAILURE: Rank them by probability and explain the hidden second-order effect I am ignoring. 2. ๐Ÿชค THE SUNK-COST TRAP: What early warning sign at Week 3 will tell me this isn't working? 3. ๐Ÿ›ก๏ธ THE PRE-EMPTIVE FIXES: For each failure mode, give me 1 concrete adjustment I can make today to de-risk it by 80%. 4. โ“ THE 2 HARD QUESTIONS: What 2 questions will my harshest critic ask me about this plan?

5. Your 60-Second Quick-Start Checklist

Do This Right Now:
  • โœ“ Pick 1 upcoming decision, proposal, or purchase you're considering this week.
  • โœ“ Paste it into the '6-Month Pre-Mortem' prompt.
  • โœ“ Write down the 'Week 3 Early Warning Sign' on your calendar.
  • โœ“ Prepare your 2-sentence answer to the '2 Hard Questions' before your meeting.

6. Pro-Tip & Privacy Safeguards

๐Ÿ’ก Pro-Tip

After the AI lists the 3 failure modes, reply: 'Now roleplay as my skeptical VP in a 3-turn Q&A and push back on my answers.' Five minutes of sparring makes you bulletproof in the real meeting.

๐Ÿ›ก๏ธ Privacy Rule

Anonymize company names, customer contracts, and proprietary dollar figures as [Company A] or [$X MRR] before pasting.

โšก Explore More Everyday AI Productivity Guides

Browse all practical personal automation guides, technical benchmarks, and architecture breakdowns in the library: