I Asked AI to Play Devil's Advocate on My Decision — It Changed My Mind
How I used ChatGPT as a devil's advocate to stress-test a big decision. A simple prompt that revealed blind spots I completely missed.
The Decision I Was Sure About
Last month, I was ready to quit my stable job and go full freelance. I had savings, a few potential clients, and that gut feeling telling me "now or never."
My friends were supportive. My family said "follow your heart." Everyone around me was validating what I already wanted to hear.
That's when I realized: I needed someone to poke holes in my plan. Not to discourage me, but to make sure I wasn't walking into a wall with my eyes closed.
So I asked AI to be my devil's advocate.
What Is a Devil's Advocate?
A devil's advocate is someone who argues the opposite side of your position — not because they believe it, but to help you see weaknesses in your thinking. It's a stress test for ideas.
The problem? Most people won't do this for you. Friends want to be supportive. Family worries about hurting your feelings. Colleagues have their own agendas.
AI has none of these concerns. It will challenge you without emotion.
The Prompt That Changed Everything
I opened ChatGPT and typed this:
prompt I'm about to make a big decision: [describe your decision in detail].
I want you to act as a devil's advocate. Your job is to:
- Find every weakness in my reasoning
- Point out what could go wrong
- Ask me tough questions I might be avoiding
- Challenge my assumptions
Be direct. Don't soften the criticism to spare my feelings. I need to stress-test this decision before I commit.
I filled in my situation — the freelance plan, my savings buffer, my client pipeline, everything.
The response was uncomfortable.
The Blind Spots I Didn't See
ChatGPT came back with questions I had conveniently ignored:
"What happens if two of your three potential clients don't convert?"
I had assumed all three would sign. But when I ran the numbers with just one client, my "six-month runway" became two months.
"Have you accounted for the mental tax of income uncertainty?"
I hadn't. I sleep badly when money is tight. That affects my work quality. Which affects my ability to get more clients. A spiral I'd never mapped out.
"Why now specifically? Is there external pressure making you feel rushed?"
This one stung. I realized I was comparing myself to a friend who went freelance last year. His success was making me feel behind. Not a great reason to make a life change.
Going Deeper
I didn't stop at the first response. I pushed back:
prompt Okay, let's say I address the client pipeline risk by waiting until I have 2 signed contracts instead of 3 potential ones. Now challenge the decision again with this adjustment.
This back-and-forth went on for about 20 minutes. Each time I fixed one weakness, AI found another angle.
It asked about health insurance (I had forgotten to budget for it). It asked about my contract with my current employer (there was a non-compete clause I hadn't read closely). It asked about seasonality in my industry (my potential clients all had budget freezes in Q1).
The Verdict
I didn't quit my job.
Not because the dream was wrong, but because the timing was. I realized I was three to six months away from being actually ready, not just emotionally ready.
That conversation probably saved me from a financial crisis disguised as a leap of faith.
How to Use This for Any Decision
This works for more than career moves. I've since used the devil's advocate prompt for:
- Whether to buy a car or keep using transit
- Choosing between two apartments
- Deciding if a side project was worth my time
- Evaluating a business partnership offer
The key is to give AI enough context. Vague questions get vague pushback. Specific situations get specific challenges.
One Tip to Make It Stronger
After the devil's advocate round, I add this follow-up:
prompt Now switch roles. Given everything we discussed, what would need to be true for this decision to be a clear yes? Give me a checklist of conditions.
This turns criticism into a roadmap. Instead of just knowing what's wrong, you know exactly what to fix.
The Takeaway
Your brain is wired to seek confirmation, not contradiction. AI doesn't have that bias.
Next time you're confident about a decision, try this: ask AI to argue against it. If your plan survives the challenge, you'll move forward with real confidence. If it doesn't, you just avoided a mistake.
That's not pessimism. That's preparation.
Did you like this?
Every Monday I send one email with a prompt or a tool I tried. No spam, unsubscribe with one click.
Subscribe to the newsletter