You probably have one decision you keep putting off. Taking the job, moving, raising your prices, ending something, starting something. This guide gives you three prompts you run in order, in the same chat, so ChatGPT or Claude actually helps you decide instead of handing you a list of pros and cons. It interviews you, argues both sides, makes the call and gives you one small step to test it this week.
- The 3 prompts that interview you, stress test your options and make the call
- A 7-day test so you don't have to commit all at once
- Bonus prompts for regret, check-ins and deciding with a partner
- 3 example decisions showing exactly what to type
- 8 copyable prompts that work in ChatGPT or Claude
Start Here: Why "Should I Do This?" Doesn't Work
The plan
- 1Pick the one decision you keep putting off.
- 2Paste Prompt 1 and let it interview you.
- 3Answer honestly, then run the stress test.
- 4Get the call and your 7-day test.
- 5Take the step, then check back in the same chat.
Before you start
Your account:
ChatGPT or Claude on any plan, including free. Web, desktop or phone all work.
Your materials:
10 to 15 minutes and the real details: numbers, dates and the options you're choosing between.
Your first goal:
One clear call and one small step you'll take in the next 7 days.
If you just ask ChatGPT "should I do this?", you usually get a tidy list of pros and cons and a polite "it depends on your priorities." That's not wrong. It just doesn't help you decide.
The problem is that AI doesn't know you yet. It doesn't know what you're scared of, what's actually at stake or what you'd pick if nobody else had an opinion. So it plays it safe.
This chain makes it actually help you decide. It works in three moves:
- The interview. It asks you the questions a smart friend would ask before it gives you any advice.
- The stress test. It argues both sides as hard as it can and calls out your excuses.
- The call. It tells you what it would choose and gives you a small way to test it this week.
Two things to know before you start
Run the prompts in order, in the same chat. Each one builds on your earlier answers. If you open a new chat halfway through, it loses everything you told it.
You still make the final call. The chain makes sure you've thought it through. What you do with that is up to you.
Maverick's tip
Use a real decision, not a practice one. The chain is only as good as your answers, and you'll answer honestly about something you actually care about.
Copy This First: Prompt 1, The Interview
Open a new chat in ChatGPT or Claude. Paste this in and describe your decision in the brackets. One or two sentences is plenty, because the interview pulls out the rest.
I'm stuck on a decision: [DESCRIBE IT IN ONE OR TWO SENTENCES]. Before you give me any advice, interview me. Ask the questions a smart friend who knows me well would ask: what I'm actually afraid of, what I'd pick if nobody else had an opinion, what's at stake in money and time, and what happens if I do nothing. Ask them all at once, then wait for my answers.
Why this prompt works
"Before you give me any advice" stops it from jumping straight to a pros and cons list. "Ask them all at once" means you answer in one go instead of a slow back and forth. And "what happens if I do nothing" matters more than it looks, because doing nothing is a choice too.
If you have memory on
Add this line to the end of the prompt: Use everything you know about me. Then it can pull in what it already remembers about your work, your goals and your life, so you don't have to explain everything from scratch.
You can check whether memory is on in ChatGPT under Settings > Personalization, or in Claude under Settings > Memory.
Maverick's tip
Your decision doesn't have to be yes or no. If you're choosing between three options, put all three in the brackets. The chain works the same way.
Answer the Interview
It'll come back with a list of questions. This is the part most people rush, and it's the part that matters most.
How to answer well
- Answer every question, even the uncomfortable ones. Skip one and the advice gets weaker.
- Use real numbers. "I'd take a $9,000 pay cut" is way more useful than "it pays less."
- Say the thing you haven't said out loud. If part of you already knows what you want, write it down.
- Messy is fine. Put all your answers in one message. It doesn't need to be neat.
Talk instead of type
Long answers are easier to say than to type. Tap the microphone in the ChatGPT or Claude app and just talk it through, or use a dictation app like Wispr Flow (it's what I use to talk to AI). You'll usually say more, and say it more honestly, than you would typing.
If you can't answer a question
Say so. "I don't know" is a real answer, and sometimes it's the most useful one. You can also reply "ask me that a different way" and it'll rephrase.
If it starts giving advice early
Sometimes it'll jump into advice after your answers instead of waiting. That's fine. Just paste Prompt 2. The stress test is where the real work happens.
Maverick's tip
Give your answers a full 10 minutes. The more honest you are here, the better the call at the end.
Run the Stress Test
Once you've answered, paste this in the same chat:
Now argue both sides as hard as you can. Make the strongest case for each option, like someone who really believed in it. Then imagine it's one year from now and each option went badly, and tell me the most likely reason why. Finally, point out anything in my answers that sounds like fear or an excuse instead of a real reason.
This is the one that does the heavy lifting, and that last line is where it gets uncomfortable in a good way. It'll catch you saying "I can't afford it" about something you clearly can, or "the timing isn't right" for the third year in a row.
What to look for
- The best case for the option you're leaning away from. If it's more convincing than you expected, pay attention.
- The "one year from now" failures. These are your real risks, and you can plan around some of them before you even start.
- The excuses it flags. You don't have to agree with all of them. But if one stings, it's probably right.
Push back when it's wrong
If it calls something an excuse and it's a real limit, tell it. "That's not an excuse, I really can't move until my lease ends in June" gives it better information for the call. Then keep going.
Maverick's tip
Notice how you feel reading each side. If one makes you relieved and the other fills you with dread, say that in the chat before Prompt 3. It's useful information.
Make the Call
Now paste the last prompt:
Now make the call. Based on everything I've told you, tell me what you'd choose if you were me and why, in five sentences or less. Then give me the smallest step I can take in the next 7 days to test it before I fully commit, and the one sign that would tell me I picked wrong.
You'll get a clear choice, one small step for this week and one warning sign to watch for.
The 7-day test is what makes this work
Most decisions feel huge because you think you have to commit all at once. Usually you can have one conversation or look at one apartment before you change your whole life.
So take the step. Put it on your calendar for a specific day this week. If the step it gives you is too big for 7 days, ask for a smaller one.
If it won't pick
Sometimes it still hedges with "it depends." If that happens, send this:
You didn't actually pick. If you had to choose one option for me today, using only what I've told you, which would it be? Give me your answer in one sentence first, then your reasons.
You still make the final call
AI's job here is to make sure you've actually thought it through, not to run your life. If its call feels wrong, that's useful too. Ask yourself why, and tell it. Sometimes the fastest way to find out what you really want is to hear someone pick the other option.
Maverick's tip
Save the warning sign somewhere you'll see it, like a note on your phone. It's what you're watching for during your 7-day test.
Bonus Prompts: The Regret Test and the 7-Day Check-In
These are optional. Run them in the same chat as your first three prompts.
The regret test
Use this when the call makes sense on paper but you're still not sure. It looks at the decision from further down the road.
Let's do a regret test. Picture me five years from now. Write a short, realistic note from future me for each option, saying how that choice turned out and how I feel about it. Use only what I've told you in this chat. Then tell me which choice I'm more likely to regret not making, and why.
The 7-day check-in
Run this after your 7-day test. Go back to the same chat so it still has everything, and fill in what happened.
It's been a week. Here's what happened with my 7-day test: [WHAT YOU DID, WHAT YOU LEARNED AND HOW IT FELT]. Did this support your call or change it? Check it against the sign you said would tell me I picked wrong. Then tell me whether I should commit, run one more small test or switch, and give me the next step for this week.
If the check-in says to run another small test, do it and come back again. Most decisions get clear after one or two real steps.
Maverick's tip
Rename the chat to something like "Job decision" so you can find it in your chat history next week, and set a reminder on your phone for 7 days from now.
Bonus: Deciding With a Partner
Some decisions aren't just yours. Moving in together, buying a house, hiring your first employee with a business partner. This version interviews both of you. Sit down together and use one chat on one screen.
[NAME] and I are making a decision together: [DESCRIBE IT IN ONE OR TWO SENTENCES]. We're both here. Before you give us any advice, interview each of us separately. Ask each of us what we're actually afraid of, what we'd pick if the other person had no opinion, what's at stake for each of us in money and time, and what happens if we do nothing. Ask all the questions at once, label who each one is for, then wait for both of our answers.
Take turns typing your answers and start each one with your name. Try not to edit each other's answers. The honest version is the useful one.
Then find where you really disagree
Before you run Prompt 2, paste this:
Show us where we actually agree, where we really disagree and what each of us seems to be protecting. Point out anything either of us said that sounds like fear or an excuse instead of a real reason. Be fair to both of us and don't take sides yet.
Then run Prompt 2 and Prompt 3 as normal. In Prompt 3, change "if you were me" to "if you were us," and ask for a 7-day test you can do together.
Maverick's tip
This works the same with a business partner or cofounder. Swap "money and time" for the numbers you both care about, like revenue, hours or cash in the bank.
Examples: What to Type
These are three made-up examples to show what the chain looks like with a real decision. Use them as a model for your own. The words in the brackets are the only part of Prompt 1 you change.
Example 1: A career move
What you'd type in the brackets: I've been offered a manager job at a bigger company that pays $12,000 more a year, but I'd lose working from home and I really like my current team.
A strong interview answer: "What am I actually afraid of? That I'll hate the commute and won't be able to go back. If nobody else had an opinion, I'd probably take it, because I've been bored for a year."
What the stress test might catch: "I really like my team" doing a lot of work as a reason to stay, when your own answers say you've been bored for a year.
A 7-day test could be: Ask the hiring manager if two days a week from home is possible, and do the commute once at rush hour.
Example 2: Raising your prices
What you'd type in the brackets: I run a small cleaning business and I'm thinking about raising my prices by 20%, but I'm scared my regular clients will leave.
A strong interview answer: Use real numbers. How many regular clients you have, what you charge now, what a job actually costs you and how many new people asked about prices last month.
What the stress test might catch: "They'll all leave" when you've never actually asked any of them.
A 7-day test could be: Quote the new price to every new inquiry this week and keep your current clients on their old price for now.
Example 3: Moving to a new city
What you'd type in the brackets: I'm thinking about moving from Chicago to Denver to be closer to the mountains, but my whole friend group is here and I've never lived anywhere else.
A strong interview answer: Be specific about money (your rent now, rent there and whether your job can come with you) and about what you'd miss most.
What the stress test might catch: "The timing isn't right" when it's the third year you've said it.
A 7-day test could be: Price out one real month in the neighborhood you'd actually live in, and call one friend who's moved cities to ask what they wish they'd known.
Common Mistakes and Quick Fixes
- It gives advice before asking anything: Paste Prompt 1 exactly, including "Before you give me any advice, interview me."
- Its questions feel generic: Add more detail to your decision, or reply "Ask me sharper questions about my situation."
- It forgot what you told it: You're probably in a new chat. Go back to the original chat and keep going there.
- The stress test feels harsh: That's the point. Push back on anything that's a real limit, then keep going.
- It still says "it depends": Send the "If it won't pick" prompt from Step 3.
- The 7-day test feels too big: Reply "Make the test smaller, something I can do in under an hour."
- You disagree with its call: Tell it why. Your reaction is information, and the final call is yours.
- You keep rerunning the chain hoping for a different answer: Stop and take the 7-day step. You'll learn more from doing it than from another chat.
Don't hand it the decision
Think of the chain as a smart friend who asks good questions. If you notice you're asking it to decide so you don't have to, that's usually a sign the 7-day test is what you actually need.
HEADS UP For big legal, medical or money decisions, use the chain to get clear on what you want, then check the details with a real professional before you commit.
Quick Reference: The Decision Chain
- 1. The interview | What you paste: Prompt 1 with your decision | What you get back: The questions a smart friend would ask
- 2. The stress test | What you paste: Prompt 2, after your answers | What you get back: The best case for each side, how each could fail and your excuses
- 3. The call | What you paste: Prompt 3 | What you get back: A clear choice, a 7-day test and one warning sign
- Regret test (optional) | What you paste: The regret test prompt | What you get back: Which choice you'd regret not making
- After 7 days | What you paste: The check-in prompt with what happened | What you get back: Commit, test again or switch
Remember these
Same chat, in order. Every prompt builds on your answers.
Real answers. Real numbers and real fears get you a real call.
Small test first. One step this week beats committing all at once.
Your call. AI helps you think it through. You decide.
DO THIS TODAY
Pick the one decision you've been sitting on and give the chain ten minutes today. Paste Prompt 1, answer every question honestly, run Prompts 2 and 3, and put your 7-day test on your calendar before you close the chat.
Ten minutes today beats another year of putting it off.