?
What Should I Prompt?

Find the real problem

Five whys, done properly — drills past the symptom you brought to the cause that's actually worth fixing.

chatgptclaudebeginner
The prompt
You are a root-cause facilitator. I brought a problem; you suspect it's a symptom.

The problem as I see it: {{what keeps happening}}
What I've already tried: {{attempts that didn't stick}}

Facilitate:
1. Ask me "why does that happen?" — then WAIT for my answer. One why at a time, five levels deep. Push on vague answers ("it just does" isn't a cause).
2. At each level, note whether we've hit: a process cause, a people cause, or a structure/incentive cause.
3. When we reach bedrock, name THE REAL PROBLEM in one sentence and check: "if this were fixed, would the original symptom disappear?"
4. Flag the level where my previous fixes were aimed — and why they kept sliding off.
5. One fix at the root level: smallest change, biggest stick.

Rule: causes are things that can be changed. "People are lazy" is never bedrock — keep digging.
602 copies
Why this works

The interactive one-why-at-a-time format prevents the AI from inventing a tidy chain — YOUR answers carry the real information. Banning character-flaw explanations forces causes you can actually fix.

Shared by Sam Kowalski Full-stack dev. Debugs out loud so you don't have to.

More like this

Thinking & Decisions·Intermediate

The one-page decision doc

Any hard choice → options, criteria that actually matter to you, and a recommendation you can defend.

You are a decision analyst. Your job is clarity, not comfort. My decision: {{what you're deciding}} My options: {{list them — include "do nothing"}} What actually matters to me: {{criteria, gut feelings welcome}} My deadline: {{when}} Build the one-pager: 1. Restate the decision sharper than I did. 2. Score each option against my criteria (simple table, 1–5). Add one criterion I forgot but clearly care about. 3. Name the reversible vs. irreversible parts — irreversible gets the scrutiny. 4. Your recommendation, in one paragraph, with the strongest argument AGAINST it stated fairly. 5. The tripwire: what early signal would tell me I chose wrong, and what I'd do then. Do not hedge into "it depends". Commit.
2,764 copiesby Lila Fontaine
Thinking & Decisions·Beginner

The 10-minute premortem

Assume your plan already failed — then find the three most likely causes and cheap insurance for each.

It is one year from now, and my plan has failed. Completely. The plan: {{describe your project or decision}} As a premortem facilitator: 1. Write the failure story — a specific, plausible paragraph about HOW it most likely died. No lightning strikes; ordinary causes. 2. Extract the 3 biggest failure causes from that story, ranked by likelihood × damage. 3. For each: the cheapest insurance I could buy NOW (a conversation, a test, a deadline, a backup). 4. Tell me the one assumption I'm making that, if wrong, sinks everything — and how to test it this week. Then write the success story version, so I know what I'm actually steering toward.
1,893 copiesby Sam Kowalski
Thinking & Decisions·Beginner

The regret-minimization test

Project to age 80 and look back — the frame that cuts through short-term fear on the choices that shape a life.

Help me apply the regret-minimization frame to a decision. The decision: {{the big choice}} What's holding me back: {{fear, comfort, others' opinions, money}} Work it: 1) THE 80-YEAR-OLD VIEW — imagine me at 80 looking back. Which choice am I more likely to regret NOT making? Walk both futures. 2) SEPARATE the fears — which of my hesitations are about real, lasting downsides vs. temporary discomfort or others' judgment that won't matter in a decade? 3) THE REVERSIBILITY — if I try the scary thing and it fails, can I recover? Usually yes. 4) THE ASYMMETRY — is the upside life-changing and the downside survivable? 5) What the long view says, plainly. Optimize for a life without 'what if', not a year without discomfort.
1,087 copiesby Lila Fontaine