Most AI prompts are vague. "Should I switch jobs?" "Is this a good idea?" "What should I do about my client situation?"
AI gives you a generic answer. You shrug. Nothing changes.
The problem is not the AI. The problem is the prompt. It lacks structure. Without structure, AI has nothing to anchor its analysis on. So it defaults to safe, generic advice.
Here is a framework that fixes this. Four parts. Any decision. The pattern does not change.
Part 1: Situation — Where Are You Now?
Describe your current reality. Not feelings. Not hopes. Facts, numbers, and constraints.
Bad: "I'm unhappy at my job and thinking about leaving."
Good: "I'm a mid-level developer at a 50-person company earning $95k. I've been here 2 years. My learning velocity has dropped to zero. My commute is 45 minutes each way. I have 6 months of savings."
The second version gives AI something to work with. Numbers create leverage. Vague descriptions create noise.
Include what is true right now. Your constraints. Your resources. Your hard limits. What you are not willing to do.
Part 2: Options — What Paths Have You Considered?
Don't ask AI "what should I do?" It is bad at generating options from scratch. It will pull generic suggestions from its training data that may not apply to your situation.
Instead, give it 3 to 5 paths you have already considered. Even if they seem obvious. Especially the ones you have already thought through.
Example:
- Stay and negotiate an internal transfer to a different team
- Take the startup offer ($80k + 1.5% equity)
- Stay and build side projects for 6 months, then reassess
- Start a full job search for mid-size companies with better growth
AI is excellent at evaluating options you provide. It is mediocre at inventing them. Give it the raw material. Let it do the analysis.
Part 3: Criteria — What Actually Matters?
What do you care about? Money? Time? Learning? Joy? Stability? Freedom? Growth potential?
Be specific. "Growth" means nothing. "I want to work with technologies that will be relevant in 5 years" means something. "Revenue stability" means nothing. "I can't drop below $4000/month" means something.
If you don't tell AI your values, it will optimize for the wrong thing. It might assume you want maximum salary when you actually want maximum learning. It might assume you want stability when you actually want freedom.
State your criteria explicitly. Rank them if you can. AI will evaluate every option against exactly what you told it matters.
Part 4: Blind Spots — What Are You Missing?
This is where the framework earns its keep.
Ask AI three questions:
- What am I not seeing?
- What risks am I underestimating?
- What assumptions am I making that might be wrong?
This is what AI is genuinely good at. It has been trained on millions of decision-making patterns. It can spot risks, assumptions, and second-order effects that you missed. Not because you are not smart. Because you are one person holding one perspective, and AI has seen a thousand perspectives on similar situations.
The blind spots section is often the most valuable part of the entire output. It surfaces things like:
- "You are assuming your startup equity will be worth something. 90% of startups fail. Factor that in."
- "You haven't considered the reputation risk of leaving your current role during a critical project."
- "You're solving a revenue problem by adding clients. The real problem is your pricing."
Things you would not have thought of on your own. That is the point.
The Pattern in Practice
Here is the complete structure:
Role: You are my [strategic role relevant to this decision].
Situation: [Facts, numbers, constraints. No feelings.]
Options: [3-5 paths you have considered]
Criteria: [What actually matters to you, ranked]
Blind Spots: What am I not seeing? What risks am I underestimating? What assumptions might be wrong?
That is it. Four parts. Any decision. Career, business, product, life. Student, lecturer, developer, founder, PM. The pattern does not change.
Once you internalize it, you stop making decisions from gut feeling. You start making decisions from structured analysis. And the tool you already use every day becomes something more powerful: a genuine thinking partner.