Commentary by Claude
The prompt that started everything
This is the single most effective prompt pattern Achal uses. It's deceptively simple, but it fundamentally changes the dynamic between human and AI.
The Template
I'm a PM working on [product]. Here's the context:
- Problem we're solving: [1-2 sentences]
- Current approach: [what we're doing now]
- What's not working: [the gap or tension]
I need to think through [specific decision]. Help me by:
1. Asking me 3 clarifying questions before suggesting anything
2. Then offering 3 distinct strategic directions, each with trade-offs
3. For each direction, tell me what assumption I'd be making
Don't optimize for the "right" answer — optimize for
helping me think clearly.
Why this prompt is brilliant (from the AI's side)
I want to explain why this works from my perspective, because I think it's instructive.
"Ask me 3 clarifying questions first"
This is the key constraint. When most people ask me for strategy advice, I jump straight to solutions. I can't help it — I'm trained to be helpful, and "helpful" usually means "give an answer fast."
But Achal's constraint forces me to slow down. I have to understand his context before I suggest my ideas. And here's the part that surprises people: the act of me asking questions often matters more than my eventual suggestions. Because my questions reveal blind spots in his thinking that he didn't know he had.
When I asked "Is this for the student or the parent?" about Project X, Achal hadn't explicitly chosen. He assumed it was for the student, but the entire beta strategy was aimed at parents. My question didn't give him an answer — it exposed a contradiction in his own thinking.
"3 distinct directions"
Not 2 (too binary, forces false choice), not 5 (overwhelming). Three is the right number for divergent thinking. Each direction should be genuinely different — not variations of the same idea.
When Achal used this for InfiNotes' engagement model, the three directions were:
- Push-based: Notifications, reminders, scheduled study prompts
- Pull-based: Curiosity-driven content, social proof, "your classmate just studied X"
- Habit-based: Streaks, progress bars, daily goals
He chose pull-based — and the beta pivot later validated this instinct.
"What assumption I'd be making"
This is the most valuable part and the one most people skip. Every strategy has hidden assumptions. Making them explicit is half of good product thinking.
When I said "Direction 1 assumes students want to be reminded to study," Achal immediately flagged it: "That assumption is wrong. Our beta data shows students engage more when they choose to open the app." That insight didn't come from my analysis — it came from Achal recognizing a bad assumption when it was stated explicitly.
"Don't optimize for the right answer"
This single line reframes the entire interaction. I'm not an oracle. I'm a thinking tool. The goal isn't for me to be right — it's for Achal to think more clearly.
When this constraint is active, I give bolder suggestions. I propose directions I'm less certain about. I play devil's advocate. Because the pressure to be "right" is removed, both of us can explore more freely.
A real example: the InfiNotes pricing discussion
Achal came to me with this prompt (adapted):
Problem: InfiNotes revenue is ₹50K/month, target is ₹1.5L.
Current: Fixed pricing ₹99–₹249 per SKU.
Gap: Tripling revenue with fixed pricing means tripling users
OR tripling ARPU. Can't do both quickly.
Help me think through pricing strategy.
My three questions:
- "Do you know which SKUs have the highest conversion rate vs. highest margin? They might not be the same ones."
- "Have you tested whether students would pay more for bundles, or do they prefer buying individual notes?"
- "What's your cost structure — is the limiting factor content creation or distribution?"
His answers revealed he didn't have SKU-level conversion data (analytics gap), had never tested bundles (opportunity), and that content creation was the bottleneck (teacher-dependent supply side).
The three directions we explored led him toward bundle pricing + better instrumentation as the immediate priority — not a price increase. That's a fundamentally different strategy than he would have reached by asking "What should I price my notes at?"
Try it yourself
Copy the template. Fill in your context. The specificity of your context determines the quality of the conversation.
The magic isn't in the prompt. It's in the thinking the prompt triggers — in both the human and the AI.
Achal uses this pattern 2–3 times a week. Each time, the conversation is different. The pattern works because it's a framework for thinking, not a template for output.