After years of building products — from Bolkar App to MentorPlus to WheelsEye — I thought I understood how to use AI tools effectively. I was asking good questions. Getting reasonable answers. Moving fast.

Then I changed one thing. And everything shifted.

I stopped asking: "What do you think?"

And started asking: "What am I missing?"

That small shift turned AI from a note-taker into a genuine challenger. And it changed how I build products.

The Problem With "What Do You Think?"

When you ask an AI "what do you think about my product strategy?" — you get agreement. Encouragement. A polished version of what you already said, handed back to you with bullet points.

It feels productive. It isn't.

You're essentially asking a mirror to validate your reflection. The AI has no skin in the game. It has no reason to challenge you. So it doesn't.

I spent months doing this. Getting AI summaries of my own thinking. Feeling like I was moving fast when I was really just running in place.

The Questions That Actually Work

Here is the framework I now use before any important product decision:

"What am I missing?"
Forces the AI to look for gaps in your reasoning, not just confirm what's there.
"What's broken here?"
Specifically asks for weaknesses in your approach — in your logic, your assumptions, your execution plan.
"What won't scale?"
The most underused product question. Your solution works for 100 users. Will it work for 1 million? What breaks first?
"What am I assuming is true?"
This one is brutal. And necessary. Every product decision sits on a stack of assumptions. Most of them are never questioned. This forces them into the open.

Why This Matters for Product Leaders

As a PM, your job is not to generate ideas. It is to make good decisions under uncertainty.

AI tools are extraordinarily useful for this — but only if you use them as a challenger, not a cheerleader.

The shift I am describing is about changing the relationship you have with AI. You are not looking for agreement. You are looking for the thing you cannot see because you are too close to your own work.

At Bolkar, we built India's first voice-led Q&A platform for the next 500 million users. One of the hardest parts of that journey was identifying our blind spots early enough to fix them. We had conviction. We had energy. What we sometimes lacked was someone willing to tell us what was broken before users did.

AI, used correctly, can be that voice. But you have to ask for it directly.

The Practical Shift

Before any major product decision, I now run through four questions:

  1. What am I missing in my analysis?
  2. What assumptions am I making that could be wrong?
  3. What breaks in this plan at 10x scale?
  4. What would a skeptic say about this in six months?

The answers are often uncomfortable. Sometimes they change the direction entirely. Always they make the decision stronger.

AI is not your second brain. It is the voice that should not let you get comfortable too early. But it will only do that job if you ask it to.

Stop asking what AI thinks. Start asking what you are missing.

That is where the real leverage is.