Every Product Manager I know is experimenting with AI in some capacity. And most have the same complaint: “The output isn’t that useful.” That’s usually not the AI’s fault. It’s because the prompt wasn’t useful either.
Generative AI is a reflection tool. It reflects the quality of your thinking. Just like a product without clear problem understanding results in wasted effort — poor prompts lead to poor output. The good news? That’s fixable. And if you already think in terms of problems and outcomes, you’re halfway there. Here’s a practical way to make AI work for your product practice — without pretending it’ll do the job for you.
1. Start With the Problem You’re Trying to Solve
Most unhelpful prompts have one thing in common: they jump straight to asking for an output.
“Write me a product strategy.”
“Generate a roadmap.”
“Tell me what features we need.”
But they skip the part we usually insist on as Product Managers: defining the problem first. A better prompt always starts with:
What’s the situation, context, or constraint?
What decision are we trying to make?
Who is this for?
For example, instead of this: “Create a go-to-market plan.” Try:
“We’re launching a B2B analytics tool in 8 weeks. We’ve validated the strongest driver is faster reporting for operations teams. Suggest 3 positioning approaches for the go-to-market strategy.”
That’s not prompt magic — that’s product thinking.
2. Use AI to Challenge and Clarify Thinking
Once you’ve defined the problem, use AI to help you pressure-test your thinking. Here are some useful prompt examples:
“Here’s a draft summary of our product direction. What assumptions am I making that might not be valid?”
“If you were a skeptical stakeholder from Finance, how might you challenge this approach?”
“What customer concerns might this direction fail to address?”
No prompt will give you a perfect answer — but a half-decent one can help you sharpen your own.
📍 Tip: You don’t need to fully believe the AI’s response. Just react to it like you would to a critical comment. That’s where the value is.
3. Use AI to Reframe and Refocus
Good prompts can help you break out of tunnel vision. If a strategy, objective, or idea has been bouncing around for too long, try this:
“Suggest 3 other ways to frame this customer problem — based on outcomes, not features.”
“Rewrite this product goal using the perspective of a user, not a stakeholder.”
“Give me 2 ways to describe the risks of not acting now.”
Even if you don’t use the suggestions, you might find one helpful angle that wasn’t obvious before.
4. Use It to Sharpen Communication, Not Just Generate It
Most strategy and roadmap discussions fail in how they’re communicated — not in the actual logic. If you’re using AI to write for you, pause. Use it to revise, improve, and clarify instead. Try things like:
“Rewrite this product update so it’s clear to an executive who cares about outcomes, not metrics.”
“Simplify this roadmap explanation for a non-technical stakeholder.”
“Turn this stakeholder update into a short summary, without losing the main point.”
Even if it just saves you 10 rewrites, that’s time better spent thinking about the roadmap itself.
Final Thoughts
If you want to get better at prompting, don’t worry about “prompt engineering.” Worry about this instead:
Am I clear on the problem I’m trying to solve?
Does this prompt reflect real product context?
Am I using AI to think better — or to avoid thinking?
Product Management is still mostly a people game. It’s about understanding people, their interests, motivations, and needs. It’s about connecting dots, shaping decisions, and communicating clearly. AI can help…but only if you’re already doing those things well.
Want to sharpen your product thinking, not just your prompting? Check out our Professional Product Management training — where we focus on the skills behind good product decisions.
Got a practical AI prompt you’ve used that worked surprisingly well? Drop me an email — I’d love to learn what’s working for you.
Contact