AI is changing product management, but not how I expected.
I sat through my first session on PM and the AI ecosystem, and one framing stuck: The PM’s value is moving from “get things built” to “decide what’s worth building — and prove it works.”
The session broke this into three layers:
1. Productivity amplifier AI can go past writing emails and summarising meetings. The real gain is systems that run reporting, competitor tracking, ticket triage and coordination without you in the loop every time.
2. Product intelligence Knowledge bases used to take months to build and validate. Foundation models have collapsed that timeline. The question is no longer “how do I build the knowledge?” — it’s “how do I use it safely for my customer?”
3. Full-stack product creation AI now touches ideation, design, dev, testing and deployment. But “give one prompt and let AI build the product” is a fantasy. The actual skill is placing guardrails, checkpoints and human judgement at the right points.
One thing I keep coming back to: not every problem needs an LLM. A narrow, repetitive task with structured data is often better served by a smaller task-based model — cheaper, faster, easier to maintain. Foundation models earn their cost when the input is messy, open-ended or language-heavy.
The session’s PM test is the one I want to keep using:
Here’s the problem. Here’s what we’re building. Here’s how we’ll know it works.
And the line that landed hardest: fall in love with the problem, not the solution.
| Session by Arun Nandewal | Masai x IIT Roorkee PM Certification |
