Unit 10: AI in product management: what changes and what doesn't

Learning objectives

By the end of this unit, you should be able to identify the specific PM tasks that AI accelerates, understand the foundational PM skills that remain human-driven, and evaluate when AI tools are helpful versus when they create false confidence.

Video script

Reading material

What AI does well for PMs

Data synthesis and pattern recognition across large datasets. Customer feedback analysis. First drafts of documents (specs, emails, presentations). Competitive monitoring. Prototyping and mockup generation. Quantitative analysis and chart generation.

What AI does poorly for PMs

Understanding context and nuance in customer feedback. Making trade-off decisions with incomplete information. Navigating organizational politics and stakeholder relationships. Developing product intuition. Saying no to compelling-but-wrong requests. Creating and communicating product vision. Reading between the lines in customer conversations.

The false confidence risk

The biggest risk of AI tools for PMs isn't that they'll replace you. It's that they'll make you overconfident. When AI generates a polished customer analysis in minutes, it looks authoritative. When it drafts a spec that reads well, it feels complete. When it produces a competitive landscape summary, it seems comprehensive.

But polished isn't the same as correct. Well-written isn't the same as well-thought-through. Comprehensive isn't the same as insightful.

I'd suggest a simple rule: the more polished the AI output looks, the more skeptically you should review it. Polish is what AI does best. Substance is what you need to verify.

Building your AI toolkit

Start with the tools that save time on low-judgment tasks. AI transcription for customer interviews. AI summarization for meeting notes and feedback. AI drafting for first-pass documents. AI analysis for large datasets.

Then experiment carefully with higher-judgment tasks. Use AI to generate hypotheses about user needs, but validate them through real research. Use AI to suggest prioritization criteria, but make the actual decisions yourself. Use AI to draft strategy options, but evaluate them with your own strategic thinking.

Over time, you'll develop a sense for where AI adds value in your specific workflow and where it creates more work than it saves. That calibration is personal and evolves as the tools improve.

Practical exercise

Exercise: AI audit of your workflow

If you're currently working in a PM or adjacent role, audit one week of your work. Categorize every task into three buckets:

1. AI could do this entirely (data processing, transcription, formatting)
2. AI could assist with this (first drafts, analysis, summarization)
3. This requires human judgment (decisions, relationships, interpretation)

If you're not yet in a PM role, do this exercise for a role you currently hold, or audit a day in the life of a PM based on blog posts or day-in-the-life content.

Write a brief reflection: where would AI save you the most time? Where would it be tempting to use AI but risky? What skills should you invest in developing because AI can't replicate them?