What AI does to the PM role
Let's be direct: a large chunk of what junior and mid-level PMs spend their time on is being automated. Writing user stories and acceptance criteria — done by AI in seconds. Synthesising user interview notes into themes — AI handles it. Creating competitive analysis documents — AI does a 10-minute job in 90 seconds. Generating A/B test hypotheses from usage data — automated. Even creating initial roadmap slides from a strategy brief is increasingly AI-assisted.
This doesn't mean PM roles disappear. It means one AI-equipped PM can do the work that previously required three. That structural compression is already visible in hiring data — PM headcount at major tech companies has been flat or declining even as product output increases.
Which PM skills are safe vs at risk
| Skill | Risk Level | Why |
|---|---|---|
| Product vision & strategic thinking | Low | Setting direction requires judgment, conviction, and organisational context AI lacks |
| Stakeholder management & influence | Low | Navigating org politics is deeply human |
| Customer empathy & discovery | Medium | AI interviews exist but humans build trust differently |
| Data-driven decision making | Medium | AI surfaces data; humans still decide what matters |
| Writing PRDs & user stories | Very High | This is now faster with AI than without |
| Roadmap slide creation | High | AI generates these in minutes from bullet points |
| Sprint ticket management | High | Linear AI, Jira AI handle routine ticket triage |
The PM archetype that survives
The PMs who are thriving in 2026 share a common profile: they spend most of their time in rooms (physical or virtual) making decisions, not documents. They have strong opinions about the market, the customer, and what the product should become. They can walk into a board meeting and defend a strategic bet. They know when the data is telling the wrong story. They are the person other people look to when the answer isn't obvious.
None of that is automatable. What is automatable is the documentation, the synthesis, the formatting — the administrative layer that wrapped around real PM work.
What's your PM risk score?
See exactly which of your product management skills are safe and which need upgrading. Free, 2 minutes.
Check my score →3 skills PMs should build in 2026
1. Strategic narrative and influence
The ability to build a compelling product narrative — a story that connects market trends, customer pain, and company capabilities into a convincing direction — is the highest-value PM skill. "Inspired" by Marty Cagan remains the gold standard for understanding what great product leadership looks like.
2. Quantitative product sense
Not just reading dashboards, but designing metrics systems, understanding causality vs correlation in product data, and building instrumentation. PMs who can interrogate data credibly with engineers and analysts command significant premiums.
3. Deep customer domain expertise
The PMs most protected from automation are those who know their customer's world so deeply that they can spot needs the customer can't articulate. This domain expertise compounds over years — it's the one thing AI genuinely can't replicate at the individual relationship level.