Future of PM: Re-organizing Product Teams

I think most SaaS product teams are still organized for a world where engineering was slow and expensive. Each product area gets a PM, designer, frontend engineers, backend engineers, QA, and a roadmap process built around handovers.

AI changes that cost structure. Non-engineers can build working prototypes. Designers can move beyond mockups. A lot of PM busy work can be automated. The bottleneck moves from delivery to discovery: finding better commercial opportunities, testing them faster, and deciding what deserves real production investment.

That makes me think product teams should become smaller and much more discovery-focused. A modern PM team can be two people: a PM with commercial and user bias, and a product builder with design, prototype, and engineering bias. Their job is to create a working pilot, test it with customers, and learn from real usage.

The handover changes too. Instead of handing production engineering a stack of specifications, Figma files, user stories, and Jira tickets, the PM team hands over a validated pilot product. It is working code. It has been tested with users. It is intentionally quick and not expected to scale.

Production engineering can then do what it is best at: scalability, quality, security, compliance, architecture, and maintenance. Vibe-coded pilots are fast, but they often come with weak QA, technical debt, poor initial architecture, and underdeveloped compliance thinking. A centralized production engineering team can rebuild the validated concept properly.

So I do not think the future is less product management. I think it is more discovery. More market testing. More working pilots. More small teams exploring opportunities, supported by a strong production engineering function that turns validated pilots into durable products.