The buzz around AI agents like Devin, which can take a prompt and handle an entire engineering task, is getting louder. It’s no longer just about code completion; we’re talking about autonomous execution of complex development work.
This isn’t just another productivity tool—it has the potential to fundamentally reshape our core working model: the product trio. For years, the magic has been in the tight, collaborative loop between PM, Design, and Engineering. If the ‘how’ of building becomes increasingly automated by an AI, what does that mean for this partnership? The feedback loop on implementation could become nearly instant. But it also means the quality of our input—the clarity of the problem, the precision of the user story, the ‘why’ behind the what—becomes exponentially more important. Our role might shift from collaborator and negotiator to ‘AI orchestrator.’
This future could free us up to focus more on strategy and discovery, or it could create a dangerous abstraction layer between the problem and the solution.
How are you preparing for a future where your engineering counterparts might be AI agents, and what skills do you think will become most critical for PMs in this new reality?