I’m not a product manager, but watching PMs lean into AI has made me think they are in a particularly interesting position.
Good PMs spend their time understanding problems, setting priorities, and making trade-offs. They decide what belongs in the first version and what can wait. These are skills that take practice, and being able to write code does not automatically give you them.
A conversation between DHH and Lex Fridman helped bring this into focus for me. DHH argues that as AI takes on implementation, the ability to decide what software should do, and who it should serve, becomes central to building it.
That sounds like familiar territory for a PM.
What changes is their ability to act on those decisions. An idea that previously needed engineering capacity can now become a working prototype. Something people can try, react to, and help improve.
I like the moment when an idea becomes concrete. It exposes things a document cannot. You discover what is missing, what feels awkward, and what seemed important until you actually used it.
But making this work takes some support from the organization.
Shopify has shared an interesting example with Remote Dev. Their AI agent, River, can build and test mobile apps in remote development environments, with tasks started through Slack. A live preview lets people interact with the running app in a browser. Shopify explicitly describes PMs testing features without having to configure a mobile development environment.
That is the kind of support I think makes this opportunity practical.
Giving a PM access to an AI tool is a start. Giving them an environment where they can explore an idea, try the result, and share it with their team takes it further. They can spend more time applying their product judgment and less time getting a development setup to work.
There is still a learning curve. Working software needs care, and technical judgment matters increasingly as an experiment becomes something people depend on. Engineers have an important role in creating these environments and helping changes find their way into the product.
I think PMs have a strong starting point. They already practice many of the decisions needed to guide the work. Learning to build with AI gives those decisions a more direct path into something real.
And organizations can help open that path.