Jotting down a couple of hypotheses on how AI affects the practice of product design. They’re thoughts developed from real observation and exposure, salted with ideas from many others. Although not ground breaking, I felt it worth capturing to a least provide a reference point for myself to test or evolve from.

Real-time translation

The benefit offered by the LLM is not that it can create code (or create a PRD, or a design, or a working prototype) it’s that it creates a medium in which experts from a variety of disciplines can collaborate on the actual object, not a model, through their own preferred toolkits.

A designer can manipulate shapes on a canvas that generates code the developer can iterate on in turn changing the shapes the designer is manipulating, all of it being informed by the business rules that can themselves by updated and change the environment, too. (An early term for this I used was a “Living PRD”)

Teams concurrently collaborate, reaching solutions sooner because there’s no handover; the outputs, and insights, are real-time.

Reviews build resilience

The explicit intent of a design review (or any similar process) is to make the end result better. The implicit benefit, though, is in upleveling the team’s own skills. By actively exploring what is meets, misses, or exceeds a goal experts pass along they’re experience to non-experts.

The quality of everyone’s work improves. This is how to make space for those entering a practice to grow and learn.

Workable designs can easily be created by AI, and maybe they are better than those anyone else on the team might individually produce. Highlighting why the design, be it human- or AI-created, works helps reveal the underlying principles; stressing it to show where the design breaks, exposes the wisdom.

The knock-on effect is that these ceremonies, done well, remind people why they matter by highlighting the skills they each uniquely have to share.