Product design in the age of AI

Photo by Michael Dziedzic.

Software product designers often describe ourselves through the artifacts we produce: sketches, journey maps, annotated mocks, interactive prototypes. As those artifacts become easier to produce, we have to begin describing the profession through the decisions behind them instead.

What is a designer responsible for when making things becomes cheap and available to almost anyone?

AI tools are still far from producing consistently great design work (as of this writing), but they are improving quickly. Given the right context and constraints, I have little doubt models will eventually be capable of producing effective, thoughtful, and scalable design work.

They may become good at evaluating those designs too. Given enough information, a tool could compare possible solutions, identify tradeoffs, and recommend the option most likely to achieve a particular outcome.

I hear people say that taste and judgement are what will differentiate man from machine in this future world, and I disagree with the premise. Human judgment is not magic. It is built from information, experience, values, and constraints. Much of that can be given to a machine too.

But every recommendation depends on how the problem is framed.

A tool might determine which design will improve completion rates. It cannot make completion rate the right goal on its own. That goal came from somewhere. Who decided what mattered most? Whose needs were represented? Which risks were considered acceptable? What happens when the system is wrong?

AI can help us think through these questions. It may even identify things we miss. But people and organizations still decide which answers to act on and which consequences they are willing to accept.

That is where I think the role of product design is heading.

Design has long been conflated with artistry, but the work has never been limited to how something looks. Design is about understanding a problem, connecting competing needs and constraints, exploring what could be done, and carrying that understanding through execution and after effects.

Designers are not uniquely capable of this process, of course. Product managers, engineers, researchers, and others all contribute to these decisions. Designers are also biased, inconsistent, and frequently wrong.

The value of the discipline of design is that it trains us to connect these different perspectives and give them a form others can inspect, challenge, and improve.

The form we present might be a prototype, a diagram, a system, or working software. The artifact is not the value by itself. It makes the thinking behind a decision visible.

As AI makes production faster, some design work will disappear. Some will move into other roles. Teams may need fewer designers whose primary responsibility is translating settled requirements into polished interfaces. It’s an uncomfortable, but plausible, future for a profession I love deeply.

What remains is the need to decide what should be made, why it should exist, what success actually means, and what happens when it reaches the world.

AI will increasingly help make those decisions. It may sometimes make them better than we do. But someone still has to examine the frame and accept responsibility for acting on the answer.

While making becomes cheaper and perhaps deciding does too, accountability does not.