Design Isn't One Thing
A practitioner’s reflection about gen AI in the design process.
JUNE 2026
A product and design leader I know recently confided in me that an engineer on their team had asked a sort-of-provocative question out loud: why hire a designer when we can use AI?
To my surprise, they got caught with an incomplete answer. The type of answer that leaves the question awkwardly open — doubt spreads like wildfire.
The question assumes design is one thing. It isn't. Neither is product. Neither is engineering.
Let me start by saying this is a totally fair question. But the absence of a considered answer from someone with design in their executive title points to something worth examining honestly: a portion of our field has been practicing design without actively thinking about what design is for. AI has made that gap visible in a way it wasn't before. This reflection is my attempt at a longer, more honest answer.
WHAT I’VE OBSERVED IN PRACTICE
There is a process that makes design work, whatever that looks like for your designer. In summary: you understand the context deeply, you explore the problem creatively, you articulate your thinking clearly enough that it can be evaluated, and only then do you arrive at a solution worth building. Skip the process, and everything downstream is weaker — often invisibly so.
I want to be clear: AI changing the design process is not the problem. We are in a moment of genuine transformation — the kind that reshapes a discipline from the inside. No one has the complete answer yet, and design practice will evolve with it. I expect that, and I welcome it. What I find problematic is the shortcut dressed up as process. The impression that thinking happened when it didn't.
What I've seen recently, when product teams use AI to prepare design requirements, is process being skipped immediately. AI-generated wireframes don't seem to consider context, the problem thinking wasn't deepened nor made communicable. It's a finished-looking artifact built on an absent foundation. AI didn't create this problem — shortcuts in the handoff predate generative tools.
AI has made it faster to produce something that resembles considered design output without doing the work that makes it one.
DEPTH OF THE QUESTION
The question has circled the field for decades, and every serious answer has only made design's responsibility larger. Don Norman¹ gave designers the language of mental models — aligning how a user thinks a system works with how it actually works. Richard Buchanan² extended that with the wicked problems framing — design as a discipline equipped to engage problems that resist clean definition. Clayton Christensen³ added Jobs to Be Done — moving design's responsibility upstream into human motivation before touching a solution.
In researching this piece I came across Roberto Verganti⁴, writing from Harvard Business School in 2020, whose argument offered the most current framing I found for what's actually happening to design right now. As AI takes over problem-solving, he argues, human design shifts toward sensemaking — understanding which problems are worth addressing at all. That shift brings design closer to leadership. You can't automate the judgment about which direction matters, or the conviction required to bring people along toward it.
Norman · 1988
Mental models
Design closes the gap between mind and system
Buchanan · 1992
Wicked problems
Design engages complexity that resists definition
Christensen · 2003
Jobs to be done
Design begins with human motivation, not solutions
Verganti · 2020
Sensemaking
Which problems are worth solving — closer to leadership
Each step hasn't replaced the one before it. It has gone deeper. And AI is the latest reason to take that expansion seriously.
ILLUSION OF DEMOCRATIZING DESIGN
Verganti's argument is compelling in one direction: as AI takes over problem-solving, the distinctly human contribution of design shifts toward sensemaking — understanding which problems are worth addressing at all. That shift brings design closer to leadership, which is inherently a sensemaking activity. I find that part clarifying. But I'd resist reducing design.
Design is craft. It is sensemaking. It is empathy, systems thinking, iteration, and the knowing that only comes from building something and testing where it breaks.
All that design is exists in tension, and that tension is generative. There is no clean way to reduce it — and that irreducibility is precisely what makes design hard to replace.
Verganti's critique of design thinking makes this personal for me. Since the beginning of my career, I blindly participated in its evangelism — democratizing the tools and the language, but not always transferring the depth. The judgment didn't transfer. And now AI is doing something similar — making the output of design more accessible without making the thinking behind it accessible. The engineer's question is in part the inheritance of that confusion.
DESIGN’S IRREDUCIBLE ROLE
Design is not one thing AI can replace. It is the irreducible tension between craft and judgment, between making and sensing, between the output and the thinking that gives that output meaning. That's not a single skill or framework. It's a disposition built over years of practice, critique, and making things that break in ways you didn't expect.
The T-shaped model has always described this better for me than "full stack" does. Deep expertise in one discipline, with enough breadth in adjacent ones to span them productively. AI is widening the horizontal stroke — and in practice, it's widening mine too. I use it often for research synthesis, problem framing, prototyping adjacent ideas. The acceleration is real. Yet the depth is not something AI widens on its own. That stays with the practice.
A FINAL NOTE
Anyone in the business could have asked what the engineer asked — why hire a designer when we can use AI. It's not a new question and answering it has always been consequential. Now it feels more urgent to have a considered, and if possible, inspiring answer — the tools have moved fast enough that the visible gap between what AI produces and what a thoughtful designer produces has narrowed.
The judgment about whether the problem was worth solving, the moral weight of what gets built, and for whom — that's where the practice of design lives. Pope Leo XIV said it plainly this year in Magnifica Humanitas: AI is not a morally neutral tool. That responsibility stays with the person who decided what to build.
None of this is an argument against AI in design. It's an argument for honest engagement with what this transformation requires. The tools will keep changing. What design practice owes — to itself and to the people it serves — is the willingness to evolve thoughtfully rather than reactively.
I hope I can inspire that leader from my opening toward a considered answer in the context of their business — one grounded in the frameworks, the thinking, and the process that define their practice.
Design isn't one thing. That's not a limitation. It's the answer.
REFERENCES
¹ Norman, Donald A. The Design of Everyday Things. Basic Books, 1988.
² Buchanan, Richard. "Wicked Problems in Design Thinking." Design Issues 8, no. 2 (1992): 5–21. Visit
³ Christensen, Clayton M. The Innovator's Solution. Harvard Business School Press, 2003.
⁴ Verganti, Roberto, Luca Vendraminelli, and Marco Iansiti. "Innovation and Design in the Age of Artificial Intelligence." Journal of Product Innovation Management 37, no. 3 (2020): 212–227. Visit