My Thoughts About Product Design in the Age of AI
I've spent a good part of the last year asking myself a question I didn't expect to be asking after 16 years in this field: does product design, as I've practiced it, still matter?
AI can now generate interfaces, write copy, and stitch together a working prototype faster than most teams can run a single round of user research. It's a fair question to sit with.
Here's where I've landed: designers aren't becoming obsolete, we're becoming architects of AI. A recent survey on the state of AI in design put it well:
Some designers are now doing more PM and engineering work, and vice versa. They're also becoming architects of AI workflows and builders of custom tools. But despite rising expectations for both volume and quality of work, few companies have updated their official policies around performance reviews, team structure, and hiring practices.
That's the part that rings true to me. The tools have moved faster than the org charts. Designers are shipping code, querying company data through agents, and owning tickets end to end, while the job descriptions and career ladders underneath them haven't caught up. But I don't think that gap is a threat to the profession. I think it's proof of the opposite: the work designers know how to do is more necessary right now, not less.
Everything traditional design practice trains into you, systems thinking, reading a business problem correctly, holding user empathy against technical and business constraints, knowing the difference between something that's functional and something that's actually good, is exactly what's required to direct AI rather than be replaced by it. AI can produce output. It can't tell you whether that output is right for the person using it, or whether it fits into everything else the product is trying to be. That judgment comes from the same muscle designers have always been building.
We also carry a specific type of context. Why a flow exists the way it does. What a user actually needs versus what they say they want in a research session. How one feature decision ripples into three others down the line. That context is what turns "AI generated an interface" into "we shipped something people actually love," and it's why designers are well positioned to be the ones directing that conversation, not standing outside of it.
As design capability spreads across product, eng, and PM, I think the designer's role shifts toward being the glue: the person who upholds the vision, protects user empathy, and keeps the craft bar high while also building the AI workflows that let everyone else work within good guardrails. That's a bigger job than the one I started with 16 years ago, not a smaller one.
That's also where our responsibility to users sits. Whatever AI can generate, we're still the ones accountable for whether it's actually good for the people using it. Delivering experiences that are consistent, natural, and predictable, and that hold up in context, isn't a nice-to-have layered on top of speed. It's the job. AI can help us get there faster. It can't decide that it matters.
Source: State of AI Design, "Teams"