AI gives product teams a chance to scale not just what they produce, but the expertise behind it.
Product design has always evolved alongside the technology we use to build things.
Early on, we designed pages. Data was structured, templates were structured, and our job was largely figuring out the best way to represent that information on a screen.
Then things became more sophisticated.
Pages became collections of reusable components. Components became design systems. Front-end frameworks gave us more flexibility to create richer interactions and more dynamic experiences.
Each step gave product teams more leverage.
AI is providing the next one, and it’s pretty exciting.
And it’s already changing how I build and manage product design teams, and how I approach product design in general.
The opportunity isn't simply to use AI to make designers faster.
It's to remove one of the biggest constraints we've always worked within:
Our own production bandwidth.
Good design has always been constrained by time
A product designer can only explore so much.
You identify a problem. Research it. Sketch ideas. Wireframe them. Explore a few directions. Build the strongest concepts out in Figma. Apply the design system. Think through edge cases. Prototype interactions. Talk to engineering. Get stakeholder feedback. Refine. Test.
That's the work.
And none of it is unnecessary.
But every step consumes time, which means there has always been a practical limit to how broadly a designer can explore.
If a project allows enough time to seriously investigate three approaches, you investigate three.
There may be ten other interesting directions, but eventually you have to move forward.
That's the constraint I'm increasingly interested in removing.
AI becomes a production partner
The way I'm starting to think about AI inside a product team isn't as a replacement for the designer.
It's as an extension of the designer's production capacity.
Imagine I'm working on a music chart.
I wouldn't ask AI:
"Design me a music chart."
That's starting in the wrong place.
I'd start with the work and knowledge the team already has.
Here's our current chart. Here's the design system. Here's how the experience works today. Here's what we've learned from users. Here's a wireframe showing where I think we could go. Here are the problems we're trying to solve. Here are the things we absolutely shouldn't change.
Now explore.
A designer who previously had the bandwidth to deeply investigate three directions might suddenly be able to investigate ten or twelve.
Not twelve random directions generated from a prompt.
Twelve directions rooted in the designer's existing work, judgment and understanding of the product.
That's an important distinction.
The AI is scaling the production. The designer is still supplying the expertise.
I think the design system becomes something larger
This is where I'm beginning to think beyond the traditional design system.
Design systems have been one of the most important advances in modern product design. They give teams a shared language around components, typography, spacing, interaction patterns and visual behavior.
But AI gives us a reason to make that language richer.
I've started thinking of this as a design vocabulary.
A component shouldn't only tell the system:
This is what I look like.
We can begin attaching more meaning to it:
This is why I exist.
This is when I'm useful.
This is the information I should prioritize.
These are situations where I shouldn't be used.
And that vocabulary can extend well beyond interface components.
A product team can bring in brand voice, brand principles, core user definitions, accessibility standards, product strategy, business objectives, research findings and other organizational knowledge that normally lives across documents, meetings and individual people's heads.
The designer's job becomes partially about curating that context.
Not simply giving AI a library of components, but teaching it the environment those components belong to.
That's a direction I'm increasingly interested in building toward.
The workflow becomes a loop
I also don't see designers moving out of tools like Figma and simply describing everything to AI.
The strongest workflow, to me, is hybrid.
A designer might sketch an idea or build a rough wireframe because visually expressing it is faster than writing a paragraph explaining it.
AI can then take that foundation and explore.
The designer reviews the output, adjusts the hierarchy, changes an interaction, combines ideas from multiple versions or takes the concept in another direction.
Then it goes back through the loop.
Designer to AI to designer to AI.
The designer remains in control of the direction.
What changes is how much production work is required between each decision.
That gives experienced designers considerably more leverage.
Experience matters more, not less
There's a tendency right now to equate the ability to generate something with the ability to design it.
Those aren't the same thing.
Anyone can describe a website to an AI tool and get something back.
But generating an interface doesn't automatically give someone an understanding of information hierarchy, interaction design, user behavior, accessibility, business constraints, technical tradeoffs or the thousands of judgment calls that go into building good products.
I've already seen the difference firsthand.
Give the same AI tools to someone with deep product experience and someone without it, and the quality of the direction, refinement and final output can be dramatically different.
AI doesn't remove the value of experience.
It gives experienced people more ways to apply it.
Prototyping can move much earlier
This also changes how I want product teams to prototype.
There has always been some friction between imagining an interaction and determining whether it's technically practical.
A designer might want a particular transition, behavior or interaction, but exploring it beyond a static mockup often means involving engineering.
AI can help move some of that exploration earlier.
Designers can prototype interactions, investigate implementation approaches and get much closer to the intended experience before asking engineering to invest significant time.
That's not about bypassing engineers.
It's about bringing engineering a more developed idea.
Instead of:
"Can we build something like this?"
The conversation can increasingly become:
"Here's what we're trying to accomplish, here's a working exploration, and here are the areas where we need your expertise."
That's a much better use of everyone's time.
Research gets leverage too
I see a similar opportunity with research.
AI shouldn't replace talking to users, analytics, testing or direct observation.
But designers have always faced the same bandwidth problem there as well.
There's more information available than anyone has time to fully investigate.
AI gives product teams another way to synthesize that information, explore adjacent research, identify patterns and ask better questions.
Again, the value isn't replacing the designer.
It's expanding the amount of context the designer can realistically work with.
This changes how I think about product teams
For years, we've improved the systems surrounding product design.
Templates gave us consistency.
Components gave us reuse.
Design systems gave us scale.
Now I think we're entering a phase where AI lets us scale something much more valuable:
The expertise behind the system.
That's where I'm heading with product teams.
I want designers building strong foundations, understanding their users deeply, creating thoughtful design systems and developing clear points of view.
But I also want them turning that knowledge into something AI can work with.
Giving it the existing product.
Giving it the rules.
Giving it the brand.
Giving it the business context.
Giving it the research.
Giving it an initial point of view.
And then using AI aggressively to explore, prototype and iterate beyond what the team's production bandwidth would traditionally allow.
The role isn't becoming less important.
The leverage is becoming much larger.
And for me, that's what makes this next phase of product design so interesting.
