Will Design Teams Shrink as AI Improves?

I've been sitting with this question for a while now. Not anxiously, but genuinely. The way you sit with something when you're not sure you have the full answer yet but you've been close enough to the edges of it to have a real opinion.

Here's where I've landed: yes, design teams will probably shrink as AI improves. But that's almost the least interesting thing about what's happening. The more interesting thing is what AI is doing to the designers who remain, and what it means for the kind of designer who thrives on the other side of this shift.

I think the answer to that second question is good news. But it's specifically good news for a certain kind of designer, and I want to be honest about that.

The stack is getting smaller

Not long ago, my toolkit looked like a full shelf. Figma for design. Photoshop for image work. InDesign for print. Illustrator for marks and icons. A handful of prototyping tools depending on the fidelity I needed. Each one had its place and each one required genuine proficiency to use well.

That shelf is mostly empty now.

Figma is still there, and it's more capable than ever. AI handles a lot of what I used to open other applications for. Every now and then I'll pull out Photoshop or InDesign for a printed asset, but the occasions are rare enough that I notice them when they happen. The consolidation has been quiet and steady and I suspect most designers working right now have had a version of the same experience.

Fewer tools doesn't necessarily mean less work. But it does mean less overhead, less context-switching, less time spent on the execution layer of a problem. And when you remove those things, something else comes into focus.

What's left when execution gets easier

When the production cost of design goes down, the thing that differentiates good work from average work becomes clearer. And what's left, what was always there underneath all the execution, is taste.

I know that word makes some people uncomfortable. It sounds subjective in a way that feels hard to defend in a business context, where everything wants to be a metric. But taste is real and it's specific and it is not evenly distributed. It's the accumulated result of years of looking at things critically, knowing what works before you can fully articulate why, and having the confidence to make decisions from that place even when no one else sees it yet.

Rick Rubin talks about this in the context of music and I think about it constantly in the context of design. He's not the one playing the instruments. He's the one in the room who knows when something is right. That's not a lesser skill. In many ways it's the hardest skill, because it can't be learned from a tutorial and it doesn't show up cleanly on a resume.

AI has made me better at wielding my taste, not by having taste itself, but by reducing the distance between what I can see in my head and what I can actually put in front of people.

The thing AI changed that nobody talks about enough

Before AI was woven into my process, there was often a gap between having a strong creative vision and being able to communicate it completely. I could describe it. I could sketch it. I could build a rough prototype that gestured at it. But the full thing, the version that made someone else see exactly what I was seeing, took time and effort that wasn't always available.

Now I can move faster. I can pull in data from our business systems, surface relevant conversations, generate a working prototype, and build something that feels real enough to share before anyone has committed to anything. The vision doesn't have to live in my head as long before it can live somewhere people can actually respond to it.

That has changed what it means to be an opinionated designer in a business setting. The friction between having a strong point of view and acting on it is lower than it has ever been. And I think that matters a lot for how design teams get structured going forward.

What actually shrinks and what doesn't

When I think honestly about where headcount will contract on design teams, I think it's in the execution-heavy roles. The work that is primarily about production, moving things through a system, generating variations, resizing and reformatting and adapting assets for different contexts. AI is already genuinely good at most of that, and it's getting better.

What doesn't shrink, what arguably grows in importance, is the role of the designer who can evaluate what gets produced. The one who looks at ten AI-generated options and knows immediately which one is closest to right, and why, and what would need to change to get it the rest of the way there. You cannot do that job without taste. You cannot do it without experience. And you definitely cannot do it with AI alone.

This is why I think the future isn't smaller design teams doing less interesting work. It's smaller design teams doing more consequential work, because the filter between a creative vision and its execution has fewer steps in it.

The designers who should be paying attention right now are the ones whose entire value proposition is production speed. That skill, on its own, is going to be worth less than it was. The designers who should feel good about where this is going are the ones who have developed genuine creative judgment and are excited about finally having tools powerful enough to keep up with how fast they think.

A more opinionated design culture

There's a version of this future I find genuinely exciting.

For a long time, design in a business context has been partly about managing the gap between what a designer knows is right and what they can actually ship given the constraints of time, resources, and organizational bandwidth. That gap has produced a lot of compromised work, not because designers lack vision but because vision is expensive to execute on.

As that execution cost drops, I think we'll see design teams that are smaller but louder. More opinionated. More willing to take a position and build toward it fully because the cost of doing so isn't prohibitive anymore.

That's a different kind of design culture than a lot of teams have now. It requires trust, the kind I've talked about building over years at one company. It requires designers who have taste and the confidence to defend it. And it requires organizations that understand what they're actually getting when they hire that kind of designer.

Not everyone will make that adjustment. Some teams will use AI to cut costs and produce more average work faster. That'll happen and it's probably already happening.

But the teams that figure out how to put real creative judgment at the center of their process, and use AI to extend that judgment rather than replace it, those are the ones I want to be on.

The double-edged sword nobody warned me about

Let me tell you about a presentation I gave recently that went really well, and what happened a week later that made me think harder about all of this.

We needed to update our checkout design. My vision for how to do it was distinctly different from what the executive team had in mind. So I leaned into AI heavily to build my case. I connected data sources, used AI to help analyze patterns and surface conclusions I might have taken weeks to find manually. I built a hypothesis, worked through the numbers, and put together a prototype in Figma Make along with a slide deck that made the whole thing feel real and defensible before a single line of production code was written.

The presentation worked. We changed course. My vision is moving forward. And honestly, the AI-assisted process was a big part of why I could show up to that room with that level of confidence and that much supporting evidence.

Here's where it gets interesting.

About a week later, the executive team came back with another idea. One that felt half-baked to me, one that hadn't accounted for some real product consequences that I could see pretty clearly. When our team started pushing back, the response was something I wasn't quite prepared for.

"Well, can't you just build this real quick with AI?"

I'll be honest, my first reaction was something between offense and whiplash. I had just used AI to build a thorough, well-reasoned case for a better product decision. And now the same technology was being invoked to skip the reasoning entirely.

But after sitting with it for a day, I realized it was actually a useful moment of clarity. What the executive team was doing, probably without thinking of it this way, was using AI as a way to compress the product design process down to execution. Skip the discovery. Skip the thinking. Just produce the artifact. Fast.

And that's the double-edged sword. The same capability that lets an opinionated designer move faster and prove their vision more convincingly also hands a shortcut to anyone who wants to bypass the process that makes that vision worth having in the first place.

This is where I think designers need to be clear-eyed and a little stubborn. AI being fast is not the same thing as thinking being unnecessary. A prototype built in an afternoon is only as good as the thinking behind it. If the thinking was skipped, you've just made it cheaper to build the wrong thing. You haven't made it right.

The opinionated designer's job in this moment isn't to comply faster. It's to hold the line on the process that actually produces good outcomes, and to be articulate about why that process exists. AI gives you better tools to do that work. It does not give anyone permission to skip it.

The honest answer

Will design teams shrink? Probably yes, in some configurations and for some kinds of work.

Will that be bad for designers? For some, honestly, yes. The transition is real and it won't be painless for everyone.

But for designers who have spent their careers developing real taste and real judgment, and who are willing to engage with these tools seriously rather than defensively, I think what's coming looks less like a contraction and more like an expansion of what one person with a strong point of view can actually do.

I've always felt that one of my greatest strengths is knowing what good looks like. For most of my career, the bottleneck between that knowledge and its output was time and execution. AI is removing that bottleneck faster than I expected.

I'm not going to pretend that's not a little exhilarating.