Designing the Shift: What It Really Takes to Build with AI
by Sara Vienna, Chief Design Officer @ Metalab
We’ve all felt it. That low hum of disruption shaking the walls of creative work. AI has crashed into the industry, loud, fast, and full of promise. But behind the flashy demos and overnight feature launches, the real shift isn’t just technological. It’s philosophical.
For those of us in design, AI isn’t just changing what we make, it’s changing how we think. How we collaborate and how we define originality. At Metalab, we’ve spent the last 18 months experimenting with AI not just as a toolset, but as a design partner. And what’s become clear is this: embracing AI isn’t about speed. It’s about mindset.
Because faster doesn’t always mean better. In fact, speed without clarity can be actively harmful. Efficiency alone doesn’t move culture forward. Meaning does. And meaning takes work.
So, what’s actually changing? And what does it take to stay relevant without losing the soul of your creative process?
Let’s talk about it.
Mindset over mechanics.
The biggest misconception I hear is that AI will replace designers. The truth? AI will replace designers who aren’t willing to evolve. In our work, we can already see the ability to turn a concept to prototype in a prompt. Every design team today should be asking “what’s the right tool for the job?” I’ll warn, though, just because you can prompt it, doesn’t mean you should.
Creative value doesn’t come from moving things around a screen. It also doesn’t come from lazy prompts with crappy outputs. It’s in knowing what’s worth making in the first place. It’s knowing what looks and feels great.
At Metalab, we’ve started treating AI like a creative co-pilot. That means less time clicking around, more time thinking deeply. More decision-making. More divergence early on. And more room to explore unconventional paths before we lock in the obvious one. Thinking of recent projects, it’s not only for going wide and diverging in visual opportunities. Prompting is now becoming part of our design systems language, making our systems more sustainable to stand the test of time.
We’re also asking harder questions of ourselves: What is the value of human contribution when so much can be automated? What are we bringing to the table beyond execution?
That kind of reflection is uncomfortable, but necessary. And it’s how we build better, more intentional work.
Designers are no longer just makers, we’re interpreters, curators, meaning-makers. In a sea of sameness, we’re the ones who decide what’s worth keeping, what’s worth remixing and what’s worth leaving behind.
Workflow: less linear, more layered.
If you’ve been inside a modern design process lately, you know how messy it’s become. TBH, I’m not mad at it. Brand work bleeds into product work. Research runs parallel to execution. Everyone’s designing, writing, and validating all at once.
In our brand practice, we’re using AI to generate moodboards, refine naming and brand explorations, prototype, and make design systems faster than ever. But it’s not about speeding through steps. It’s about looping through them with more fluidity. Testing how ideas hold up across different contexts, finding their edges, and refining with more confidence. It helps us push past safe ideas faster, and gives clients a clearer view of what their brand could grow into.
We’re building internal tools that plug AI into our existing Figma flows, letting us reframe use cases, rewrite microcopy, or auto-generate variant UI in seconds. That means more room to explore, refine, and adapt, without waiting for a creative bottleneck to break.
We treat this stage like a sandbox. We test, tweak, break things, and rebuild. Because experimentation isn’t a nice-to-have, it’s the cost of staying relevant. We even have to get more comfortable with mistakes, or time “wasted” with emerging AI processes, because the old school way would have been faster and better. In one recent project, we f-ed around, tried it, and binned the results because they sucked.
Messing around is also how we future proof our thinking. Rigid systems break under pressure. Generative systems flex and evolve. Products now need to respond in real time, to people, to context, and to culture. In turn, so does our process.
Personalization is finally practical.
One of the most promising areas we’re seeing is in content scaling. Everyone talks about personalization, but most brands never had the infrastructure, or the budget, to pull it off meaningfully.
Now, with the right design systems and AI models in place, we’re seeing teams unlock entire content pipelines that feel custom-built for their audiences.
The impact isn’t just cost savings, it’s relevance. That’s the real power of AI at scale, not quantity, but contextual quality.
We’re moving from the attention economy to the meaning economy. And in that shift, personalization isn’t about volume, it’s about depth. The brands and experiences that thrive will be the ones that feel like they actually see you. Take our client Suno, they’re the fastest growing AI music app in the App Store. Why? They’ve unlocked the joy of making, discovery, and fun with Gen AI.
Client relationships are evolving, too.
Here’s something that doesn’t get talked about enough: how AI is changing agency–client dynamics.
Historically, we were hired as thought partners, with an output of execution. Now, we’re being brought in as operational collaborators, helping clients navigate which tools to trust, how to integrate them, and what principles to lead with. It’s not just about handing over design anymore. It’s about imagining a better future, defining the strategy to get there, and executing alongside their internal team.
Some of our most exciting work lately has been in co-building AI-assisted products with clients, where we’re designing not just the interface, but the very role AI plays in the experience. That requires deep alignment on user expectations, ethical guardrails, data usage, and failure states.
One guiding principle we hold close: AI should never obfuscate intent. If a machine is influencing outcomes, that should be clear to the user. We prioritize transparency, fallback states, and a human-in-the-loop mentality wherever possible. No matter if it’s an internal tool/B2B, or B2C.
We also remind clients that AI isn’t a shortcut to differentiation. It’s a multiplier. If your brand voice is fuzzy, your AI voice will be, too. If your product value is unclear, AI won’t magically fix that. You have to get the fundamentals right first.
It’s a new kind of relationship. One where we’re not just delivering work, we’re helping shape how companies internally define creativity, accountability, and innovation in real time.
That means advocating for transparency. For opt-out paths. For trust that’s earned, not assumed. We’re not here to hand over control to machines. We’re here to design experiences that deserve it.
Creativity isn’t dying. It’s expanding.
We’re in a weird, wonderful moment. One where the tools are getting stronger by the day, but the rules are still being written and broken.
What’s exciting is that creativity isn’t going away, it’s just becoming more layered. More strategic. More interdisciplinary. More about what you make, less BS to get there.
The teams who thrive won’t be the ones who master every prompt or automate every process. They’ll be the ones who stay curious. Who think critically. Who know when to go wide, and when to go deep.
At Metalab, we’re not pretending to have it all figured out. But we are showing up with intention. We’re building the systems, habits, and shared language to make AI part of our practice, not a disruption to it.
Because in the end, great work still comes down to the same thing it always has: clarity of thought, quality of taste, and a team brave enough to try something new.
In a world of infinite output, the edge isn’t speed or polish. It’s taste. It’s meaning. It’s knowing what deserves to exist.
And that? AI can’t do for you.
Sara Vienna, Metalab
