How AI leaders are borrowing from the design playbook

As AI changes how we work

, organizations are racing to implement AI widely, govern it responsibly, and make sure it delivers meaningful impact. Companies are discovering how critical—and complex—it is to transform not just workflows and systems, but the teams that use them. Someone needs to connect all the dots.

That’s why roles focused on AI innovation and acceleration are on the rise. They exist to make workflows faster, help teams launch AI-powered features sooner, and speed up tool adoption across the business. Lately, I’ve been talking to leaders who are stepping into these new roles. One person I met wrote a memo to their team about using AI—and now they’re running an executive committee figuring out how it shows up across product, support, and internal workflows. Another has become an AI advisor to their company’s CIO, managing technology investments that ripple across thousands of employees and millions of users.

These leaders have an important mandate, but not an easy one. It’s common to get stuck in what technical leader Kaitlyn Hova calls "performative progress"—integrating new tools to check a box without evolving the systems underneath. But the task goes beyond rolling out tools and products. Leaders need to design a new way of working

. In fact, one pattern from my conversations stands out: The most effective AI leaders aren’t just thinking like technologists. They’re thinking like designers. Here are the design principles they’re using to align teams faster, make better bets, and turn speed and experimentation into something real.

1. Learn the material by using it yourself

Whether it’s pixels, plastic, or AI, one of design’s core tenets is to know your material. And understanding a material requires using it intensively. For leaders with titles that center acceleration and innovation, it’s tempting to default to strategic work. But if you’re only operating at a 30,000-foot view, you’re missing the day-to-day challenges, trade-offs, and tactics that determine what actually sticks. In order to drive tool adoption, you need to dig into those tools yourself.

Most AI leaders I know are prompting and building their own agents as often as they can, in every tool they can find, often at home. Exploring outside of work is one way to push AI’s limits while keeping your company’s systems compliant. For example, I met one leader who was using AI to manage everything in their personal life: planning family vacations, redecorating the house, organizing birthday parties, and even volunteer work. Using AI in this way isn’t just about curiosity or experimentation—AI fluency is part of the job. You can’t lead a probabilistic shift if you haven’t seen how probabilistic tools behave in the real world.

2. Observe in order to understand

It’s not enough to use AI on your own—you need to understand how teams across the business are using it, too. If you don’t know what their process is, you don’t know what their pain points are. Too many leaders stay at the altitude of tools and outputs, but the real signal is in the workflow—where things break, cause friction, or unlock problem solving.

In practice, this means paying attention to signals from across your company: what excites people, what frustrates them, and where they get stuck. These signals show up everywhere, from Slack threads, to survey responses and usage patterns. For example, you might roll out a new AI automation that works perfectly on paper, only to watch adoption flatline. Look closer, and you may find that the issue isn’t the capability; it’s the friction it introduces into an already complex workflow. Teams are routing around it, stitching together their own solutions. If you’re not observing that behavior directly, you’ll miss it.

3. Turn ideas into something real

Design is a balance of observation and action. In fast-moving AI work, ideas don’t fail because they’re bad; they fail because teams can’t visualize them. That’s where prototyping comes in. By turning early concepts into tangible prototypes using Figma Make

, for example, ideas quickly become artifacts teams can react to, refine, and rally around. They’re not just outputs, but coordination and alignment tools that help teams see where things break, what’s missing, and what’s actually worth pursuing. I know more and more AI leaders who are prototyping with their teams

3 ways product teams are building conviction faster with Figma Make

From communicating complex behaviors to pushing the edges of an idea, product managers at ServiceNow, Ticketmaster, and Affirm are using Figma Make to prototype their way forward.

as a way to communicate. Whether it’s a custom AI marketing workflow, or an agent that automates reporting, you’ll get a better sense of the tools people need—and how they should work—by riffing on them together.

And this goes beyond just prototyping. Visual roadmaps, diagrams, mockups, idea maps, and demos can help illustrate not just an end state, but all the steps you’ll need to get there. In fact, I’ve spoken with leaders who can recall specific moments when these types of visuals made everything click. Plus, you can point to something real when you’re making the case for time, resources, or investment.

4. Make critique part of the work

As ideas take shape, how teams refine them matters just as much as the ideas themselves. Leaders who are tasked with shaping AI tools and workflows are turning to a common design practice: critique. Of course, crits aren't just for design—engineering crits

help technical teams align early to solve complex problems.

In practice, implementing a critique process looks like creating spaces—like standing meetings or Slack channels—to evaluate what the team is building and how well it’s working. To start, try sharing a prototype you’re testing, a workflow you’re exploring, or an infrastructure question you’re wrestling with. Ask for feedback and model the tone. The key is to create an environment where feedback feels supportive: Not everything needs to be actioned, just considered. Over time, teams develop a shared sense of what “good” looks like—what outputs are useful, what interactions feel intuitive, what trade-offs are worth making. That shared standard reduces bottlenecks and helps teams move faster without losing coherence. At the same time, coming together as a group can illuminate information that only some have, but everyone needs.

5. Invest in community

AI transformation doesn’t have a finish line. The technology keeps evolving, which means learning has to be continuous—and shared. Design has always thrived on community: working in the open and building on each other’s ideas. It’s why we host Config

; leaders and practitioners come together to expand their perspectives, and leave with a shared momentum and sharper ideas. Meanwhile, participating in internal or cross-industry Slack groups and leadership networks helps you learn from others in real time—so you can see what’s working, what’s not, and where the edge is moving next.

The AI leadership playbook is still being written, but I believe design thinking will remain at its center. AI transformation can’t happen in the abstract—you need to dig into the tools, visualize the details, and invest in ongoing discussion. Design offers a way to bring structure to AI’s complexities and uncertainty so you can move your company’s AI initiatives forward as the landscape continues to shift.