Case study // Embrace an inquisitive mindset
Leading an AI-native design practice, human first.
- Role
- Group Design Manager
- Team
- Adobe, Extensibility Platform + Partnerships / 9 product designers
- Timeframe
- 2024 - Current
The challenge
Have you ever had to lead a team through a shift you were still figuring out yourself? That's where I landed when AI tools rolled out across the company and industry-wide, forever changing the shape of our work. My team was facing the same tension every design org is facing right now: move too fast and cheap, and AI-generated work starts eroding the craft bar. Move too cautiously, and you get left behind. There was no shared playbook for any of it. Some designers on my team were ready to dive in immediately. Others were quietly anxious about what it meant for their skills, their careers, even what “good design work” would mean going forward.
While I did also share tops-down mandates from company leadership on AI usage, I most importantly built an environment that made it fun and safe to try out this new technology and build new practices around it.
Blocky is our team mascot, originally built using Figma components to create customizable characters for internal communication. In my journey to learn how to build using AI, I built a fun “builder” for our beloved mascot.
My approach
Leading by example
My own hands-on moment wasn't one single thing, it was a habit. I built a team dashboard myself using Claude Code. I built several skills myself and shared them out with the team. I'm even using AI to build this portfolio site right now. Along the way, I shared what worked and what didn't with my team as I went. That's the part I think matters most: they didn't just hear “Execs want us using AI in our work more.” They watched me try them, experiment, and share back openly what I learned.
Meeting people where they're at
Leading by example only works if you also leave room for people to arrive on their own timeline. Two designers on my team show just how wide that range can be.
The tinkerer/builder
My most AI-forward designer came in with some light development experience and moved fast from day one. He spent weekends testing tools we didn't even have internal access to yet, then brought back what he learned in team meetings and Slack. Pretty quickly, he became the person others turned to when they got stuck on the more technical parts of vibe-coding. I asked him to lean in to this new area that sparked joy for him and I gave him room to lead. He lead learning sessions for the design team and our cross-functional partners on Git, Claude, and Cursor. I knew that designers learning from their peers was going to be more impactful than any sessions I could run.
The steady learner
On the other end, one designer on the team kept treating AI-assisted prototyping as an afterthought. She'd tack it on at the end of a project, mostly to check a box, while still doing the real work of exploring interactions by hand in Figma. Instead of pushing harder, I got curious about why. Turns out she didn't feel like she had the technical footing to move fast enough with the new tools under normal deadline pressure, so she defaulted to what she already knew. I gave her permission to slow down: pushed her timeline back just a tiny bit, connected her with my most AI-forward designer to work through the technical blockers, and gave her one clear, low-stakes goal: start the next project with a vibe-coded prototype instead of ending with one.
She's since gone further on her own, using AI to map out the variations and edge cases across the framework she owns. She was also the first person on the team to use my Impact Clarifier skill for her mid-year self-reflection.
So what changed?
Peer-led learning alongside top-down training
Within the first month of the company-wide rollout, 100% of my team was using AI tools in some form. Not because I mandated it, but because the learning happened sideways. My most AI-forward designer ran most of our learning sessions himself, things like git fundamentals for designers who'd never touched the command line. It wasn't just a favor to the team. It gave him real reps at presenting and leading in front of a group, the kind of visibility that actually moves a career forward.
Slide from presentation shared with the design team.
Code as a design deliverable
We cut the time from an approved plan to a usability test from several days down to three, just by prototyping directly in code instead of Figma. The clearest example was our Embed Framework. Instead of static mockups, we built a vibe-coded template with every Embed SDK module wired up, plus live controls so internal stakeholders could actually manipulate the configurability of each module themselves. Sharing that internally unblocked weeks of back-and-forth with product and engineering over what our framework's modularity actually needed to support.
Wireframe version of screenshot created using Claude AI to obscure sensitive company information.
Tools that compound the team's leverage
Jira could tell me the status of one epic, or one design ticket. It couldn't tell me whether design work across a dozen concurrent projects was on track against the timelines product needed. So I built a dashboard that pulled every design ticket, matched it to its parent epic, and auto-flagged each one as on track, at risk, behind schedule, or missing a date, broken out by designer across the whole team. Before, I only found out a timeline had slipped after it had already slipped, always reactive. Now the moment something dips to “at risk,” I can go straight to the designer and PM and get ahead of it. Proactive instead of reactive, which sounds small until you're the one no longer getting blindsided in a status meeting.
That same effort produced a roadmap planning view the team still relies on. And because a design team without a little joy isn't sustainable, it also includes a character builder for personalizing our team mascot, Blocky.
Wireframe version of dashboard screenshot created using Claude AI to obscure sensitive company information.
AI for judgment and communication, not just production
I built a Claude skill called Impact Clarifier to help designers turn a scattered list of work into a clear, level-appropriate impact statement for performance check-ins. It doesn't write the statement for them. It asks the questions that force clarity, the same way I would in a 1:1.
I built a parallel skill for myself and other managers, one that turns our emails and Slack threads into a clean weekly rollup for our design director. She, in turn, uses her own version to roll every manager's update into one high-signal briefing for our SVP of design. What used to be manual synthesis, repeated at every level of the org, now takes minutes at every level.
Wireframe version of a screenshot from Claude where I used connectors to pull information on a workstream into a weekly brief that I share up to my leadership.
“It really kept pushing me to dig deeper. Reminded me of when Shannon says 'say more.' I appreciated the push to find metrics and quantifiable impact... it broke down my impact according to my level and gave actionable details on what to push on based on my career goals.”
— Taylor G.
“... the /impact-clarifier skill is SO HELPFUL for mid-year check-in...”
— Julia C.
“I also liked the targeted questions and that it asked about my level! I'm going to continue using this to draft impact statements for future check-ins.”
— Elissa W.
So what's my hot take on the future of our jobs as designers and AI?
AI is not going anywhere, but I also do not think it will take over our design jobs as the fear-mongering all throughout LinkedIn say it will. Once the novelty parts wear off, what will be left are the AI tools and practices that people are actually willing to use daily and pay for. AI as a tool that integrates seamlessly within our workflows while keeping humans in the loop for decision-making will become the norm. The designers who treat it like a novelty, something to tinker with, only open when they remember to, are going to get outpaced by the ones who treat it as part of how the work actually gets done. Built into research, prototyping, brainstorming, and communication until it disappears into the workflow instead of sitting on top of it. Understanding that distinction now, while the tools are still new, is what will let designers do their best work in this next era of the industry. Not the specific tools themselves.
A caveat, my builder background
In the early years of my career, long before this AI wave, I was a front-end developer and designer. Most of that specific knowledge has gone rusty by now, but the fundamentals (reading code structure, understanding git, knowing what good front-end craft actually looks like) meant vibe-coding was never a cold start for me. It's a quiet part of why I could lead this by doing the work myself, not just cheerleading it from the sidelines. I have always had a tinkerer mindset, both in and outside of work. I think this growth and learning mindset and resilience to keep trying is key for any designer in this new era ofo ur industry and profession, just as it has been to centuries.
Design is not dead. It's just entering a new era.
Next case study
Modernizing Adobe’s developer console→