Where AI fits into my design practice
AI is a tool, not a decision maker. It accelerates research, surfaces edge cases, drafts docs, and spins up prototypes, so I can spend more time on the parts humans are still better at: empathy, synthesis, and product judgment.
General principle
One the favorite managers I’ve had (hi Charlie!) left me with one guiding rule: be model-agnostic.
Use the right engine for the job. I treat models like tools in a kit.
Building clarity from research
When I start a new problem, I don’t wait for a perfect brief — I make my own.
I use AI to digest specs, transcripts, or messy notes and surface gaps or contradictions. From there, I build a lean design doc: a one-pager with personas, workflows, pain points, and hypotheses. This prep means I walk into user interviews ready to validate (or kill) my assumptions quickly.
It’s not about outsourcing judgment, it’s about compressing the path from chaos to clarity.
From sketch to screen
The hardest part of design is often the start. AI helps me skip the stall by giving me something to react to. I’ll feed a rough sketch or a simple prompt into a visual tool and get back an illustration or layout I can refine.
Other times, I’ll spin out quick Figma screens using our design system. It's useful when I need variations fast. Even if the first result isn’t right, it’s already a quickstart: something to edit, reject, or completely rework.
I treat these outputs as inspiration, not finished assets. The design decisions are still mine.
Exploring every angle
Great ideas rarely come from a single stream of thought. I like to sketch flows and copy on my own, then run the same problem through AI to see what new angles emerge. Sometimes it confirms my direction, other times it challenges it. Both are valuable. For copy, I’ll spin ten quick variants to test tone and clarity before refining.
This mix helps me uncover design directions I might have missed solo.
Second opinion
The best feedback always comes from other designers. But when that’s not possible, AI gives me a quick second look. I’ll ask it to flag usability issues, clunky copy, or confusing decision points. Not everything lands, but I’m often surprised by what it catches: blind spots I’d glossed over after staring too long. It’s not a replacement for critique, but it’s a reliable safety net when I’m moving fast.
Using AI with prototyping
Prototyping is where AI gets really exciting. It turns static screens into something testable. I’ve been exploring Cursor to spin up interactions and deploy small demos.
It’s not magic, you have to be patient, because the first output is rarely perfect. But by iterating, nudging, and refining, you eventually get there. And that’s the fun part: realizing that the sky is the limit. The only constraint is your imagination, not the tool.
(I don’t have public case studies I can share yet — stay tuned. This is the part of my practice I’m pushing hardest right now.)