How I’m Finding My Own Rhythm with AI
For a long time, the public conversation around AI in creative work felt a bit flat to me. It was almost entirely about speed, prompt tricks, or knowing which shiny new generator to open. But that misses the point. For me, it isn't about output shortcuts, it’s about changing how I think from the second an idea takes hold.
My background doesn't fit into a tidy box. I’ve run a fashion discovery startup, built interactive installations, and created motion graphics for live concerts. On paper, these projects look like totally different worlds. In practice, they all start in the exact same state: total uncertainty. A vague impulse, a pile of open questions, and zero clarity on what the finished piece will actually look like. That messy, uncomfortable starting line is where I’ve brought AI into my process.
Rather than treating it like a vending machine at the very end of a project to spit out polished images or copy, I bring it in right when I’m most lost - when I’m still figuring out what problem I’m actually trying to solve.
Case Study 1: Sift Landing Page
When I was building my fashion startup Sift, my biggest challenge wasn’t the code or the UI, it was clarity. Sift is an app designed to help people build visual moodboards and find their right outfit. But explaining that fluid, interactive experience in a few seconds to someone who had never used the app before was a real design puzzle.
I first threw raw concepts at AI to test different messaging directions and brainstorm how that carousel should be framed. Should we lead with the creative, tactile fun of building moodboards, or the practical payoff of finding the perfect outfit? How could the interactive carousel walk someone through that experience without feeling overwhelming?
Along with AI, I gathered references and set up visual moodboards in Canva to anchor the aesthetic direction for the app’s landing page. Once the narrative and visual direction felt clear, I took those refined concepts into Figma to design the carousel, build out the UI, and test it with real people. Eventually, we decided using an interactive carousel that would instantly show visitors how the app helps them curate moodboards and pull together outfits.
I wasn't looking for AI to build the layout for me. I was using it to sharpen the narrative and user journey so I knew what I was designing before spending hours pixel-pushing in Figma.
Case Study 2: Motion Graphic in Live Music Concerts
Designing graphics for live music performances always begins with deep listening and musical analysis. Before touches of software ever come into play, I spend time interpreting each music piece to understand the core concepts behind it - digging into why the composer wrote it in a specific way, and why the performers choose to play it with a particular energy or expression. My goal is to best explain and translate those abstract musical ideas visually.
Once I understand the heart of the music, I use AI at that translation point to gather ideas for how those concepts could manifest visually. One direction might explore organic textures and subtle lighting shifts; another might play with rhythmic compositions, high-contrast forms, or abstract color landscapes.
To bring those possibilities into focus, I assemble visual moodboards in Canva, organizing color palettes, visual themes, and key frames for the performance.
However, a striking visual moodboard or generated still image doesn't automatically translate to compelling live graphics. It tells you nothing about rhythm, pacing, transitions, or how the visuals will respond to sound in real time. That’s where tools like TouchDesigner come in - where I take the established visual mood and turn it into motion graphics that respond dynamically to the performance.
AI helps ground abstract musical analysis into concrete ideas I can see. But bringing the performance to life remains a deeply intentional, hands-on craft.
What Connects These Projects?
Looking across product design, publishing, and live visual performance, the deliverables couldn't be more different. But the underlying process is remarkably consistent:
Define → Explore → Evaluate → Build → Validate → Iterate
AI supports almost every stage, but its role shifts. Sometimes it challenges an assumption; sometimes it breaks through a creative block; and sometimes, the most important decision is ignoring its input altogether.
A good prompt isn't a silver bullet. The quality of the outcome depends entirely on the context you bring, the questions you ask, and your ability to critically evaluate what comes back.
For me, becoming AI-native isn't about using AI at every turn or replacing reliable tools just because something new launched. Speed alone is a boring metric.
What excites me is how dramatically it shrinks the distance between a vague hunch and something tangible I can test, critique, or share. It lets me wander down four paths instead of one, move between disciplines fluidly, and make abstract ideas real much earlier.
At the same time, these tools make personal judgment more critical, not less. AI can generate options, but it can't supply taste, intent, or resonance.
The question I ask myself is no longer, "What can I make with AI?"
It’s becoming: "How does working alongside AI change the way I think, create, and bring an idea to life?"
Comments