There’s a version of this story where I say I saw the future, immediately understood the potential of AI, and seamlessly integrated it into my design practice.
That’s not what happened.
If I’m being honest, my initial reaction was… indifference.
Not because it wasn’t interesting—but because we’ve been here before.
Every few years there’s a new tool, a new technology, a new headline declaring that design is either saved or doomed.
And somehow, we’re still here. Designing.
So I treated AI the same way I treat any new tool: cool—let me know when it actually makes my job better.
Cut to six-ish months later… It wasn’t a training, a keynote, or a think piece about the future of work.
It was a hackathon.
A company-sponsored, time-boxed, slightly chaotic hackathon where we were encouraged to use tools like Microsoft Copilot and Figma Make to build something new.
And that changed everything.
Not because I suddenly became “good at AI.” But because I had the space to actually sit with it. To try things. To fail without consequence. And, if I’m being honest, just enough pressure from my own ego to make something work.
Up until that point, I already had a full toolbox I trusted. I knew how to produce good work. There wasn’t a real incentive to struggle through a new tool that felt uncertain.
The hackathon gave me a different kind of constraint: use this—and still deliver something you’re proud of.
So I had to figure it out.
What actually changed
Not everything. And that’s kind of the point.
AI didn’t rewrite my process. It didn’t replace the work. But it did change the shape of it—and the speed at which I can move through it.
The biggest shift is how quickly I can get to the core of things. What’s actually being asked. What’s actually being said. What matters, and what’s just noise.
There’s a lot of noise.
AI has become a way to cut through it—not by deciding for me, but by giving me something I can react to faster.
It’s also helped with something I’ve struggled with for a long time (and likely you do too): the blank page.
AI gives me a starting point. Not perfect. Not final. Sometimes not even that good. But enough.
What I actually use it for
In practice, it shows up in pretty unglamorous ways. Not as a replacement for the work, but as support in very specific moments:
When I’m staring at a pile of notes or inputs, it helps me find a thread—something resembling a through line—faster than I would on my own. Not the truth, but a version of it. And then I do the work of deciding if it holds up.
There are times when I know what I think… but I can’t quite say it yet. And staring at a blinking cursor on a blank page is paralyzing. AI gives me language I can react to. I can feel when it’s off. I can feel when it’s getting closer. That back-and-forth has become part of how I think things through.
It’s also become a kind of critique partner. One that, left unchecked, is almost aggressively supportive. Because left to its own devices, AI will happily tell you you’re brilliant. I’ve learned to ask it to push back, to be more critical, to give me something I can actually respond to. It’s… nicer than my professors were in design school but still pretty useful.
Lastly, there are some great rapid prototyping tools. You had to see that one coming, right? It’s faster than ever to make something that feels real enough to react to, to share, to align around. Not production-ready, but directionally useful.
Where it falls apart
It’s not magic. And it’s definitely not reliable in the ways people sometimes imply.
It lies. Confidently.
There are moments where it says something with its full chest that is just… wrong. Not slightly off—completely fabricated. The confidence is impressive. But I’ve learned to ask it to prove what it claims it’s done.
Show me receipts.
It gets close—but not quite there
It gets close enough to feel usable, but not close enough to actually keep. Whether it’s writing or visuals, there’s always a layer of refinement that can’t be skipped.
It defaults to groupthink
You can also feel when it’s defaulting to something generic. Something polished and reasonable and… forgettable. Like it was assembled out of the internet’s most common (read as boring) opinions. That’s usually when I know it’s time to step back in.
If everything sounds right,
but nothing feels interesting…
we’ve probably lost something.
The part that still matters most
Using AI well isn’t really about prompting.
It’s about judgment.
It’s knowing when something is useful, when it’s misleading, when it needs to be pushed further, and when it needs to be thrown out entirely.
AI can generate options. It can suggest directions. It can help you get unstuck. But it doesn’t (by default) know your users. It doesn’t know your product. It definitely doesn’t know your company politics or the context behind a decision.
That’s still on you.
And honestly, that part doesn’t go away. It just becomes more important.
A moment that stuck with me
There was a time I used AI to help articulate a point of view I hadn’t fully figured out yet. I wanted early feedback—before I invested too much in one direction.
I spent an afternoon, conversing back and forth with Copilot to get something close to what I was trying to say—a shitty first draft—and sent it out to design, research, and content peers.
Leave it to my content buddies to call it out as "AI slop." Instead of reacting to the idea, they started reacting to the writing—the phrasing, the structure, the tone. Which, fair. And accurate.
I could have asked them to focus on the ideas instead of the writing. But why?
"Yes, this was 100% written by AI based on some back and forth—but yes, you're right."
That exchange stuck with me. The way the use of AI shaped not just the work, but how the work was evaluated.
I don’t have a clean takeaway from that yet. But I do think more carefully about what I send—and what I’m actually asking people to react to.
The “AI expert” thing
This is the part I struggle with the most.
We’re early—really early—in this technology. AI is still an infant (or maybe a know-it-all pre-teen, but less moody…is it actually a college freshman? An infant adult?).
And yet, there’s a rush to claim expertise. To define best practices. To speak in absolutes about where this is all going.
I don’t know how you become an expert in something that’s still changing this fast.
That doesn’t mean people don’t have valuable perspectives. They absolutely do. But I’d rather be someone who is paying attention, doing the work, forming a perspective in real time—and being open to being wrong.
That feels more honest.
Where AI actually makes me better
Despite everything, I do think AI makes me better at my job—but not in the way I expected.
It hasn’t made me a better designer by doing the work for me. If anything, it’s made me more aware of how I work.
It’s highlighted the gaps—what I don’t know, where I need more context, where I’m making assumptions. It’s helped me get up to speed faster in areas outside of design, whether that’s finance, engineering, or whatever space I’m operating in. But more than that, it’s forced me to be clearer—in how I frame ideas, how I explain them, how I pressure-test whether they actually make sense.
What I didn’t expect is how much it would change how I think, not just how fast I move.
It made something obvious that I hadn’t really named before:
I do my best thinking when I’m in conversation.
When I have something to react to. To push against. To reshape. That’s when ideas sharpen, when they become something I can actually stand behind.
Before, that conversation had to come from another person. Now, it can start earlier. It can happen faster. It can be a little messier and lower stakes.
And sometimes, that’s all AI really is for me:
a conversation starter.
Not the answer. Not the source of truth. Just something to get the thinking going.
What I’m still protective of
There’s a kind of thinking that feels very human to me. The strange connections. The outliers. The moments where something feels interesting before you can explain why. The ideas shaped by lived experience, by taste, by curiosity—by the very specific way someone’s brain happens to work.
AI is trained on what already exists. That’s part of what makes it useful. It can synthesize, remix, and surface patterns across an enormous body of work.
But that also means it’s working from what’s already been done.
Maybe that changes over time. Maybe it gets closer to something that feels truly novel. I don’t know. But right now, there’s still something uniquely human in surprise. In delight. In creating something that didn’t quite exist before.
And for now, I think that’s worth holding onto.
Final thought
There’s a lot of pressure right now to have a strong opinion about AI. To take a stance. To be certain about what it means. I’m less interested in that. I’m more interested in staying curious. Using what’s useful. Questioning what isn’t. And making sure the work is still good.
Because at the end of the day:
AI isn’t the work.
It’s part of the work.
And everything else—we’ll figure out.