When I see a designer using an AI tool like Claude Code to build their portfolio, it signals to me that they're comfortable with a skill companies are asking for. More importantly, it gives me a window into how they'd use it on the job.
Over the past few months I've reviewed a lot of portfolios built with AI. Some were impressive. Others weren't. The difference had nothing to do with whether AI was involved and everything to do with judgement, the decisions a designer made about what to build, and their ability to make it feel like theirs.
Here's what I've learned from looking at both, and what I'd tell any designer building their portfolio with AI right now.
1. Don't overdo it
AI slop is easy to spot, and when I see it, it tells me the candidate doesn't have the judgement to know how to use AI effectively.
I looked at one portfolio recently that had a music player embedded on the homepage, AI summary buttons attached to every case study, and generic UX jargon scattered throughout. None of it had anything to do with the actual work. It was distracting, and worse, it pulled focus away from the one thing I was there to learn, who this person is and what they can do. At a certain point I found myself questioning whether there was a real person behind the portfolio at all.
If a feature doesn't help me understand your work better, it doesn't belong there, no matter how easy AI makes it to add.
2. Show your personality
What's your story, and how does it come through in the way you present your work?
One of the biggest challenges I'm dealing with right now as someone who evaluates candidates constantly is figuring out whether the person behind a portfolio is real. AI makes it easy to produce something polished and generic in an afternoon, which means polish alone doesn't tell me anything anymore. The more your portfolio sounds and feels like you, the easier it is for me to trust that what I'm looking at is yours.
3. Obsess over the details
The bar is higher now, both in design and in implementation. Mistakes stand out more than they used to.
I've looked at a few portfolios recently with grammar errors, inconsistent spacing, and navigation that left me unsure where to click next. AI is good at getting you to a clean-looking starting point, but it won't catch everything, and it definitely won't catch how something feels to actually use. Don't rely only on AI to check your own work. Run it by a friend or two and get a human read on it before you submit it.
4. Don't rely solely on an existing design system
Show me your own creativity, not just your ability to assemble what someone else already built.
A lot of client case study work is built on top of an existing design system, which makes it hard to see a candidate's own creative instincts in that work alone. Your portfolio is where that changes. Where can you add a detail, a flourish, a small moment of delight that's entirely yours? That's often where I learn the most about what you're capable of when nobody's constraining you.
5. Make your case studies personal
Tell the story behind the decisions and who the key players were, not just the outcome.
It's difficult for me to tell what's AI generated and what's yours anymore. So make it easy for me. Look at your own case studies and ask yourself if they read as generic, the kind of write-up that could belong to anyone on any team. If they do, the fix is usually the same, tell the story behind the decisions. Talk about who you worked with and what made the project hard. That kind of detail is nearly impossible to fake, and it's what helps me see what you're capable of.
The work and the stories have to be yours
None of this is really about AI. It's about whether the work in front of me tells me something true about the person who made it.
AI can help you build faster, write cleaner code, and get a portfolio online in a fraction of the time it used to take. What it can't do is make the decisions that were yours to make, and it can't tell the story of how you got there. That part is still entirely on you, and it's the part I care about most.
So use AI. Just don't let it speak for you. The work has to be yours, the reasoning has to be yours, and when I ask you why you made a choice, the answer has to be yours too. That's what I'm looking for, and no tool can hand that to you.
Get the next one by email
One short email a month on hiring and design.