“So I built an A.I.-powered portfolio…” I've been saying that a lot lately. The truth is, I have no idea if it'll actually work the way I want it to. With the tech industry shifting rapidly and A.I. reshaping everything from workflows to entire job roles, A.I. has quickly become less “future” and more “now”—I knew I had to dive in.
A.I. is no longer just a buzzword floating around Silicon Valley—it's in everyday conversations. Even my dad, who's more likely to be at home under the hood of a car than behind a screen, started referencing ChatGPT. That's when I realized: instead of diving into a brand-new project in healthcare or finance—industries I've always found fascinating—maybe the smartest thing to do was redesign something much closer to home: my portfolio which admittedly had been quietly collecting dust.
Portfolios are like sandcastles—meticulously crafted, beautiful for a moment, and then slowly washed away by time and trends. Most designers never feel like they're done. Give it a year or two, and we're already dreaming of starting from scratch.
As I looked at the portfolios that inspired me (and probably inspired hundreds of others), I noticed a pattern. Sure, the content and style change, but the experience? Pretty identical.
Image: [Photo of designer looking more professional than they really are]
H1: “Hey, I am [insert designer name here]”
H2: “I'm a [Job Title] with a passion for design, research, and writing user acceptance criteria…”
New Section: This is my work.
Multiply that by 200+ portfolios, and yeah—it's a blur.
Thinking about my portfolio I started to think about what its purpose is and what I want it to say about me. My job as a Product Designer is to create intuitive AND innovative experiences. I still needed to highlight my work, but maybe I could rethink the vessel that delivers it.
I wanted users landing on brianprouty.com to say “wait whaaaaat?!”. My goal was to break away from the cookie-cutter template while still making it easy to explore my projects. I decided to put the AI front and center—trained on my work and ready to answer questions. It's a risk, I know. Attention is short, and unfamiliar interfaces don't always help. But I also believe that if the experience is clear, intuitive, and just different enough, it might just spark curiosity—and that's all I need to get someone to take the next step.
I split my audience into four types of visitors:
To better understand how each group might interact with my site, I mapped out a user journey for each:
With the user journeys in place, a solid design system carried over from my previous site, and a clear vision, I set out to create something new.
Initially, I defaulted to what I always do: fire up Figma and start designing a prototype. But this time, I found myself wondering what kind of application I wanted to build—something more interactive, more dynamic. That curiosity led me down the rabbit hole of OpenAI's APIs. The more I explored, the more I started to believe I could actually pull this off.
I've dabbled in front-end development for a few years, but I still consider myself a novice. I've leaned heavily on Stack Overflow to get through past projects, but everything I'd been hearing about AI tools suggested they could help me get unstuck faster. So I decided to give it a go.
And like any aspiring developer in 2025, I turned to ChatGPT...
It all started with one prompt:
“Assume the role of an experienced web developer well-versed in implementing OpenAI APIs for the rest of this conversation.”
From there, ChatGPT walked me through everything. I explained my idea:
“I am a product designer and a beginner developer that needs help creating a new AI driven website page. The idea behind this is you would be able to arrive on the page and ask questions about the 'Brian' and get responses. This would only reference information about Brian such as professional experience, hobbies, skills, values, things Brian has worked on in the past and nothing else. An example question a user would ask: 'What is Brian's professional experience'. Can you walk me through, in great detail, how I would do this?”
And it did. In detail. I mean, I'm not saying it handed me a finished app, but as someone who still considers himself a novice, I got about 70% of the way there with barely a hiccup.
But then my luck, or more accurately, my skill ran out.
I was having an issue where the API couldn't find the JSON file. I was in too deep in chatGPT land, buried in folders I'd created via terminal commands I barely understood. I could not figure out why, no matter where I moved the data, the program couldn't find it. After a week of frustration, I was ready to scrap the whole thing.
Taco Tuesday saved my side-quest project.
Over tacos, I told my friend Sam—an actual computer science grad and far more legit developer—about my issue. He asked, “Have you tried using Cursor?”
I blinked. “What's Cursor?”
TL;DR: Cursor is an AI code editor that understands your entire codebase and helps you debug and build in real time. Within minutes, it solved the JSON issue. Then it helped me add things like streaming, animated text, and mobile responsiveness. It was like having a senior dev sitting over my shoulder—minus the judgment.
And just like that, the project was back on track.
Once the build was far enough along, I knew I needed to get some outside eyes on it before throwing it into the ether. So I started cold-messaging recruiters and design leaders to get some honest feedback.
Huge shoutout to those who took the time—especially considering how weird it probably felt to get a cold message from someone asking, “Hey, wanna try my AI-powered portfolio?”
Some of the responses were positive, but a lot pointed to areas where I needed to step it up. The biggest issues:
A few quotes that stuck with me:
Tough love—but 100% necessary.
After gathering all that feedback, I rolled up my sleeves and made some updates:
The end result? A portfolio powered by AI that can answer questions about me, using only the info I fed it.
Give it a spin—I'd love to hear what you think (or what you ask it)!