How "Vibe Coding" with AI Supercharged My Creativity as a Designer

A few weekends ago, I signed up for a “Code with AI” workshop in Singapore, thinking it would be a quiet class filled with engineers.

To my surprise, the room was buzzing, filled with Product Managers, Engineers, Consultants, Entrepreneurs, Educators, and people from every walk of life. The energy was electric. Everyone was there for the same reason: to learn how to create with AI, even if they weren’t professional coders.

During the class, we were introduced to Vercel 0, Cursor, and Claude.ai as our core tools although along the way, a few other AI tools were also mentioned when the audience asked. The two founders who led the workshop were incredibly experienced, knowledgeable, and passionate, making the whole learning journey approachable even for beginners.

This article is a reflection of how that experience reshaped my thinking as a product designer, and why I believe AI isn't something to fear, it's something to embrace to unlock new creative possibilities.

The AI Stack for making our projects come alive in 2 days.

How It Started: PRD Written by AI

For my project, I created UXniverse, a flashcard web app that helps users test their knowledge of Web3 UX patterns in a fun, bite-sized way.

So I opened up Claude.ai and asked:

“I want to build a Web3 UX flashcard platform. The goal is to help people (and myself) learn Web3 design principles in a fun, visual way. Can you just give me features and app description? Before we write the PRD, can you ask me 5 clarifying questions?”

Within minutes, Claude organised a lightweight PRD outlining goals, features, and user journeys. It wasn't perfect, but it gave me momentum and that made all the difference.

Writing Prompts: Learning to Converse with AI

One thing I learned quickly is that how you talk to AI really matters. Good prompt engineering is whereby things like being clear, using the right language, and giving enough context makes a huge difference in the kind of output you get.

If you're vague or too open-ended, the AI can get confused and take you in random directions. But when you're specific about what you want, how you want it to sound, and who it's for, the AI becomes a way better creative partner.

It’s honestly like learning how to have a good conversation, the better you express yourself, the better the collaboration gets.

Some key lessons I picked up:

Instead of saying "build a flashcard app," I'd prompt:

"Create a simple flashcard web app with a start screen, flashcard flipping animation, and progress tracking."

Instead of "fix this," I'd ask:

"Explain this error to me, I'm new to coding and show an example of the correct version."

The instructor demonstrated an example of his project, asking the agent to explain what is going on with his codebase without writing any new code.

The better I communicated, the better AI collaborated. It was less about coding and more about "creative directing".

Rapid Creativity: Prototyping on the Fly

Using Vercel 0 (v0), I generated UI screens in less than 10 minutes. The designs weren’t final, but they were fuel for more ideas.

This was the first time design felt like: Idea → Prototype → New Idea → Refine (in real time).

From prompt to product: code and designs in under 10 minutes.

After generating the first draft, I exported the codebase to Cursor and continued refining there. In Cursor, I learned to run simple developer commands like:

When I got stuck?
I asked the AI agent inside Cursor to troubleshoot with me.

Instead of fearing mistakes and getting sooooo done with the errors popping up from time to time (I think I can finally understand a developer's pain and patience), I started learning through momentum.

How AI Changed My Creative Thinking

This whole experience taught me something powerful. AI doesn’t replace creativity. It amplifies it.

I could focus energy on exploring ideas, not technical bottlenecks and felt freer to experiment without worrying about "getting it wrong."

AI turned "I can't code" into "I can create." However, there are some pros and cons of using AI in design. Do not expect perfection in your design immediately.

Pros and Cons of Using AI in Design

Pros:

Cons:

The screenshots below highlight some of the limitations of using AI in design, the output feels generic and overly “safe,” and the UI definitely needs further refinement.

That said, my current goal is to gather feedback on the interaction flow, user experience, and whether the design accurately reflects the intended user journey. So this is a solid starting point!

Home page of UXniverse. A design concept purely generated from the first try. It covers the basic requirements and functionality.

Handling error scenarios is crucial. In this case, I had to craft a few prompts to trigger the right logic, aiming for the most concise and elegant solution that required minimal code changes throughout the project.

Flashcard of question and multiple choice answers. Images in flashcard are created by me.

While AI made it fast to prototype everything, there were still some hits and misses. After completing the ‘Easy Level’ quiz, the flow abruptly ended. There was no closure.

As a user, I still want guidance even after a task is complete. So I prompted the AI agent to include a simple congratulatory message at the end to wrap up the experience more thoughtfully.

Choose to continue the next level or return Home.

I wanted to make the platform more engaging by adding a chatbot feature. But as you can see, I ran into some API connection errors. I’m still working through the setup with the AI agent. Might be time to phone a friend from engineering!

My chatbot uses the Gemini API. Supabase was mentioned during the workshop and in my case, I’d consider using it if I want to store or analyse user messages for learning, personalisation, or insights.

A collection of best practices that designers can consider when building web3 user experiences & interfaces. Content adapted from web3designprinciples.com

Closing Thoughts

This journey made me think about the future of product design:

• How do we design intuitive experiences that partner humans and AI?

• How can we lower the barrier for non-technical creators?

• How do we design ethically, ensuring AI amplifies human creativity rather than replacing it?

As a designer, I want to learn more about:

• Prompt engineering as a form of UX

• AI agent interaction design

• Foundations of ethical AI product design

I know many designers talk about that mix of excitement and anxiety about AI. We all feel it, and honestly, I’m no different. But at the end of the day, I always remind myself: Use it to expand your creativity, not replace your thinking. That to me is the magic of "vibe coding".

It was such a fulfilling session, not just because of what I learned, but also because of the incredible people I met. The energy in the room was contagious, with everyone bringing different perspectives and backgrounds. If you're curious about trying vibe coding or experimenting with AI tools in your own creative process, I’d genuinely love to hear your thoughts!

An amazing bunch of curious minds, all exploring what’s possible with AI and creativity! (Special thanks to Agrim & Sherry for facilitating this workshop.)