vcard
VCARD
The simpler way to pay: virtual cards for everyone
VCard lets users create instant virtual cards—both single-use and recurring — that help them manage their finances with more control and privacy. It’s built for international transfers, holding money in different currencies, and making online shopping safer. Plus, it gives users an easy way to manage subscriptions and keep their main bank account details to themselves.
Project Goal
Create an MVP for the virtual card feature that prioritizes a clean user experience while leaving plenty of room to expand functionality over time.
Key Features
At the MVP stage, VCard lets users create up to five virtual cards in euros or dollars, each with a unique 16-digit number, CVV, and expiration date linked to the main balance. Cards can be funded via bank transfer, debit card, or cryptocurrency. Security features include instant card freezing. Virtual cards work for online purchases, and users can order a physical card for in-store use and ATM withdrawals.
CLIENT
AltDev
SKILLS
UX Research
UI Design & Design System Management
Interactive Prototyping
Integrating AI Into Real Design Work
The 4-Day MVP: My AI-First Design Sprint
It was the kind of request that usually guarantees missed deadlines and sleepless nights: five working days to deliver a clickable prototype, a full UI kit, and developer specs—with initial sketches due to the client in just two days. To make it more interesting, my own team was still debating the core functionality of the MVP. With the clock ticking, I knew a traditional linear process wouldn't cut it. Instead of seeing AI as a futuristic novelty, I treated it as my core design partner, tasking it with everything from research to prototyping to documentation. The goal wasn't just to finish on time, but to use the leverage of AI to deliver a higher fidelity, more thoroughly explored solution than the timeline should have allowed.
AI as Research Synthesizer and Ideation Partner
The first hurdle was breaking the team's deadlock on MVP functionality and simultaneously gathering inspiration. I started by interviewing our product manager and developers to capture their vision, then fed that context into an AI workflow.
I used Gemini to analyze the most successful virtual card apps I'd sourced from Gemini and Mobbin, prompting it to synthesize their core interaction patterns. Where a human researcher might spend hours categorizing, Gemini provided a structured breakdown of UX patterns in minutes. I then used that analysis as a springboard, asking Gemini to generate several distinct implementation options that mapped those patterns to our specific team goals. This didn't give us a final answer, but it gave us concrete, well-researched options to debate, instantly moving the conversation from abstract opinion to tangible preference.
Visual Exploration and Rapid Prototyping at Scale
With the team coalescing around a direction, I moved into high-speed execution. Instead of manually building every screen, I used Figma Make to generate several clickable prototype versions based on the approved user flows, allowing the team to test and validate the interaction design almost immediately.
Simultaneously, I tackled the look and feel. While the team reviewed the flow, I used ComfyUI to generate multiple high-fidelity visual concepts. I wasn't asking AI to replace my aesthetic judgment; I was using it to rapidly iterate on color, texture, and ambiance, creating a range of options I could curate and refine. This parallel process meant that by the time the client saw the final prototype, we could also present a curated set of visual directions, making the feedback cycle incredibly efficient.
Closing the Loop: From Client Feedback to
Developer Handoff
The client's revisions arrived, and this is where the AI-powered workflow truly paid off.
The product manager's list of changes became direct prompts in Figma Make, which intelligently updated the prototype’s structure and flow. This transformed what is usually a tedious manual adjustment phase into a rapid, semi-automated refinement loop. The updated, detailed prototype was then easily tweaked in Figma to lock in the client’s chosen visual concept.
For the final handoff, I used Gemini and DeepSeek to co-write a series of highly specific prompts for Figma Make. The result was a comprehensive, fully documented UI Kit and a complete set of developer specifications, generated in a fraction of the time it would have taken to assemble manually.
The Designer's Leverage: A Retrospective
The project was delivered to developers in just four days. This case study isn't about AI working miracles in isolation; it’s a story of strategic augmentation.
AI acted as my research synthesizer, ideation partner, and documentation assistant. It didn't replace my design thinking—I provided the context, strategy, curation, and final polish—but it did remove the bottlenecks. By handling the heavy lifting of exploration and variation, AI freed me to focus on the critical 20%: the nuanced decision-making, user behavioral patterns, and emotional brand resonance. The result was a compelling, well-documented solution that the client loved and the developers could build from immediately.
As AI continues weaving itself into traditional design workflows, the boundaries between technology and creativity will only grow more fluid. Yet this fusion isn't about replacement—it's about leverage. The designers who flourish won't be the ones who resist AI or or those who rely on it completely. They’re the ones who develop AI literacy: They’re the ones who develop AI literacy: understanding what these tools do well, where they fall short, and how to orchestrate human creativity with machine efficiency.
