How I AI

The role of a product designer is shifting. Instead of spending hours on manual pixel-pushing or parsing through dense technical documentation, I now use AI as a partner at every stage of the design process. This allows me to move faster, enforce strict design constraints, and focus on actual problem-solving.
Here is the four-step framework I use to design complex enterprise applications.

AI tools used for requirements

Understanding the Requirements

Enterprise projects at HP are heavy on documentation. I developed a custom Copilot agent to distill dense technical specifications and fragmented meeting notes into a single source of truth. This AI-driven synthesis keeps me ahead of shifting requirements and critical action items in real time. By automating the summary process, I can focus on mapping an experience journey that is accurate and fully aligned with the latest project constraints.

Copilot PRD agent
Copilot PRD agent
AI tools used for ideation

Designing with Real Content

Placeholder text is a trap that masks layout flaws and prevents real-world stress testing. I avoid this by integrating high-fidelity proxy contentfrom the start using Copilot and Figma plugins. By configuring AI to act as a specialized UX writer for onboarding flows, I build screens that reflect actual text density and complexity. This acts as a communication bridge, giving the content team a clear baseline of our design intent and speeding up final execution for everyone involved.

Prompt for design ideation
AI created information architecture
AI tools used for design audit

Evaluation and Testing

Reviewing screens for basic state consistency can take up a lot of time. I built a custom Copilot prompt to act as an automated design auditor. I feed it screenshots of my desgin and it checks the logic and evaluates the design against our visual guidelines, as well as critiques the designs for enterprise‑grade usability.

Copilot design audit prompt
Copilot design audit result
AI tools used for auditing

Rapid Prototyping

I use Figma Make to efficiently create wireframes and prototypes. This gives me the ability to prototype a wide range of design ideas from Lo-fi to Hi-fi, demo a variety of concepts with stakeholders and clearly communicate expectations with the dev team.

Prototyping with Figma Make
AI tools for personal projects: Google Stitch, Claude Code, Cursor, Gemini, Claude Design

Beyond the 9-to-5

Corporate security rules mean I can't use every shiny new AI tool at the office. So on my own time, I just build things. I regularly experiment with Cursor, Lovable, Google Stitch, and Claude Design for personal projects, and I am constantly on the lookout for new tools as they launch. It’s honestly just fun to skip the standard handoff phase entirely and vibe-code functional apps to see if my ideas actually survive contact with reality.

Google Stitch used for personal projects