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.
Enterprise projects usually start with massive PRDs. I use Notebook to quickly synthesize these documents. For example, during a recent printer cloud onboarding project, the AI extracted key insights about Zero Trust networks and dual onboarding straight from the documentation. This gives me a highly accurate requirement summary and helps map the experience journey.
I rely on a mix of AI tools to generate initial concepts and build the structural foundations. I use Copilot to assist with generating initial design ideas and concepts, and FigJam with Nano Banana as an AI-powered tool for brainstorming and ideation. By using highly structured prompts that define the exact product context and user goals, I can quickly generate viable information architecture and low-fidelity layouts to iterate on, rather than generic UI.
Placeholder text is a trap that hides broken layouts. I integrate real copy from day one using plugins like Frontitude and Copilot. I set up custom instructions for the AI to act as a specialized UX writer, specifically focusing on error recovery. It takes dense, technical system faults and translates them into plain, actionable human instructions. Designing with real words clarifies the user flow immediately.
Reviewing screens for basic state consistency takes up too much 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.
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, communicate expectations with the dev team and work directly from the design code level through to deployment.
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.
