Skip to content

AI + Design

AI does not replace judgement. It expands what I can explore, build and check.

AI sits inside my design process rather than beside it — widening the option space early, compressing the distance from concept to working prototype, and catching accessibility and consistency issues while they are still cheap to fix.

The six-stage flow

  1. 01

    Explore

    AI widens the option space

    • ChatGPT
    • Perplexity
  2. 02

    Structure

    AI as a thinking partner on IA

    • FigJam
    • Miro
    • ChatGPT
  3. 03

    Generate

    AI accelerates making

    • Figma Make
    • Midjourney
    • Adobe Firefly
  4. 04

    Prototype

    AI-assisted build

    • Cursor
    • Figma Make
  5. 05

    Evaluate

    AI catches what review misses

    • WAVE
    • Axe
    • ChatGPT
  6. 06

    Decide

    The designer decides

    • Designer judgment (no AI tool — intentional)

AI Skills

  • AI-Assisted Product Design
  • AI-Assisted Discovery
  • Research Synthesis
  • Prompt Engineering
  • AI-Assisted Prototyping
  • Custom AI Agents
  • Human-AI Collaboration
  • AI Design Operations

AI Tools

  • ChatGPT
  • Claude
  • Microsoft Copilot
  • Cursor
  • Figma Make
  • FigJam AI
  • Google Gemini
  • Google Stitch

Exploring

  • MCP
  • API-Based Design Workflows
  • AI Workflow Automation

The principle

AI accelerates options. Design judgment decides what ships. The designer stays responsible for every decision.

  1. Step 01

    Explore

    AI widens the option space

    Generate alternative flows, structures and framings quickly, then judge them against product constraints.

    • ChatGPT
    • Perplexity
  2. Step 02

    Structure

    AI as a thinking partner on IA

    Pressure-test information architecture, naming and grouping before committing to an interface.

    • FigJam
    • Miro
    • ChatGPT
  3. Step 03

    Generate

    AI accelerates making

    Move from structure to screens without spending the day on scaffolding.

    • Figma Make
    • Midjourney
    • Adobe Firefly
  4. Step 04

    Prototype

    AI-assisted build

    Turn concepts into interactive experiences that can be used, not just reviewed.

    • Cursor
    • Figma Make
  5. Step 05

    Evaluate

    AI catches what review misses

    A custom accessibility agent in Cursor flags accessibility issues and inconsistencies in the emerging build.

    • WAVE
    • Axe
    • ChatGPT
  6. Step 06

    Decide

    The designer decides

    Strategic, ethical, contextual and product decisions stay human. AI expands exploration; it does not replace judgement.

    • Designer judgment (no AI tool — intentional)

In practice

What this looked like on a real project

LabSync AI

On LabSync AI, AI was used to generate 8 layout variants for the role-based dashboard in under 20 minutes. The final design combined elements from 3 variants — filtered through accessibility requirements and clinical workflow logic.

AI agents

Purpose-built agents, not a chat window

I have explored building specialized agents that hold a specific point of view in the design process — a practical way to reduce bias and keep quality checks continuous.

User-focused agent

Argues from the user's position during exploration, keeping usability questions present in early decisions.

Business-owner agent

Represents commercial and operational constraints to reduce bias toward a single perspective.

Accessibility agent in Cursor

Reviews implementation for accessibility issues and flags them while the work is still cheap to change.

The line I hold

AI generates options. Product context, user needs, accessibility, ethics and trade-offs decide which option ships — and that decision is design work.

Contact

Let’s talk about the problem, not just the solution.

I’m open to senior product design, UX leadership, design systems, and AI-native product opportunities.

Get in touch