AI Work

AI as a layer in product delivery systems.

I use AI as part of an evolving product design practice — supporting research, strategy, prototyping, specification, and implementation. The focus is on designing workflows where AI becomes a structured, reliable component of how products are built, rather than an isolated tool.

Practice

Active · Ongoing system development

An AI-enabled design practice.

AI is embedded into my product design workflow — not as a separate discipline, but as part of how I structure and deliver product work. My practice spans research synthesis, design exploration, localisation, workflow automation, and early-stage product definition.

Increasingly, this is supported by structured AI-assisted systems built around reusable context, documentation, and iterative feedback loops. The shift has been from prompt-level experimentation to system-level design — where value comes from how tools, context, and decision-making are connected.

Projects

Three pillars

  1. End-to-End Thinking

    Product design remains a full lifecycle discipline. AI changes how stages connect, not whether they exist.

  2. Systems Over Prompts

    Value comes from structured workflows, reusable context, and feedback loops — not isolated interactions.

  3. Enhancement, Not Replacement

    AI supports synthesis, exploration, and iteration, but does not replace product thinking or judgment.

Continue reading

Projects

  • Coming Soon

    Hobby Paint Inventory App

    An AI-supported product system for managing paint inventories, cross-brand colour matching, and recommendation logic. The underlying product behind the PaintRack case study.

    Role

    Solo product design and development.

    Stack

    • Claude Code
    • React Native
    • SQLite
    • Anthropic API
    Coming soon
  • Live

    This portfolio site

    The Next.js production site you're reading now — designed and built as a structured AI-assisted product. Case-study MDX pipeline, embedded portfolio assistant, brand-led design system. AI works as a structural layer in the build, not the centrepiece.

    Role

    Solo product design, copy, and engineering direction. Wrote the brief (CLAUDE.md), made every product and visual call. Claude Code as the engineering pair across every commit.

    Stack

    • Claude Code
    • Anthropic API
    • Next.js 16
    • Tailwind CSS v4
    • MDX
    Currently in use
  • In Development

    Personal Financial Dashboard

    An AI-assisted financial visibility tool that consolidates policies and investments into a unified view. Uses AI to extract structured insights from policy documents and present simplified financial summaries.

    Role

    End-to-end product design and build.

    Stack

    • Claude Code
    • React Native
    • Expo
    • TypeScript
    In development
  • In Development

    Specialized Rider Rewards

    A loyalty and rewards system for cyclists in South Africa. Focus on structured reward mechanics and service design across physical and digital engagement loops.

    Role

    Product designer — end-to-end contribution from discovery through to high-fidelity design system.

    Stack

    • Figma
    • User research
    • Loyalty mechanics
    • Service design
    In development

Deep dive

How AI shows up inside my design process.

End-to-End Thinking

Product design remains a full lifecycle discipline. Research, strategy, information architecture, prototyping, specification, and implementation all still matter — arguably more than ever.

AI changes how those stages connect, not whether they exist. The work is integrating AI as a structured layer across the lifecycle so that quality and clarity carry through, rather than treating it as a shortcut around the thinking.

Systems Over Prompts

Value comes from structured workflows, reusable context, and feedback loops — not from isolated clever prompts. The quality of any AI output is bounded by the quality of the system surrounding it.

I focus on building workflows where context compounds over time: documentation patterns, knowledge structures, connected tooling, and review loops. As the system becomes more refined, the AI inside it becomes more useful.

Enhancement, Not Replacement

AI supports synthesis, exploration, and iteration. It does not replace product thinking or judgment.

The most useful applications are the ones that reduce repetitive overhead and create more space for deeper decisions — better synthesis of research, structured documentation, clearer specifications. Speed is a side-effect; the aim is more thoughtful products built with less friction.