warrenbuilds.ai

A personal space for AI-built experiments, prototypes, and things in progress.

> always iterating
experiments
Agentic Commerce Readiness Audit
A framework that scores how ready a merchant is for AI-driven buyers, then turns the gaps into a prioritised plan. Built to make agentic commerce readiness measurable instead of theoretical.
> for e-commerce merchants preparing for AI buyers
Shopify Digital Shelf
When you ask an AI to shop for you, who decides what you see? A demo built on Shopify's live Global Catalog over MCP: real retrieval, a look at the selection, and the brands competing for a slot they can't see.
> for merchants wondering how AI agents surface their products
Who shops for you?
One household job seen across three horizons: today's fragile browser automation, the merchant's own agent, then the platform's agent. An interactive anatomy of each, and what the merchant keeps — or loses — at every step.
> for merchants weighing what an agent keeps and takes
Hypothesis led agentic AI blueprint
From a rough hypothesis to a structured agent blueprint. AI drafts the first pass, you critique it, and a stress tester hunts the gaps before anything gets committed.
Most agent ideas fail because they're under-specified and over-trusted. This forces a hypothesis through critique and stress-testing first, so the spec is de-risked before a line of code is written.
> for anyone scoping an agent before building it> Build in progress
Chip, the Woodworking AI assistant
A woodworking assistant that produces shop drawings and cut lists, not just generic advice. You describe the piece, it returns something you can take to the bench.
Generic AI gives vague advice you can't act on. Woodworkers need buildable specifics — Chip returns shop drawings and cut lists you can take straight to the bench.
> Personal hobby> Build in progress
enablement
Build playbook
A living repository of the operational patterns behind shipping these projects, from scaffolding to deploy. Written for enablement: turning the working notes I reach for into something others can pick up and build from.
> for learners and builders with Claude Code
Skill library
A browsable explainer for a working set of Claude skills: what each one does, when it fires, and how the pieces fit together. Built to make a folder of skills legible to someone who didn't write them.
> for builders assembling their own skills for Claude Code
Parallel design exploration
Fan out from one brief into many genuinely different design concepts, score them against that brief, then refine the survivors. Built to start design from the problem rather than the pixels.
> for designers and product owners
about

Warren Everard. Leading Product Experience Design, AI, and agentic commerce consulting at Visa across Australia and New Zealand.

Fifteen years across financial services, government, and telco. Building teams, embedding human-centred design in matrixed organisations, translating customer research into experience strategy.

Increasingly interested in how service design principles extend to agent orchestration: designing systems where humans and intelligent agents collaborate to deliver better customer outcomes. This site is where that thinking gets prototyped on personal time.

I'm also enrolled in MIT's Applied AI and Data Science Program, deepening the technical understanding that sharpens how I think about these problems strategically.

disclaimer

Personal projects only. No employer affiliation, endorsement, or proprietary content.