Notes from the work

Ideas worth
thinking through.

What I’m learning about products, design, and the systems that make them work.

The Operating Contract Behind a Trustworthy AI Workflow

The Operating Contract Behind a Trustworthy AI Workflow

A practical operating contract for AI workflows: clear jobs, visible states, approval boundaries, and evidence of completion.

· 6 min read
Eight Questions to Answer Before Automating a Business Process

Eight Questions to Answer Before Automating a Business Process

Eight questions about ownership, decisions, exceptions, and recovery to answer before choosing your automation tools.

· 4 min read
Human-in-the-Loop Is a Product State

Human-in-the-Loop Is a Product State

Why approval, rejection, editing, and recovery belong in the product state model of every human-supervised workflow.

· 4 min read
Agent Permissions Are UX

Agent Permissions Are UX

Identity, access, scope, approvals, and history: the product decisions that make AI agents understandable and trustworthy.

· 3 min read
Prototype ≠ Production System

Prototype ≠ Production System

A seven-point launch-readiness checklist for founders moving from a convincing prototype to a dependable production system.

· 6 min read
The Difference Between Designing Screens and Designing Systems

The Difference Between Designing Screens and Designing Systems

What changes when you stop designing individual screens and start designing the systems, responsibilities, and people behind a product.

· 3 min read

More of my writing on Medium ↗