Notes

Field Notes

I write about AI, business, and product management — the stuff I actually work with day to day. Short, useful reads on things I think matter.

Subscribe via RSS

Sort

DOC-012BusinessOct 3, 2026 · 1 min read

Think in Systems. Transform with Purpose

Many digital transformations stall before a single tool is bought, and not because of the technology. Six steps for thinking in systems and bringing people along, plus what a transformation really needs.

Read entry

DOC-011AISep 5, 2026 · 11 min read

Measuring an AI System, Stage by Stage

There is no single score for an AI system. What you measure changes four times between choosing a model and running one in production — what an eval actually is, which metrics earn their place at each stage, and why the governance ones belong on the same schedule.

Read entry

DOC-010BusinessSep 5, 2026 · 9 min read

From an Open-Ended Problem to a Shipped Solution

Building an agent is the easy half. The hard half is choosing what to build and closing a vague business problem into a bounded one — with TOSCAR — before a single prompt gets written. Then the four stages that follow.

Read entry

DOC-009AISep 5, 2026 · 5 min read

RAG Isn't Training — It's What You Put In Front of the Model

Retrieval-augmented generation changes nothing about the model — it changes what the model is looking at when it answers. The mechanism, worked end to end on one real question.

Read entry

DOC-008AISep 3, 2026 · 4 min read

A Framework for Building Agents That Don't Fall Over

Six parts every working agent has, whether or not the team building it named them. A practical structure for scoping, building, and containing an agent — before the demo becomes a product.

Read entry

DOC-007AISep 1, 2026 · 4 min read

MCP: The Boring Standard That Makes AI Tool Use Maintainable

Every AI product used to write its own tool-calling glue. Here is what the Model Context Protocol actually is, what a server exposes, and one small server walked through end to end.

Read entry

DOC-006AIAug 31, 2026 · 3 min read

Workflows, Agentic AI, and Autonomous Agents — and Why RAG Exists

Three terms get used almost interchangeably in AI conversations, and they mean genuinely different things. Here's the actual distinction, plus why retrieval-augmented generation matters underneath all of them.

Read entry

DOC-005BusinessAug 31, 2026 · 1 min read

A Simple n8n Workflow: Form Submission, Logged and Actioned

Not every automation needs a model. A small, real workflow — three fixed steps in n8n — that turns a form submission into a logged record and a sent email, automatically.

Read entry

DOC-004BusinessAug 10, 2026 · 2 min read

What Business and AI Together Look Like: The Value vs. Effort Matrix I Actually Use

Most 'AI and business' conversations skip the part where you decide what to build first. Here's a simple value-vs-effort framework for sorting real initiatives from expensive demos.

Read entry

DOC-003AIJul 15, 2026 · 3 min read

How Large Language Models Actually Work, for Non-ML People

A mental model of what's happening inside an LLM, written for someone who needs a clear picture without an ML background — not a machine learning course.

Read entry

DOC-002AIJun 28, 2026 · 2 min read

Prompting Techniques That Actually Move the Needle

Not a list of magic words. A short, practical set of prompting patterns that consistently produce better output, and why each one works the way it does.

Read entry

DOC-001AIJun 5, 2026 · 4 min read

A Short History of AI, for People Who Skipped the Hype Cycles

What the word 'AI' actually contains — machine learning, deep learning, generative AI and LLMs are nested, not synonyms — and the short version of how the field moved from the outer ring to the inner one.

Read entry