Field notes · AI systems in production

Notes from building AI
inside real businesses

Things I've learned making AI actually useful in day-to-day operations — context engineering, agent tooling, and the parts nobody warns you about. Each note is interactive: you can manipulate the idea, not just read about it.

Note 01 · Context Engineering

The Context Layer

Why business AI gives generic answers, why feeding it more data makes it confidently wrong instead, and the curation discipline that fixes it. Two interactive demos.

Note 02 · Operating Model

Why Your Company Wiki Died

Documenting loses to the work every time, because the person with the knowledge pays and someone else benefits. What changes when the deposit stops being a separate chore. Two interactive demos.

Note 03 · In progress

Anatomy of a Spec That Steers an Agent

What separates a specification an agent executes correctly from one it wanders away from.

Note 04 · Planned

Connecting Agents to Real Systems

Turning an assistant that talks about your business into one that can look things up and act.

Note 05 · Planned

Designing Skills That Don't Break

Packaging repetitive work so it runs the same way every time — and the iteration loop that gets it there.


Neal Meinke

I build AI systems inside businesses — context layers, system integrations, and internal tools that replace manual work. Currently a senior lead application developer working on enterprise AI assistants; these notes are the methods, generalized.

nealm682@gmail.com · LinkedIn · GitHub