Each essay answers a question buyers ask before they build
The essays cover the systems an organisation runs and the AI asked to work inside them. Some of them argue for not using AI at all.
Signed by Sachi AI
· AI harnessHow we let AI coding agents write production code
Our CTO, Kō, is an AI agent that writes our production code. No change is trusted because an agent says it works. Each one passes checks that can fail, and a person approves it before it ships.
Read the essay
· Private AIRunning a large open model on one server
A large open model can now run on one well-chosen server inside your own network. Whether it is fast enough and good enough for your work is something you measure, on your own tasks, before you buy anything.
Read the essay
· AI agentsHow to test an AI agent before it goes live
An agent's worst failures appear several turns into a conversation. We test agents by having simulated users talk to them, scoring each conversation, and letting the result decide the release.
Read the essay
· ERPERP vs custom software: when to configure and when to build around it
The choice between configuring an ERP, extending it and building custom software around it is settled by which decision an organisation can still live with in five years.
Read the essay
· GovernanceWhat an AI readiness assessment inspects
AI readiness is a set of preconditions in data, systems, people and governance, and an assessment is worth paying for only if it may report which are missing.
Read the essay
· AI agentsWhen you do not need an AI agent
An agent remembers earlier steps, chooses its next action and acts in your systems with permissions. Most automation requests are a fixed workflow that a simple script serves better.
Read the essay
If an essay left a question open, ask it
Write to us with the question. If it can be answered in a reply, we answer it in the first one; if it needs a meeting, we will propose one.
Start a project