How agent governance works under the hood.
Notes on mechanisms and operations, written by the people who built them. Each one explains an engineering decision and ends by saying what it doesn't solve.
What goes into an audit trail
Eight fields, and what happens when the most expensive one is missing at audit time.
Sep 14, 2026 6 min
Where policy is evaluated, and why it happens first
The difference between observing and governing is when the rule runs, and what that costs.
Sep 14, 2026 6 min
Governing an agent you didn't build
Registering isn't migrating. What changes, what doesn't, and the technical condition for it to actually work.
Sep 14, 2026 7 min
The math of a parallel runtime
When the runtime's foundation stops being a preference and becomes an infrastructure cost, month after month, on the invoice.
Sep 14, 2026 7 min
The agent nobody remembers deploying
Every organization has at least one. It becomes a problem in one of four ways, and you can see it coming.
Sep 14, 2026 6 min
Why we call it an operating system
The word isn't a metaphor. And the analogy breaks at one specific point worth saying out loud.
Sep 14, 2026 7 min
AI agent governance: a guide for banks
Five questions governance has to answer about any agent, and where to start when none of them has a quick answer today.
Oct 6, 2026 7 min
Human in the loop for AI agents: when to stop the run
Three questions for deciding where an agent waits for a person, and the signal that approval has stopped protecting anything.
Oct 13, 2026 6 min
Why AI agent pilots stall before production
Four questions the committee asks that the pilot never answered, and why every new agent starts from zero when the answers live inside each project.
Oct 20, 2026 7 min
