Blog
AI agents, under your rules.
Deep dives, field notes and guides on putting AI agents into regulated work: where the rules live, where controls belong, and how to prove every decision.
All posts
Getting AI agents into production.
Why agent projects stall after the pilot, what a wrong answer costs, how the alternatives compare, where controls belong, and what we heard from banks.
Agents don’t fail on intelligence. They fail on rules nobody sealed.
Why most agent projects stall after the pilot, and the three steps that get them into production.
For AI leaders, CEOsWhat does an AI agent mistake cost? A pricing model for the wrong answer
Teams budget for tokens and forget to budget for error. A way to price it.
For CFO, COO, riskSix ways to keep an AI agent on policy, compared honestly
Prompts, guardrails, evals, observability, hard-coding and a sealed rulebook: where each one wins.
For AI leaders, architectsWhat 19 conversations, mostly with banks, taught us about agents in regulated work
Five patterns from the field, and what we changed because of them.
For Bank COO, operationsA dashboard is not a gate: where controls belong when AI agents act
Monitoring tells you what an AI agent did. A gate decides whether it may act. How to sort your agent decisions and where each control belongs.
For Process owners, risk, AI leadersTen questions your examiner will ask about your agents
Which rule applied, who approved it, could anyone stop it first. Ten questions, and what a good answer contains.
For Risk, compliance, bank COO
Bring one process. See it Factified.
A working session with our team on one of your real processes: where its rules live today, where they conflict, and what it looks like sealed.