Weekly newsletter
Every Tuesday, a short field note on what was approved, who owns the decision, and what evidence survives after deployment.
June 30, 2026
Seven primary sources between December 2025 and June 2026 each named a different surface of the same governance gap. ISACA called it an authorization crisis. FINRA called it a traceability problem. The OCC excluded it from model risk guidance entirely. None of them named the gap itself. Edition 16 names it: the Chain Authorization Gap.
Read issue 16 →June 23, 2026
Three major AI governance announcements landed in the first week of June 2026: Executive Order 14409, Microsoft's Agent Control Specification, and Microsoft's updated AI Red Team taxonomy. All three address what agents do while running. None answers who authorized the agent's scope before it ran the first task.
Read issue 15 →June 16, 2026
An internal auditor asks who approved the ten thousand dollar auto-approval threshold inside an accounts payable agent, and no one in the room has the record. Microsoft's Agent Governance Toolkit, released April 2, 2026, can prove the rule was enforced. It cannot prove the rule was ever reviewed under the accountability model the AI agent was supposed to operate under.
Read issue 14 →