FRAMEWORK / IN DEVELOPMENT
The Agentic RFX Framework
Procurement criteria for agentic systems: mandate boundaries, tool and memory governance, identity, and alignment with intent. In development — extending the eight AI-RFX criteria to systems that act.
01 — WHAT THE EXTENSION WILL ADDRESS
What the extension will address
The extension will request evidence for agent mandates and oversight under P01 Human Augmentation by Oversight, alignment with intent under P09 Human Alignment by Intent, and agentic security aligned to the ASI taxonomy under P08 Safety by Security. This closes the classic framework’s known ninth-principle gap without changing its established criteria.
The criteria are being drafted against the nine principles. They are not yet ratified.