Immutable axioms defining the boundary between human authority and autonomous agent execution.
A hierarchical architecture from strategic governance to real-time execution supervision.
ACF classifies systems by autonomy level. Level 2 is the recommended deployment target.
Fixed rules, no ML. Human intervention for any modification.
Agents analyze and recommend. Every final decision remains with a human.
Agents decide within strict governance. Non-delegable zones locked.
Agents decide and learn. Maximum governance. For mature organizations only.
A sequential path deployed progressively over 6–18 months. Each module builds on the previous.
In 2026, autonomous agents are already executing decisions in commercial environments, setting prices, qualifying leads, processing claims, managing contracts. Most organizations have no framework to govern them.
ACF was created to fill that void, before a governance failure becomes a legal, financial, or reputational crisis.
Six complementary products operationalizing the ACF Standard.
The platform that turns EU AI Act obligations into a defensible, audit-ready trail: classification, regulator-ready documentation, and a court-grade audit trail for every agentic decision.
Free for teaching and non-profit use. A 130-page Manual, 30 printable cards across three families (Decide, Organise, Operate), calibrated cases by academic track and a human grading methodology. Designed for business schools, engineering schools, MBA / EMBA, AI master's, bar schools and continuing education.
Governance insights, framework updates, and agentic intelligence research.
56 questions across 8 categories: framework, agentic commerce, EU AI Act, ACF tools and more.
View the full FAQ →