Microsoft’s Humanist AI Code of Conduct addresses an increasingly urgent question: what constitution should govern an autonomous AI actor? Its hierarchy of authority, bounded delegation, environmental constraints, human-legible conduct, and commitment to human control give the Code a genuinely constitutional character. But the Code also brings a second question into view: what constitution governs the institution in which that actor is allowed to act?

​

A model may comply fully with its constitution while operating inside an enterprise where commitments are ambiguous, evidence is fragmented, exceptions lack legitimate adjudication, and responsibility dissolves across organizational boundaries. This paper distinguishes between a model constitution that governs actor behavior and an Enterprise Governability Constitution that governs institutional behavior. It also introduces CoordaOS as a working reference architecture for making that institutional constitution computable—and proposes a practical inquiry into how Microsoft’s emerging model constitution might interface with the constitutional requirements of the Operator environment.

​

Robert Long
Founder, Enterprise Governability Lab
September 2026

READ OR DOWNLOAD THE PAPER

​

For research, consultation, or potential collaboration:

Contact Rob Long
rob@enterprisegovernabilitylab.com