FIELD NOTES
Clear thinking for complicated governance.
Short perspectives on data, AI, controls, accountability, and the gap between policy and operation.
AI GOVERNANCE
AI governance without data governance is mostly theater.
An AI inventory and policy may create the appearance of control. But if the underlying data lacks ownership, lineage, quality expectations, access discipline, and an auditable chain of accountability, the governance layer is built on presentation rather than proof.
Real trust in AI begins with trusted data—not a stage.
CONTROL DESIGN
The policy is not the control.
A policy expresses intent. A control assigns an action, owner, trigger, evidence, exception path, and consequence. Confusing the two is how mature-looking programs fail audits.
ACCOUNTABILITY
Committees do not own risk.
Cross-functional participation matters, but collective discussion is not accountable ownership. Someone must have the authority and obligation to decide.
READINESS
Can you prove the control operated?
If evidence is assembled only when an auditor asks, the control environment is already telling you something important.