Articles by Hans Schabert
-
Authority First, Governance Follows
When every team defines its own agent authority, the organization has no authority at all. Centralized authority through SOP-MCP makes decision rights explicit, scoped, and enforceable — governance follows.
-
Give Agents Autonomy. Give Humans Reason to Trust.
AI agent governance should focus on trust, not control. By encoding SOPs with MUST/SHOULD/MAY requirement levels, organizations create a dual-purpose trust contract.
-
No System Is 100% Right
Every automation decision rests on a cost baseline. Most baselines are structurally incomplete. That incompleteness distorts every decision that follows.