The CEO-CIO Compact for Agentic AI: Who Owns Risk, Spend, and Business Outcomes?

The article argues that agentic AI creates an accountability problem before a technology problem, as AI agents can autonomously interpret goals, retrieve data, select tools, spend money, and initiate workflows. It proposes a CEO-CIO compact with a dual-key operating model: the CEO owns business decisions, risk tolerance, investment thesis, and outcomes, while the CIO owns the technical system enforcing authority boundaries, data protection, identity control, evidence, and service operation. The compact requires named decision rights for budgets, data access, human approval, incidents, scaling, and decommissioning. AI spend should be measured against business units of value (e.g., cost per case resolved) rather than just tokens or API calls. The article references NIST, ISO/IEC 42001, and FinOps principles, emphasizing that an AI committee is insufficient; executive leadership must establish clear roles and responsibilities.