Effective AI governance in a large organization requires a structured, yet adaptable, framework. The board's request for "clear guardrails and accountability" is best met by establishing central principles that guide AI development and deployment. This includes defining roles and responsibilities (accountability), classifying projects by risk level, creating clear approval workflows, and setting up continuous monitoring. Integrating this AI-specific framework into existing enterprise risk and compliance processes ensures consistency, leverages existing expertise, and avoids creating isolated, inefficient silos. This approach balances innovation with responsible oversight.

A. This decentralized approach leads to inconsistent standards, duplicated effort, and a lack of central accountability, directly contradicting the board's request.

B. A single, static document is impractical for the rapidly evolving field of AI and creates a rigid, bureaucratic process that stifles innovation.

D. Governance must extend beyond procurement to cover the entire AI lifecycle, including data usage, model behavior, user access, and ongoing monitoring.

1. Microsoft. (2022). Microsoft Responsible AI Standard, v2. "Governance: We will establish and maintain a robust and adaptive governance system to enable oversight, accountability, and compliance with our Responsible AI Standard." (Section: "Our Principles and How We Put Them into Practice", p. 6).

2. Microsoft Learn. (2023). Establish an AI governance strategy. Cloud Adoption Framework for Azure. "Effective governance helps your organization empower teams to work toward business goals... Your governance strategy should include: Roles and responsibilities, Risk management, [and] Compliance." (Section: "Key elements of an AI governance strategy").

3. Brynjolfsson, E., & McAfee, A. (2017). The Business of Artificial Intelligence. Harvard Business Review. This article emphasizes the need for new organizational structures and governance to manage AI's impact, moving beyond simple technology adoption. (Section: "What Companies Need to Do").