Two Layers of a Single Control Problem
AI governance and LLM governance are frequently used interchangeably in vendor material and internal policy documents, which creates confusion about where controls should live. AI governance is the broader organizational discipline: it establishes risk classification schemes, model inventories, approval workflows, and review cadences that apply across every AI system an enterprise operates, regardless of architecture. It answers questions like which systems require review, who owns risk acceptance, and what documentation is required before deployment. LLM governance is a subset of that discipline, scoped specifically to the behavior and outputs of large language models and the agents that wrap them. It answers a different set of questions: what a model is allowed to generate, which tools an agent can invoke, and how those actions are constrained at the moment of execution. Both are necessary. Neither substitutes for the other.