Organizational Behavior for Non-Human Actors: Why Every Existing Leadership Model Fails in Agentic Enterprises — and What the Missing Doctrine Actually Requires

The departure of major enterprise CEOs citing AI as a factor in their decisions to step down is not a talent signal. It is a structural signal. Coca-Cola’s James Quincey and Walmart’s Doug McMillon are not leaving because they failed to adopt AI. They are leaving because the leadership model under which they were selected — optimized for predictability, hierarchical control, linear workflows, and human-executed accountability — is structurally incompatible with the operating environment that agentic AI creates. This paper argues that no existing leadership doctrine knows how to govern a workforce where workers are agents, teams are dynamic clusters, tasks are decomposed and recomposed in real time, accountability cannot be assigned to a single actor, judgment must be human but execution is machine-distributed, risk propagates at machine speed, and oversight must be substrate-level rather than managerial. The five broken assumptions of legacy leadership, the four structural gaps in board governance, and the five components of the missing agentic leadership doctrine are each addressed in turn. The central claim: what is needed is not AI leadership. It is organizational behavior for non-human actors.