Why agent identity needs a full lifecycle model, not just governance

The article argues that the market for governing AI agent work is real but depends on a more basic problem: defining agent identity across its lifecycle. Current governance systems assume a principal already exists, while autonomous agents raise earlier questions about how identity begins, how ownership attaches, and how delegated authority spreads. The author highlights an intermediate 'orphan' state, where an agent acts before a human claims it, as a key reason identity systems must go beyond governance. The piece also presents lineage tracking of which agent spawned which as part of identity itself, and separates identity lifecycle, governance, and runtime enforcement into distinct but interacting layers.