The ability of an AI system to perform multi-step tasks and make decisions without human intervention at each stage.
In agentic AI, operational autonomy allows the system to decompose a high-level goal into sub-tasks, select tools, and execute actions independently. While this creates massive efficiency, it also complicates governance, as the system may take irreversible actions (like deleting data or moving funds) before a human can intervene, requiring automated escalation triggers.
A procurement agent is given the goal restock office supplies. It autonomously checks inventory, compares vendor prices, selects the best deal, and executes the purchase without a human clicking approve for each step.
