Agentic AI is artificial intelligence that pursues a goal by planning and taking a sequence of actions on its own, instead of producing a single response to a single prompt. An agentic system breaks a goal into steps, uses tools or data to carry each one out, observes the result, and adjusts its next move. The distinguishing trait is autonomy over a series of actions rather than one answer.
Given an objective like preparing a weekly performance summary, an agentic system might query several data sources, compile the figures, identify what changed, and draft the summary, deciding the order of steps itself rather than following a fixed script. It can adapt when a step returns something unexpected.
The distinction matters because autonomy shifts what software can be trusted to do without a person directing each step. A passive model answers a question and stops, while an agentic system can carry a task from start to finish. That capability raises real questions about oversight, guardrails, and accountability, since a system that acts on its own needs clear boundaries and the ability for people to review and correct what it does.
