Autonomous agent
An autonomous agent is an AI system that pursues a goal over multiple steps with minimal human intervention — deciding what actions to take, executing them via tools, observing results, and iterating until the goal is achieved or judged unreachable. Autonomy lives on a spectrum: from one-shot completion to fully open-ended multi-day operation.
The autonomy axis is the defining 2024-2026 progression in AI applications. At one end: a coding assistant that proposes a diff for human review (low autonomy, high oversight). At the other: an agent given a vague brief that researches, drafts, tests, and iterates over hours (high autonomy, low oversight). The right point on the spectrum is task-dependent: high-stakes work (financial transactions, code that ships to production) typically warrants lower autonomy and tighter oversight; exploratory or low-stakes work can tolerate more. Production deployment patterns commonly include async-with-checkpoint (the agent runs autonomously but pauses at defined checkpoints for human review).
Related terms
- Agent loop
An agent loop is the orchestration pattern where an LLM iteratively reasons, calls tools, observes results, and continues — until a terminal condition is met (task complete, max iterations reached, error).
- Agentic workflow
An agentic workflow is a software process in which one or more LLM-driven agents drive the work, making decisions and invoking tools at each step rather than following a hardcoded script.
- Multi-agent system
A multi-agent system uses multiple LLM-driven agents working together — typically with role specialisation (planner, executor, critic) or domain specialisation (researcher, coder, writer).