Autonomous Agent
An AI system that takes a goal and runs the whole way to done without a human in the loop for each step.
Think of it like
A contractor you hand a key and a punch list — they let themselves in, do the work, and call you when it’s finished.
Example
You say “file my expenses for the trip,” and it reads the receipts, categorizes them, fills the form, and submits — no clicks from you.
How it actually works
It’s the same loop — think, act, observe, repeat — just with the human approvals removed. Autonomy is a dial, not a switch: more autonomy means more speed and less oversight, so the failure modes get bigger too. The honest catch is that a wrong turn early can compound into a very confident, very wrong result before anyone notices.
For product teams
Higher autonomy trades human effort for risk — decide per task how much you’re willing to let it run alone.
For engineers
The control loop iterates without blocking on human approval between actions, bounded only by stop conditions and guardrails.
Related
- Agent Loop — It’s the agent loop running unsupervised.
- HITL — The opposite of keeping a person on every step.
Read anything AI without the jargon
Look up any term in plain English, or save terms as you read with the free Chrome extension.
Open DecoderAdd to Chrome