Autonomous Agent
An autonomous agent is an AI agent that can complete a multi-step goal without human intervention between steps, deciding its own sub-tasks and tool calls.
In plain English
An autonomous agent runs on its own. Once you set the goal, it plans, acts, observes, and iterates without asking for permission on every step. It may still pause on high-stakes decisions — send a payment, email a prospect, deploy to production — but the routine work in between runs end to end.
Autonomy sits on a spectrum. A low-autonomy agent asks before every tool call. A high-autonomy agent runs a ten-step plan and only surfaces the final result. Most production agents land in the middle: run freely for low-risk actions, require approval for the expensive ones. The right setting depends on the task's cost of failure and the strength of the guardrails around it.
Why it matters for Black Box
Black Box defaults to high-autonomy execution with an approval inbox for irreversible actions. The owner sets a goal, the system runs, and only decisions that cannot be undone — paid ad spend, public sends, production deploys — stop and wait for a click.
Examples
- An agent that researches, drafts, edits, and files a weekly report with no check-ins.
- A lead-gen agent that scrapes, scores, enriches, and queues outreach — but asks before sending.
- A coding agent that implements a feature, runs tests, and opens a PR for review.