Supervised autonomy
Giving an AI agent independence in stages: it starts by proposing actions for approval and runs on its own only as its work proves reliable.
A new procedure might begin with a person reviewing every action. As accuracy is confirmed, routine cases run automatically, while sensitive actions, such as write-offs or credit decisions, stay gated for approval.
This lets a team adopt automation without giving up control, and move an agent back to supervised mode if its results slip.
Further reading
Related terms
- Human in the loopA design in which an automated system hands defined decisions to a person to review or approve before it acts.
- Agent operating procedureA written procedure that tells an AI agent how to handle a type of work, encoded from the standard operating procedures a finance team already follows.
- Audit trailA chronological record of every action taken on an account, who or which agent took it, and why, so decisions can be reviewed later.