Closed-Loop Automation

Closed-loop automation is the control pattern underneath most modern network self-management, and it's worth understanding as a literal feedback loop: monitor the network, analyse what the data means, decide on an action, act, then go back to monitoring to confirm the action worked. The "closed" part is the key word — an open loop would act and walk away; a closed loop verifies and adjusts.

The loops come in degrees of autonomy, which is where the nuance lives:

• Human-in-the-loop — the system recommends, a person approves before anything changes.
• Human-on-the-loop — the system acts on its own, with people supervising and able to intervene.
• Fully autonomous — the loop runs end to end with no routine human involvement.

Most operators deliberately start at the supervised end and earn their way toward autonomy as trust builds, because a bad automated decision can propagate across a network faster than any human can catch it. This pattern is the engine inside ZSM, intent-based networking, and AIOps alike.