Glossary
What is Human-in-the-loop?
Human-in-the-loop is a design where a person reviews or approves key steps in an automated process before they run.
What it means
Automation handles the routine, and a person makes the call on risky actions. A payment, a message to a customer or a spending decision are good places for a human check. The machine proposes, and the person decides.
Where to put the checkpoint
Put the check before any action that is hard to undo. Sends, payments, data deletion and public posts need approval. Routine steps such as tagging or sorting usually do not.
Keeping it fast
Show the full action in the approval request, not a vague summary. Set a time limit, so requests do not wait forever. Let people approve from their phone, so a decision takes minutes.
Example
A Gurugram insurance agency’s AI agent drafts a renewal reminder. Before anything goes to customers on WhatsApp, the agent asks the account manager to approve the exact text and the recipient list. The manager approves in the morning, and the messages go out at the scheduled time.
In Montr, the AI agent stops for your approval before risky actions. These include WhatsApp sends, voice calls, campaign sends and destructive CRM changes. Workflows can also pause and wait for a person to fill in a form or approve a request.