What does human in the loop mean?

Human in the loop describes a process where automation or AI does part of the work and a person checks it at defined points. The AI might draft, sort, extract, or recommend, but a human approves the result before it is sent, published, paid, or saved. This keeps judgment and accountability with a person while still removing much of the repetitive effort.

For example, an automation might read incoming invoices, pull out the supplier, amount, and due date, and prepare a payment entry. Instead of paying automatically, it sends each entry to someone in finance for a quick approval. Anything unusual, such as a new supplier or a total that does not match the purchase order, is flagged for a closer look.

The right amount of human review depends on the stakes. Low-risk, easily reversed tasks, such as tagging internal notes, may need only spot checks. Tasks that affect customers, money, legal commitments, or people's jobs usually need a person to approve each result. Some automation tools include approval steps, where a workflow pauses until someone approves or rejects it.

Human review only works if the reviewer genuinely checks. When an AI is usually right, it is easy to start approving without reading, a tendency known as automation bias. Keeping review screens simple, showing the source next to the AI's output, and occasionally auditing approved items all help keep the check meaningful.

An example

An AI drafts replies to customer complaints, and a support lead reads and approves each one before it is sent.