Glossary

Human approval (human in the loop)

Human approval means a named person must approve a step before an AI agent carries it out. You choose which steps those are, and the decision goes on the record.

In a run

The agent reaches a step you marked as sensitive and stops. The person asked sees the step, the limit it crossed and their own role. Approve and reject both go on the record, with who decided and when.

Our word for it

Sensitive steps go to a person.

Which steps wait by default

Four kinds of step already wait for a person: widening a permission, touching another agent’s identity, reaching a higher-tier system, and keeping data beyond the job. You add the steps that matter in your business, such as an order above a threshold or a change to a delivery date.

An example

In the requisition run, the agents check stock and draft the order. The order is above the threshold, so it stops and waits for the budget holder. Nothing is placed until that person decides.

Start supervised

A person can approve each sensitive step until you choose otherwise. You decide which steps stop needing approval, and when.

Approve and reject are both records

A rejection is as useful as an approval. It shows that a person stopped a step, what the step would have done, and why it reached them. Both stay on the record with the rest of the job.

Where it stops

  • Ares does not decide which steps are sensitive. You do.
  • The record holds the request the person saw beside their decision. Training the people who approve is yours.

Related terms

  • Workflow automationWorkflow automation is software that carries a business process through its steps without a person doing each one by hand.
  • AI agent audit trailAn AI agent audit trail is the record of what an agent did, for whom and under what permission, written as each step happens.

Tell us about one workflow.

We will say whether an agent should do it, and what the first step costs.

A senior member of the team replies within two business days.