11 docs tagged with "human-in-the-loop"

How to answer when your AI Agent uses the Ask User Questions tool — wizard-style, one question at a time.

What the approval card looks like in a conversation, and how to approve or dismiss a pending tool call.

Let your AI Agent pause mid-conversation to ask you single-choice or multi-choice questions before continuing.

Require your AI Agent to pause and ask for your approval before running a specific tool.

The mental model behind Human-in-the-Loop — why your AI Agent pauses, how cascade resume works, and what happens in AI Teams.

Keep a human in the loop — let your AI Agent pause, ask, and wait for you before it does anything that matters.

Approve, edit, or reject every approval waiting for you — with feedback, parameter edits, and a link back to the conversation.

Handle approvals in AI Teams — when one AI Agent calls another, and the inner one wants to use a tool that needs your OK.

Every status a task can be in, what it means, and what causes the next transition.

A task looks stuck — here's how to work out what's blocking it and get your AI Agent moving again.

Build an AI Agent that creates HubSpot contacts — and have it pause for your approval before every write.