A structured lesson workspace with readable content, hands-on examples, and a clean path to completion.
Give an agent too much freedom, and it makes expensive mistakes. Give it too little, and it is just a slow chatbot. The key is designing the interface so humans stay in control while the agent handles the heavy data lifting.
Human-AI interface
This video requires an access code. Enter your code below to unlock it instantly.
Don't have a code? Contact us to get access.
| Pattern | Description | Best Use Case |
|---|---|---|
| Human in the Loop | Human edits/approves before execution | Pricing, external emails, compliance |
| AI in the Loop | Agent assists a human-led process | Data gathering, internal drafting |
In n8n, use the 'Wait' or 'Human in the Loop' nodes to pause the workflow. The agent prepares the quote and drafts the email, but the execution halts. A Slack notification is sent to the sales desk. A human reviews the HTML rendering and clicks approve.
Design with a high level of human oversight at launch. Once the evaluation metrics prove the agent is consistently accurate over thousands of iterations, you can begin removing the approval gates for low-risk actions.
Finish the lesson once you have worked through the material. This awards ★ 30 XP.