How Can I Disable n8n Bot Responses When a Human Joins the Conversation?

Based on our documentation, there isn’t an out-of-the-box setting to automatically disable all AI responses once a human intervenes. However, you can design your workflow to achieve this behavior.

Custom Workflow Logic

A common approach is to introduce a flag or state indicator (for example, a field like “humanIntervention”) that your workflow checks before sending an automated response. When your human support (for example, via Slack) replies, you can update that flag to indicate that the conversation is now under human control. Subsequent triggers or messages would then bypass the AI response and follow a different path (or simply do nothing).

Example Strategy

1. Catch the Human Intervention:
   Use a node (for instance, an HTTP Request or a webhook) that listens for a human response from your chosen communication channel (like Slack).

2. Update the Conversation State:
   Once a human response is received, update a field or flag (using a Set node) in your conversation context or database record.

3. Conditional Logic:
   In your AI Agent workflow, add a condition (with an If node) early on that checks for the “humanIntervention” flag. If the flag is set, route the flow away from the automated response logic.

Workflow Components to Consider

• Chat Trigger Node:
  This node starts the conversation. You can configure it to check or include context from previous conversation states.

• Custom Workflow Tools or Set Node:
  These allow you to modify the ongoing conversation state by setting or updating fields that indicate human intervention.

Additional Resources

You might find the human fallback flow example (which integrates a fallback to human assistance via Slack) useful as a reference. While it focuses on sending a human notification when the AI can’t respond, the same principle applies to marking a conversation for human takeover.

---

By implementing this custom logic, you’ll be able to prevent further automated responses once a human has stepped in. Unfortunately, a fully built-in toggle for this behavior isn’t available in n8n, so using workflow nodes and conditional logic is the recommended solution.

Let me know if you need further details on setting this up!