How to identify conversations requiring human intervention in AI Assistant interactions

When the AI Assistant cannot give a satisfactory answer or solution, it signals the need for human input. This ensures that issues or queries are promptly addressed by the right personnel.

Identifying Messages Requiring Human Intervention

When the AI Assistant is enabled, messages that need human intervention are clearly labeled and come with specific indicators:

  • Sidebar Labels: Conversations in the sidebar will be labeled the same way as Unread messages to indicate that human intervention is necessary. This visual cue helps you quickly distinguish between automated responses and those requiring personalized assistance.
  • Inbox Individual Conversations: Messages needing human intervention will be visibly labeled within your Inbox. Look for the red notification message: "Human response needed! AI Assistant is paused." This label ensures these conversations are easily identified for your attention.
Indication Within Conversation When Human Intervention Is Required 
Indication Within Conversation When Human Intervention Is Required 

Once you've identified a conversation that requires human intervention, take the following steps:

  1. Follow-Up: Read the history of communication that AI had and continue the conversation as needed, helping your client get their answer(s).
  2. Resume AI functionality: Use the blue "Mark As Solved" Call-to-Action (CTA) to indicate that the conversation has been addressed by a human and can now be resumed with the AI Assistant. When this is done, the AI Assistant will resume its functionality, providing further assistance when the client needs it - following up after client responds to your previous message (managed manually)

Initial AI Assistant Message When Human Intervention Is Required 

When the AI Assistant identifies a need for human intervention, it will initiate the conversation with an acknowledgment message. This message serves to inform the user that their query has been forwarded for human review, ensuring transparency and proactive communication.

Example initial message:

  • "I've passed along your message to the relevant team. They'll reach out to you shortly to assist further."

However, you can manage this functionality in the AI Assistant Settings page, and disable it/switch off if you prefer AI not to respond at all.

By default this functionality is disabled.

AI Assistant Settings Page - Manage AI Response When No Right Answer Is Available
AI Assistant Settings Page - Manage AI Response When No Right Answer Is Available

Was this article helpful?

How to recognize messages sent by the AI Assistant?
How to set booking policies to be followed by AI Assistant?