Collect instant user feedback on every AI agent message

End users can share quick feedback on individual AI agent messages, helping your team continuously refine model quality and conversational accuracy. With a simple thumbs up/down control and optional short comment, users can instantly flag helpful or inaccurate responses. Feedback is visible to human agents and admins directly within the Dashboard, empowering data-driven improvements to both AI and human performance.
 

What’s new

• Per-message feedback: Users can rate each AI agent generated message (👍 / 👎) and add short notes (up to 100 characters) to explain the rating. Feedback can be modified or deleted before the chat is resolved.
• Flagged messages view: Access a consolidated list of all rated messages (filter by 👍 or 👎) with timestamps and conversation links. Monitor patterns over time through a trend chart that visualizes positive vs. negative feedback, helping identify opportunities for AI model improvement.

Why it matters

This creates a lightweight, actionable feedback loop between your users and your AI agent so you can better understand where the AI performs well, where it falls short, and how human or model interventions impact satisfaction over time.