Summary
Introduce a built-in feedback mechanism for AI-generated replies in Jitbit Helpdesk, allowing end users (customers) to rate chatbot responses directly within the ticket.
Background / Problem
With the increasing use of AI-generated replies via automation rules, customers may receive responses that are not always accurate or helpful. Currently:
- Users have no direct way to evaluate the quality of an AI-generated response.
- Negative feedback is often misdirected to the overall ticket or support agent.
- Agents may correct AI mistakes later, but customer dissatisfaction may already be recorded unfairly.
Proposed Solution
Add a lightweight feedback feature specifically for AI-generated replies:
- Display thumbs up / thumbs down buttons under AI-generated responses.
- Optionally allow users to leave a short comment when selecting thumbs down.
- Track feedback separately from agent satisfaction ratings.
Key Functionalities
- Feedback is linked to a specific AI reply, not the entire ticket.
- Feedback is visible internally (admins/agents) for monitoring AI quality.
- Optional: flag negatively rated responses for review or improvement.
- Optional: analytics dashboard with AI feedback trends.
Benefits
- More accurate measurement of AI performance.
- Prevents penalizing human agents for AI mistakes.
- Provides actionable insights to improve chatbot rules and training.
- Enhances transparency and trust with customers.
Optional Enhancements
- Ability to disable feedback per project or rule.
- Trigger automation (e.g., notify agent) when a reply gets negative feedback.
- Export/reporting on AI performance metrics.
Use Case Example
A chatbot responds automatically via automation. The customer finds the answer incorrect and clicks 👎, adding a short comment.
Later, an agent corrects the issue. The ticket can still be rated positively overall, while the AI feedback remains accurately captured.