Measure and improve

Reviewing Conversations to find knowledge gaps

In brief: Regularly review negatively rated conversations and monitor ratings on the AI insights page to spot recurring problems in AI responses and prioritize training improvements.

Conversation ratings can reveal patterns of dissatisfaction that may be missed when interactions are considered one at a time. Reviewing them regularly helps you focus on repeated issues and maintain accurate, helpful answers.

Review negatively rated conversations

Check negatively rated conversations regularly to identify patterns in user dissatisfaction. Pay attention to recurring problems in AI responses, since these can point to areas that need improvement.

Use patterns to guide improvements

Recurring response problems can help you prioritize improvements to your AI training. Use those priorities to support more accurate and helpful answers.

Monitor conversation ratings

Check the AI insights page regularly to monitor conversation ratings. This helps you keep track of user feedback and maintain high-quality AI responses.

Regular reviews turn conversation ratings into a way to spot repeated problems and guide training priorities. Continue monitoring ratings to support helpful, accurate responses.