AI Conversation Reports
View how well your AI chatbot handles customer conversations and identify patterns in your support interactions.
Understanding the Metrics
Your AI chatbot's performance is measured across all customer conversations:
Successful Responses
Conversations where the AI provided helpful answers, connected users to agents when requested, or triggered automated actions.
Neutral Responses
Basic interactions like greetings or simple questions that didn't require specific knowledge.
Failed Responses
Conversations where the AI couldn't provide adequate help, leading to automatic agent requests or errors.
What You Can See
Overall Performance
View success rates across all AI interactions to understand how well your chatbot performs.
AI-Handled Conversations
See which conversations were fully managed by your AI without needing human intervention.
Response Breakdown
For each conversation, see how many responses were successful, neutral, or failed.
Individual Chat Access
Click on any chat ID to view the full conversation and understand what happened.
How Conversations Are Counted
These are conversations that your AI managed completely without any human agent involvement. A conversation counts as "AI-handled" if it didn't require agent intervention.
Long conversations are divided into segments that are at least 180 minutes apart. Each segment is counted separately to give you accurate insights into conversation patterns.
Each AI response in a conversation is categorized as successful, neutral, or failed based on whether it provided value to the user.