Logs
The Logs section allows you to monitor and review how your AI Agent interacts with users.
It provides full visibility into conversations, including:
User messages
Agent responses
Agent reasoning
Confidence scores
This helps you understand agent behavior, troubleshoot issues, and improve performance.
What You Can See
In the Logs section, each interaction session includes:
Conversation History
Full message exchange between the user and the AI Agent
Organized per session
Agent Reasoning
The internal reasoning behind each response
Shows how the agent interpreted the request
Indicates whether knowledge, skills, or instructions were used
Confidence Score
A score indicating how confident the agent is in its response
Helps identify:
Reliable answers
Uncertain or risky responses
How It Works
Select a session from the left panel
View the conversation details
Expand reasoning for each message
Analyze confidence scores
Identify issues or improvement areas
Supported Channels
Logs in this section are available for:
Playground
Mate Chat
External Channel Logs
For the following channels:
WhatsApp
Microsoft Teams
Webhook
Logs are not displayed in this section.
Instead, you can monitor these interactions through:
When to Use Logs
Use the Logs section when you:
Want to understand how the agent responds in real scenarios
Need to debug incorrect or unexpected answers
Want to verify knowledge or skill usage
Analyze user interactions
Improve agent instructions and behavior
Best Practices
Review sessions regularly
Check reasoning for incorrect responses
Monitor low-confidence answers
Use insights to refine:
Instructions
Knowledge
Skills
What’s Next
After reviewing logs, you can improve your agent by:
Updating Agent Instructions
Enhancing Knowledge sources
Refining Skills configuration
Summary
Logs provide visibility and control over your AI Agent.
Understand what happened
See why it happened
Improve future responses
Last updated
Was this helpful?