Logs
The Logs page centralizes monitoring of conversations, messages, and usage across the AI Assistant. Track activity, analyze performance, and control data storage and access.
Conversation Saving Policy
The conversation saving policy determines how the AI Assistant handles and stores user interactions within the AI Assistant's chat task. Choose from three approaches:
All conversations are saved and accessible to users of this assistant.
Preserves the complete conversation history. Ideal for:
- Team collaboration - All team members can see the full context
- Knowledge continuity - Previous discussions remain available for reference
- Training purposes - Complete interaction history for analysis
- Customer support - Full context of user issues and solutions
Visibility Settings for Logs
After choosing what to store, control who can view the logs. This setting is independent of the saving policy:
- Share with All AI Assistant Users — All users of this assistant can view the logs and conversation history
- Not Shared — Logs are completely private and not visible to anyone
- Share with the AI Assistant Owner — Only the assistant owner (typically the creator) can view the logs
Quick Reference Matrix
How saving policy and visibility settings interact:
| Saving Policy → / Visibility ↓ | Store All | Store Anonymized | Discard All |
|---|---|---|---|
| Share with All AI Assistant Users | Everyone sees full logs | Everyone sees anonymized logs | No logs exist |
| Share with the AI Assistant Owner | Only owner sees full logs | Only owner sees anonymized logs | No logs exist |
| Not Shared | No one sees logs | No one sees logs | No logs exist |
Why Use Logs?
Logs provide insights into AI Assistant performance and usage through its dedicated chat interface:
- Activity Monitoring — Track message and conversation volumes to understand usage patterns
- Data Source Analysis — Identify which data sources are most frequently accessed and useful
- Quality Assurance — Find and address unanswered messages to improve user experience
- User Engagement — Analyze usage patterns and trends to optimize your assistant
- System Health — Monitor performance metrics and identify potential issues
- Compliance — Maintain records for audit trails and regulatory requirements
Logs Dashboard Features
The logs dashboard monitors and manages AI Assistant activity:
Overview & Analytics
The main dashboard displays:
- Key Metrics — Total conversations, messages, and active users
- Activity Charts — Visual trends over time to spot patterns
- Current Settings — Your active Saving Policy and Visibility Settings at a glance
- Real-time Updates — Live monitoring of current activity
- Log Search — Search for specific conversations and messages for in-depth analysis
Message Sources Analysis
Tracks which data sources are used most effectively:
- Source Usage Stats — See which files and documents are accessed most frequently
- Export Capabilities — Download usage data for further analysis
- Customizable Columns — Configure the table to show the metrics that matter to you
- Performance Insights — Identify which sources provide the best answers
Messages Without Answers
Identifies messages the assistant could not answer satisfactorily:
- Failed Responses — See messages that didn't receive satisfactory answers
- Context Analysis — Understand why certain questions weren't answered
- Q&A Creation — Convert problematic interactions into new Q&A entries
- Continuous Improvement — Use this data to enhance your assistant's knowledge base
User Management
Monitors and manages the assistant's user base:
- User Directory — Complete list of users who have interacted with your assistant
- Usage Statistics — Conversation counts and engagement levels per user
- Geographic Data — Location information (when available) for user analytics
- Account Management — First-seen dates and user activity tracking
- Data Control — Delete logs for specific users when needed
Best Practices
Apply these guidelines to maximize log value while maintaining security and compliance:
Choosing the Right Settings
- Store All — Best for internal teams where full context is valuable
- Store Anonymized — Ideal for public assistants or when privacy is important
- Discard All — Perfect for testing environments or strict privacy requirements
Visibility Guidelines
- Set Visibility to match your compliance model (follow the principle of least access)
- Consider your organization's data governance policies
- Regularly review who has access to sensitive conversation data