Skip to main content

Observability Overview

The Galene.AI Platform provides comprehensive monitoring and auditing capabilities for all agent and chat interactions through its Observability features. This system enables super administrators to track, analyze, and audit every conversation, ensuring transparency, quality control, and compliance with AI Act regulations.

Super Admin Access Required

Observability features are exclusively available to super administrators who need to monitor platform interactions across all organizations, review agent behavior, and maintain audit trails for compliance.

Why Observability Exists

The Observability system is designed to meet AI Act and GDPR requirements for high-risk AI systems.

Regulatory Compliance

The platform complies with AI Act regulations:

ArticleRequirement
Article 12Maintain comprehensive logs of AI system operations
Article 13.3(d) and 3(f)Ensure transparency and enable oversight
Article 14Enable meaningful human oversight of AI behavior
Article 19Implement automatic logging systems
Article 26.2, 26.5, 26.6Meet deployer monitoring and logging obligations
Article 72Support continuous performance monitoring

Conversations contain personal data protected by GDPR. Access requires legitimate business justification.

Legitimate reasons for access:

  • Security incident investigation
  • Policy violation reports
  • Quality assurance reviews
  • User complaint investigation
  • Compliance audits
  • Legal or regulatory requirements
Legal Responsibility

Accessing conversation traces without legitimate business purpose may constitute a GDPR violation. Every access to personal data must be justified, documented, and logged. The Audit system ensures accountability and protects all parties involved.

The Two Systems

Traces

Traces provide comprehensive records of conversations between users and AI agents, including the full conversation thread, agent processing details, Generative Shield activity, and performance metrics. Use Traces to monitor agent quality, investigate user complaints, and analyze conversation patterns.

Learn More: Traces Documentation →

Audit

Audit automatically logs every access to conversation traces, recording who viewed what, when, and from where. These immutable logs are essential for GDPR compliance and demonstrating lawful data processing. Use Audit to track access patterns, investigate security incidents, and support compliance audits.

Learn More: Audit Documentation →

How They Work Together

Traces capture what happened in conversations, while Audit captures who viewed them and when. When you access a trace, the system automatically creates an audit log entry, providing complete accountability for compliance reviews and regulatory audits.

How to Use This Section

If you need to...

  • Review a specific conversation → Go to Traces
  • See who accessed a conversation → Go to Audit
  • Investigate a user complaint → Start with Traces
  • Monitor access for compliance → Go to Audit
  • Analyze agent performance → Go to Traces
  • Demonstrate GDPR compliance → Go to Audit

Next Steps

  1. Traces - Complete guide to conversation monitoring and analysis
  2. Audit - Detailed access logging and compliance tracking