ElevenLabs Voice Agent Observability with Galileo | Multi-Turn Evaluation Tutorial

See how to monitor and evaluate ElevenLabs voice agents with Galileo's comprehensive observability platform. This demo shows real-time tracking of voice agent conversations with out-of-the-box metrics for completeness, efficiency, and request fulfillment. In this video, our Field Engineer, Al Chen, walks you through: Setting up Galileo handler for ElevenLabs voice agents Automatic session tracking for multi-turn conversations Turn-by-turn conversation analysis in the dashboard Agent-specific metrics: efficiency, completion, and correctness Filtering sessions to identify agent performance issues Debugging incomplete responses with detailed rationales 🔗 Try Galileo: https://app.galileo.ai/sign-up?utm_medium=organic&utm_source=youtube Galileo provides 9 purpose-built agentic metrics, including Action Completion, Agent Efficiency, Conversation Quality, and Tool Selection Quality—helping teams ship reliable voice agents faster. Perfect for teams building voice assistants, conversational AI, and multi-turn agent workflows with ElevenLabs. 📚 Docs: https://v2docs.galileo.ai/ 💻 Check out the example repo here: https://github.com/rungalileo/sdk-examples/tree/main/python/chatbot/elevenlabs-chatbot 0:00 - The Voice Agent Observability Challenge 0:09 - Out-of-the-Box Metrics for Voice Agents 0:21 - Demo Setup: Product Marketing Voice Agent 0:36 - Live Voice Agent Conversation Example 1:24 - Real-Time Session Tracking in Galileo 2:09 - Implementation Code Walkthrough 2:27 - Automatic Agent & Human Speech Tracking 2:45 - Understanding Metric Rationales 3:03 - Inspecting Individual LLM Spans 3:21 - Additional Agentic Metrics Available 3:33 - Build Your Own Voice Agent Integration