Beyond latency: The art of building a truly great voice agent | Decagon
Apr 9, 2026
Summary
This post details Decagon's approach to building great voice agents by emphasizing conversational design principles (efficiency, naturalness, recovery, empathy, user knowledge) and the importance of a synchronized voice profile (TTS characteristics and persona guidelines). It highlights the need for A/B testing and metric validation to discover the optimal voice profile and conversational design, moving beyond initial technical fundamentals.
More from the Voice Activity Detection (VAD) and Turn Detection trail
- Audio-native semantic speaker-change detection | Decagon
- Why speech-to-speech models aren’t ready for the enterprise (yet) | Decagon
- "Agent, agent, agent": designing conversations to reduce barge | Decagon
- What I learned building voice agents at Decagon | Decagon
- Conversational IVR from definition to business impact | Decagon