Krisp VIVA 2.0 Aims to Fix Voice AI's Real-World Audio Challenges
Event summary
- Krisp launched VIVA 2.0, a voice AI infrastructure layer designed to improve real-world audio handling for voice agents and IVRs.
- New features include Turn Prediction v3, Interrupt Prediction v1, and Signal Detectors for synthetic speech, accent, and gender identification.
- VIVA 2.0 processes over 12 billion minutes of voice AI agent traffic annually and is embedded in over 130 voice AI products.
- Customers report 3.5x improvement in turn-taking accuracy, 50% fewer dropped calls, and 30% higher customer satisfaction with VIVA.
- VIVA 2.0 will be showcased at Twilio Signal 2026 on May 6-7 in San Francisco.
The big picture
Voice AI adoption surged 9x in 2025, but real-world audio challenges like background noise and misfires in voice activity detection have limited effectiveness. Krisp's VIVA 2.0 addresses these gaps by providing a dedicated infrastructure layer that enhances the reliability of voice agents in messy, real-world environments. The solution is part of a broader trend toward improving the quality of audio signals in AI-driven conversations, which is critical for the scalability of voice AI applications.
What we're watching
- Adoption Pace
- How quickly voice AI developers and enterprises will integrate VIVA 2.0 into their existing systems.
- Performance Metrics
- Whether the reported improvements in turn-taking accuracy and customer satisfaction will be sustained at scale.
- Competitive Response
- How competitors in the voice AI infrastructure space will react to Krisp's enhanced real-time audio capabilities.
Related topics