Case Study: telli achieves more reliable enterprise voice AI with ai-coustics

Telli reduces voice agent errors across 5 million calls with ai-coustics

The customer, telli, a European AI startup automating high-volume phone operations for enterprises, faced the challenge of unreliable voice agents due to poor audio quality from noisy environments and telephony compression. This was causing misunderstandings, breaking turn-taking, and escalating costs. They worked with vendor ai-coustics and its Quail Voice Focus model to address this core issue.

ai-coustics integrated its SDK and models directly into telli's LiveKit pipeline to optimize audio specifically for machine understanding, not human listening. This solution dramatically improved speech detection and transcription accuracy for short utterances. The results were significant, enabling telli to deploy voice agents reliably at scale for major clients and formalize call quality into a benchmark-driven process.

Telli