Case Study: Nurix achieves enterprise-grade reliability for high-volume voice AI workflows with Cekura

Nurix cuts voice AI call failures by 30% with Cekura

Nurix, an enterprise AI company, provides critical voice AI infrastructure for large enterprises. They faced significant challenges in deploying a flagship voice agent for a major client, including handling complex mixed-language inputs, preventing disconnections on short user answers, and supporting hundreds of concurrent calls without latency or failure. To move beyond manual testing, they partnered with Cekura and its simulation-driven testing platform.

Cekura's solution involved high-velocity, persona-driven simulations to conduct multilingual chaos testing and identify fragile Voice Activity Detection thresholds. The results were substantial: Nurix successfully patched its model to handle language transitions, eliminated a bug that dropped calls on short answers, and optimized webhook responses to reduce latency from 2.65 seconds to sub-second levels. This ensured enterprise-grade reliability and saved thousands of potential customer interactions from being abandoned.

Nurix