Case Study: Leading Health Insurance Provider achieves improved transcription accuracy and key-term recognition with Voci Technologies' V-Blaze ASR

Leading Health Insurance Provider - Customer Case Study

Leading Health Insurance Provider, which serves nearly 15 million members, needed to improve its automatic speech recognition (ASR) performance—particularly accurate transcription of healthcare-specific terms—to better track customer experience and NPS. The provider selected Voci Technologies’ V-Blaze ASR speech-to-text solution to address these challenges.

Voci Technologies tuned custom language and acoustic models using roughly 10,000 hours of call audio (about 4 million manually transcribed words) and created separate agent and customer models. The work delivered relative word error rate improvements of over 35% for agent speech and over 20% for customer speech, plus ≥20% recall gains for 650 agent key terms and for more than 2,000 customer key terms (many terms improved by several hundred percent, some exceeding 1,000%). These results enabled the Leading Health Insurance Provider to extract higher-quality analytics, better identify trends, improve agent training, and drive improved customer experience and reduced churn.