Case Study: Sharpen achieves 90%+ accurate, real-time contact center transcription with Deepgram

Sharpen Elevates the Contact Center Customer Experience with Deepgram

Sharpen, a cloud-native, agent-focused contact center platform, needed a more accurate and affordable automatic speech recognition (ASR) solution after its legacy tri-gram–based provider produced poor transcriptions that threatened compliance and customer satisfaction. To address this, Sharpen evaluated alternatives and selected Deepgram’s automatic speech recognition platform for the enterprise.

Sharpen fed hours of calls into Deepgram’s general audio model and built custom trained models, leveraging Deepgram’s end-to-end deep learning approach to deliver fast, reliable transcripts. The move produced measurable gains — transcription accuracy above 90% even in noisy, multi‑speaker conditions, improved voice-search and QA workflows, and a far more cost-effective solution compared with legacy Big Tech options (which Sharpen estimated would cost roughly eight times more). Deepgram’s real-time capabilities also freed Sharpen to focus engineering resources on new features and deeper analytics for customers.

Sharpen

Adam Settle

Vice President of Product