Customer Stories

Deepgram and Red Box: 100% audio capture and transcription boost insights

Increasing Transcription Accuracy and Offering Bespoke Customization While Maintaining a Market Leading Low Compute Footprint is a Game-Changer for Red Box Customers

The Landscape

Unlocking Voice, the Most Powerful Dataset in the World

Major technological advances in voice technology have enabled organizations to more seamlessly extract insights from conversations at scale by turning audio files into structured datasets for artificial intelligence (AI) and machine learning (ML) engines to analyze. At the same time, with new opportunities to leverage voice, established call recording practices that historically served companies well for compliance and quality purposes, are now holding them back from fully maximizing ROI benefits from speech analytics and enablement investments.

Red Box is uniquely positioned to address this issue and the growing demand for organizations to capture, secure, and unlock the value of enterprise-wide voice data. As the leading voice specialist for 30 years, Red Box provides enterprises with open access to, and control of, high-quality unstructured and structured audio data and the freedom to leverage it in any application they choose. Conversa, Red Box’s new and the industry’s first truly open, microservices-based enterprise voice platform, has been engineered for flexibility and scalability and offers low compute footprints, market-leading total cost of ownership, and flexible deployment options.

The Challenge

Meeting the Additional Requirements 
of High-Profile Customers

With voice datasets increasingly seen as a strategic asset awash with rich insights, timely access to high-quality transcripts became critical for Red Box’s customers, which include leading organizations across financial, contact center, government and public safety sectors. Automatic speech recognition (ASR) has been a part of Red Box’s offering for the last three years, but once the company started working with more high-profile customers—such as six of the world’s top banks—it needed a platform that offered customization and could scale accordingly.

“An urgent need for improved scale and performance drove us to find a new ASR solution,” said Pete Ellis, Chief Product Officer at Red Box. “We could see that the requirements for our more high-end users, who employ 30 to 40 thousand call center agents, needed a more elegant and performant architecture in their software solutions.”

The Solution

Deepgram is Unmatched in Accuracy, Speed 
and Customization

When Red Box started evaluating the other ASR solutions on the market, Deepgram blew its competition out of the water in terms of accuracy, speed, customization and compute needs. “We pride ourselves on how fast we can capture audio and deliver it to any best of breed analytics application, and Deepgram had a similar mentality when it came to the speed and high-quality of its transcriptions,” said Ellis. “Especially during the COVID-19 pandemic, we’ve seen great growth in our healthcare customer segment, and customizable transcriptions are especially critical for those organizations.”

Deepgram was also the perfect fit because it uniquely targets and meets the needs of each of Red Box’s key customer personas. Deepgram’s AI speech platform is built from the ground up to optimize GPU performance with up to 300 real-time streams on one GPU and hence can reduce the overall computing and carbon footprint.

“IT teams love Deepgram’s speed and accuracy, while technology teams appreciate how the platform doesn’t use the same open-source space that other vendors do, which helps with the total cost of ownership,” said Ellis. “Compliance users are laser-focused on ensuring that their customer-facing users adhere to compliance rules, which Deepgram’s high-quality transcripts enable them to do. And on the contact center side, the platform allows them to examine customer experiences both from a sales and an inbound service perspective, which helps them with training and coaching.”