Capio: AI-Powered Speech Recognition

Dr. Ian Lane, CEO and Co-Founder, CapioDr. Ian Lane, CEO and Co-Founder
Nearly every Marvel Universe fan has fancied the idea of having a full-length conversation with J.A.R.V.I.S., the highly advanced artificial intelligence (AI) system created by Tony Stark. While J.A.R.V.I.S. has several capabilities, the speech recognition feature is the most captivating. Over the years, several industries have taken to speech recognition in an attempt to break the walls that separate them from the future. One of the most common grievances faced by the organizations that use speech recognition is the lack of solutions that are specific to their use cases. Besides, most speech-to-text solution providers are cumbersome, with half the accuracy as human transcribers. Capio is enhancing speech recognition with machine learning to create a best-in-class cognitive solution. “Our products are the torch-bearers into a world where voice-enabled devices not only take commands but also understand natural language and interact humanly,” gleams Dr. Ian Lane, the CEO and co-founder of Capio. Here is how.

Capio offers an AI-based automatic speech recognition (ASR) solution that provides human quality transcription with the capability of continuous learning from user feedback. The company’s ASR solution can function either as a speech recognition component within a larger system or as a stand-alone AI-powered transcription tool.

Our products are the torch-bearers into a world where voice-enabled devices not only take commands but also understand natural language and interact humanly


It continuously learns from corrections made by native speakers. New words and dialects are learnt from these corrections, allowing the user to customize the ASR specifically for their use case. Capio’s proprietary GPU-engine makes it possible to leverage thousands of processing-cores to transcribe speech significantly faster than real time, in a fraction of the time of their competitors. “In the past, it would take several months and many engineers to customize a speech recognition system; at Capio, we can do this in a completely automated manner in just few hours,” says Dr. Lane. Mounted on the pillars of machine learning, the solution’s accuracy bears semblance with professional human transcribers. The solution can extract speech data from phone calls and media sources such as videos and audio clips, and put them to best use.

Dr. Lane emphasizes the extensive focus on call centers where ASR can record and transcribe conversations between two or more individuals in real time. These abilities have proved beneficial to Capio’s clients who require solutions for recording and analyzing both in-bound customer support calls and outbound sales calls. Similarly, a broadcast analytics company searching for an economical and apt solution to analyze live broadcast media is using Capio’s offering for round-the-clock transcription of live broadcast media across a wide variety of sources.

Capio’s solution is flexible as it provides both cloud and on-premise options for deployment. The solution’s engine can run across a diversified range of devices—from large servers to laptops to cell phones. Capio is the first provider that enables data center-level accuracy locally on mobile devices.

Marching into the future, Dr. Lane foresees the launch of Capio’s analytics platform that can provide real-time transcription and analysis of inbound and outbound calls. Learning from customer data, the platform could even coach call center agents on the fly. On the geographical front, expansion in Asian countries such as Japan, Korea, Indonesia, and Singapore is expected to help Capio spread its wings globally.