Deepgram Raises $47M to Enhance Its AI Speech Recognition Tech
For those of us who can’t type at lightspeed and have indiscernible penmanship, accurately transcribing audio is best left to modern technologies. Deepgram is one such tech solution working to achieve this and more with its artificial intelligence-powered deep learning platform. The company announced on Tuesday a $47 million addition to its latest funding round.
Deepgram’s new injection of capital led by Madrona brings its Series B round up to $72 million. To date, the company’s total equity funding reaches $86 million that it plans to put toward innovating the high-performance speech tech sector.
Whereas natural language processing solutions have historically taken a text-based approach, according to a company release, Deepgram prioritizes audio with the goal of incorporating intonation and inflection into sentiment analysis, speaker labeling, summarization and other services.
The company’s enterprise product is built to facilitate automatic speech recognition made to understand the nuances of human speech. It recently added new features for topic and language detection as well as translation to provide enterprise clients with data that’s easier to transcribe and analyze.
Rather than companies having to task employees with gathering insights from audio data directly, organizations can use Deepgram to analyze audio data and structure it into text and metadata. Additionally, they can use Deepgram’s API to build the solution directly into their tech stack to assist with voice-based customer experiences, according to Scott Stephenson, Deepgram’s co-founder and CEO.
Stephenson and his co-founder arrived at the idea for Deepgram while studying particle physics in college. While situated in an underground bunker, the pair continually recorded themselves and the audio surrounding them.
“In the hours not devoted to research, we life-logged,” Stephenson told Built In via email. “When we tried to go back and find key conversations and specific moments in those audio files, we felt the very real pain of not having a good tool available to help process the recordings and pinpoint valuable timestamps. We decided to build one using the same AI we were using to find dark matter particle events and that was how Deepgram was created.”
As the platform continues to evolve, Deepgram will keep working to let companies analyze their audio data accurately, quickly and at a scalable cost. The company’s new equity funding equips it to expand its research and engineering resources. Deepgram wants to develop solutions for elements of speech such as emotion detection, intent recognition, summarization and more.
The company is also hiring for several roles. Deepgram is currently looking to fill remote positions across teams including engineering, customer success, research and sales.