Verbit Software Ltd., a provider of automated video and speech transcription services powered by artificial intelligence, is hoping to take on giants such as Google LLC after closing on a new $23 million round of funding led by Viola Ventures.
Vertex Ventures, HV Ventures, Oryzn Capital, Vintage Venture Partners and Clal-Tech also participated in the Series A round, which brings Verbit’s total funding raised to $34 million following an earlier $11 million seed round.
Verbit is fighting for business in an extremely competitive transcription space that’s growing in importance to enterprises for obvious reasons because it saves them huge amounts of time and money. Verbit’s problem is that it faces some serious competitors in Google, Microsoft Corp. and Amazon Web Services Inc., all of which have developed their own automated video and audio transcription capabilities in recent years.
Going up against established competitors of that nature is no easy task, but Verbit offers a key differentiator in that it combines its AI speech recognition software with an on-demand network of human transcribers in order to boost the accuracy of its services. The company makes some bold claims, saying it can transcribe video and audio with 99 percent accuracy when human transcribers are involved.
“Our solution involves an extensive network of freelancers, who edit and review the transcripts that are generated by our AI technology,” Tom Livne, Verbit’s chief executive officer, told SiliconANGLE. “This hybrid approach enables Verbit’s customers to optimize their workflows and meet the increasing demand for transcription services, which is reflected in the company’s high customer retention and renewal rates.”
When Verbit’s speech-to-text technology works alone, it’s able to transcribe one hour of speech in just five minutes, Livne said. But the transcriptions may be less accurate with software alone.
Verbit’s software also allows data in speeches to be analyzed for potential insights. This is because it incorporates natural language processing technology, supported by adapted speech recognition models, Livne added.
“Natural language process querying provides the ability to obtain complete insights on all the data in context, not just isolated segments of it,” Livne explained. “It also helps uncover patterns and trends in the vast amount of data, making it possible to conduct more complex analyses.”
The AI/human combo seems to have some appeal in niche industries, with Verbit claiming more than 100 customers in legal and higher education alone. Its services are said to be very popular in the legal industry to help fill the court reporter talent shortage, for example.
“Transcription is a good example where progress in AI and cheap compute services from the cloud enable a more automated approach,” said Holger Mueller, principal analyst and vice president of Constellation Research Inc. “It helps reduce human error rates and comes with 24/7 availability. That’s the market Verbit is after.”
Fresh with new funding, Verbit says its main aim now is to accelerate its expansion in the U.S. The plan is to grow its sales, marketing and product teams, and funnel more money into the development of new capabilities for its transcription platform.