Source: GreenKey Technologies
GreenKey Technologies, creator of an award-winning, team-based, voice collaboration tool for financial and regulated markets, announced today that the U.S. Patent and Trademark Office has granted the company a second patent for its latest innovation: speech-to-text for the financial markets, incorporating custom processes for language model creation; model training; error correction, and accelerated transcription to allow for fast, accurate transcription of domain-specific conversations and instructions.
The technology was designed to help firms comply with regulations such as MiFID II, which requires the capture of all voice orders and quotes intended to lead to an execution, even if they do not result in a trade. With GreenKey’s patented speech-to-text solution called GreenKey Scribe, orders, quotes and trades can now be extracted directly from real-time conversations taking place on any platform and fed into any execution, analytics or compliance system. The tool was designed to function and improve learning when interacting with very nuanced, “tribal” dialects, such as those that exist within voice-executed markets. Partnering with market participants across all major asset classes has given GreenKey access to thousands of hours of actual sales, dealer and broker conversations, resulting in accuracy rates above 95 percent - up to twice as high as ubiquitous offerings from Google and Microsoft, or other capital markets specific solutions.
Anthony Tassone, GreenKey Technologies Chief Product Officer, said: "GreenKey is the market leader in innovation, and our investments in artificial intelligence (AI) represent the bleeding edge of voice technology. Our world-class data science team has built everything in-house, as customizable containers which enable our large bank, broker and trading customers to leverage the security of an on-premise deployment with the flexibility to create their own language models, all while supported by a data team of PhDs.”
Tejas Shastry, Ph.D., GreenKey Vice President of Data Science, said: “We are absolutely thrilled to have been awarded a patent for our unique training, prediction and deployment methods. We are working with the largest financial firms in the world, and they’re starting to recognize the power that AI-driven speech recognition has to transform many of their workflows and business units. Beyond just regulatory compliance, we’re now partnering with customers on cutting-edge concepts around speaker recognition, real-time call transcripts and customer relationship management (CRM) integrations. It is extremely rewarding to be able to help them extract powerful, actionable data from their voice conversations.”