Related to the 2010 Dodd-Frank Wall Street Reform and Consumer Protection Act (commonly referred to as Dodd-Frank), Harvey Spencer Associates (HSA) has announced the launch of a new multi-client study to identify opportunities for the use of voice recognition technologies to reduce the costs associated with compliance discovery of trading records in the banking and financial services industries. Basically, Dodd-Frank requires that traders in stocks and bonds record all communication related to their business of trading. And those records have to be made available for audits for a year, and delivered fairly quickly when asked for. Deadline for complying with this is Dec. 31. Thus, the current opportunity for implementing voice recognition software.
Spencer is known to most of us for his work in the document capture software market. Here’s how he describes his expansion into voice recognition (a move which capture software ISV Nuance actually made several years ago): “It is part of expanding HSA to defining capture as ‘capturing transactional information from unstructured big data elements,'” Spencer stated in an e-mail to DIR. “Defining the inputs as documents; voice; photographs; and video, the intent is to build expertise in these different underpinning technologies which all use pattern based understanding combined with business rules (it’s really forms processing with different inputs) to extract meaningful data.
:In the case of the Dodd-Frank compliance issues, it is a matter of being able to add enough metadata to the trader’s voice recordings in order to reduce the time spent on searching and discovery. As with forms processing in the past, we do not expect that technology can deliver 100%, there will always be manual effort – what we want to do is identify the value of partial recognition/understanding and the supplements to voice recognition that can achieve this.”