Harvey Spencer Associates, Inc., announced last week the launch of a new study aimed at identifying opportunities for the use of voice recognition biometric technologies to cut back on compliance costs under Dodd-Frank.
The 2010 Dodd-Frank Act requires major swaps and derivatives traders to implement policies and procedures that allow for the monitoring of individual traders to reduce “unauthorized risk” and the recording of all oral communications concerning an actual or potential trade to be kept on file for one year.
Records must be readily available for research and compliance checks, and regulators expect trading institutions to respond expeditiously to requests for information—generally within 24 hours. Entities have until Dec. 31 to comply with the oral communication records requirement.
An individual trader could potentially be involved in hundreds of calls per day, and current voice recognition software does not have the ability to comply with the regulatory discovery requirement because of the recordings’ complexities. The CFTC said hiring a compliance specialist to translate voice to text could cost a firm as much as $779,800 per event.
“The technological state of voice recognition is similar to that of Optical Character Recognition 10 years ago,” Harvey Spencer said. “Voice recognition software currently works very well for command and control, and it works reasonably well for converting voice to text. We believe that correctly configured voice recognition, combined with other software tools, can improve the efficiencies of identifying voice transactional records and thus significantly reduce the cost of Dodd-Frank compliance.”