Managing audio data files
Discovery Insights: Five questions
Many organizations, particularly those in the financial services industry, have massive and rapidly growing data stores including recorded phone calls and other audio files that are essentially unsearchable. It is important for CIOs and CLOs to understand how they can mitigate compliance risk, while achieving cost savings.
An interview with Sean Riley, principal, Paul Yackinous, specialist leader, and Joshua Uhl, senior manager—Deloitte Transactions and Business Analytics LLP, Deloitte Consulting LLP and Deloitte & Touche LLP.
Regulatory and legal requirements have led organizations to retain large volumes of audio files, primarily voice call recordings. Recently, the Dodd-Frank Act has led to the creation of Commodity Futures Trading Commission (CFTC) requirements for the recordkeeping of swaps trading records (Regulations 23.201-203). Organizations are now required to have the ability to quickly and accurately reconstruct trades, including related pre- and post-trade communications. Unfortunately, most in-place Voice Recording Systems (VRS) do not include technology to match recordings with specific trades. Further, most technologies do not include the ability to index the calls. As a result, many organizations use manual methods to identify specific recordings associated with a trade—at great cost and effort.
Further, many organizations store the recordings in the systems that created the recordings, making it more difficult to support a single source search for eDiscovery collection.
What is driving the management of audio data files?
Most traditional data files like emails and word processing documents can be easily indexed to make them searchable using keywords. Unfortunately, most Voice Recording Systems (VRS) do not support indexing or transcribing audio files in a manner that makes them searchable. Compounding the problem are the myriad languages, dialects and accents that comprise voice recordings.
In addition, many VRS store files in proprietary formats that require expensive vendor-supplied plug-ins to transform into standard formats like mp4 or .wav
Why is managing audio files so complex vs. "traditional" data files?
Many companies do not currently have the technology to automatically index or transcribe audio files. As a result, these companies manually review files for responsiveness and manually tag them for production based on crude date range and device information that may be available. This process is time consuming, costly and subjective.
How are companies addressing the management of audio files now?
Emerging technologies including speech analytics and advance phonetic indexing are being coupled with traditional file archiving tools to provide a more seamless and automated environment from which companies can search for and access audio files. Unlike comparable technologies for typical data files, these technologies are currently immature and their accuracy requires improvement. Even in cases where these technologies are in use, companies often choose to include manual quality assurance testing when collecting audio files.
What technologies exist to help organizations with this challenge?
Companies should consider beginning the process of reviewing their Voice Recording Systems, consolidating where possible to vendors that support a current or future solution that includes automated indexing or transcription of audio files. Companies should consider evaluating vendor solutions as they become available. Companies should also consider disposing of low-risk, low-value audio recordings that are not subject to a statutory or legal retention requirement and imposing a retention period on audio files moving forward that meets the regulatory requirements.
What do companies need to do now?
Managing audio data is much bigger than simply identifying and gaining visibility to the files. The defensible disposition of low-risk audio files also presents an opportunity to reduce costs and risks, while improving the efficiency of responding to regulatory requests. In addition, the practice of continually culling and managing audio archives will likely be critical to maintaining compliance in the current regulatory environment. The benefits of a continuous monitoring system can be particularly effective for companies that are generating audio files from global offices.