How Machine Learning Can Revolutionize Your Document Review Process
The Problem with Manual Document Review
Document review is, by far, the most expensive component of electronic discovery. According to the RAND Corporation, review consumes about 73% of electronic discovery costs. It’s easy to see why. Traditional document review relies on legal teams of attorneys, paralegals and investigators reviewing documents one at a time until a review of the entire document collection is complete.
For instance, let’s say the goal of legal and compliance teams at a global financial firm is to find electronic documents relevant to a securities fraud litigation across the firm over a three year period. The number of electronic documents to be reviewed could be well over a million. Assuming after bulk culling there are 1,000,000 documents to be reviewed, at a manual review rate of 100 items per hour that comes out to 10,000 billing hours or 200-250 reviewer weeks. One can easily see that in this case, the traditional linear review will be extremely expensive to perform in terms of costs, resources and time involved. Moreover, review accuracy rates may vary widely as cognitive processes of humans have well-documented biases.
How Technology Can Help
In cases where a high volume of documents is to be reviewed, Computer Assisted Review can be a savior. Also known as Technology Assisted Review (TAR), it is a process in which an algorithm (predictive coding model) leverages categorization decisions made by an expert reviewer (attorneys, paralegals, and investigators) on a randomly selected subset of documents to make predictions regarding the relevancy of each document in the entire collection.
Technology-Assisted Review (TAR) is a judicially endorsed process and has come a long way since 2012 when US Magistrate Judge Andrew J. Peck issued the first judicial opinion endorsing the use of TAR in Da Silva Moore, et al. v. Publicis Groupe. TAR is now used widely in e-discovery and widely accepted by judges. Because TAR leverages the computing power of high-performance servers, instead of manual review by humans, employing TAR can lower the cost of review in “large-data” cases.
How Proofpoint Accelerates eDiscovery
Organizations that regularly face regulatory or eDiscovery requests, can now utilize Proofpoint Enterprise Archive’s native Computer Assisted Review capabilities for first pass document review on “large-data” cases. With the ability to perform Computer Assisted Review right within the archive, organizations need not export extraneous (non-relevant) documents out to third-party review platforms to perform first pass document review. By automating the task of producing relevant documents ready to be exported to third-party vendors or to the requesting party for further analysis, Proofpoint Enterprise Archive’s native Computer Assisted Review capabilities enable IT, legal and compliance teams to meet court-sanctioned deadlines and reduce the soaring costs of third-party document review.
Learn how to accelerate eDiscovery. Register for the Archive Live Demo Today!