Kurt Leafstrand, Symantec: Clearwell eDiscovery Platform To Offer Transparent Predictive Coding To Enable Rapid and Defensible Automated Document Review

January 25, 2012

Symantec plans to introduce Transparent Predictive Coding to its Clearwell eDiscovery Platform in order to help organizations defensibly reduce the time and cost of document review. With Transparent Predictive Coding, Symantec will help organizations accelerate the review process by leveraging the intelligence of expert reviewers to train the software so that it can automatically predict document relevance in order to streamline document review and increase accuracy. The net result is expected to be a more defensible automated document review process at significantly reduced cost.

"Since document review is usually the most expensive and time-consuming step in e-discovery, an efficient, legally defensible review process is critical to cost control," said Ralph Losey, partner and national e-Discovery counsel, Jackson Lewis, LLP. "Our method for controlling the costs and quality of productions uses estimates, the legal doctrine of Proportionality, and the latest technology assisted review methodologies – including review using discussion threads, near duplication, concept clustering, and my favorite, predictive coding. The combination of new legal methods and the latest technologies delivers tremendous savings to the review process.”

“The stakes are high when it comes to eDiscovery requests. It is vital for businesses to be able to use advanced technology to support their needs, and also to be able to understand how the technology works so they can defend their process,” notes Kurt Leafstrand, director of product management, Symantec Corp. “With Transparent Predictive Coding, Symantec will take the mystery out of predictive coding by providing a new level of visibility into the reasoning behind document decisions, and by documenting the entire process This can help organizations defensibly and significantly reduce the time, risk and cost of eDiscovery review.”

Benefits of Transparent Predictive Coding

Predictive coding helps organizations reduce the time and cost of document review by using machine learning technology to leverage the intelligence of expert reviewers. As reviewers tag documents in a training set, the software identifies tagging criteria common across those documents. This enables the software to “predict” the reviewer’s coding across all case documents, requiring fewer documents to be reviewed manually and thereby accelerating the review process.

Now, with Transparent Predictive Coding, Symantec’s Clearwell eDiscovery Platform is expected to open the black box of predictive coding by providing unprecedented visibility into this training and prediction process and delivering context for more informed decision-making. Reviewers will have complete visibility into how predictions are generated in order to make consistent review decisions and improve the accuracy and defensibility of the document review process.

Symantec’s Transparent Predictive Coding is expected to further accelerate the review process by providing an intuitive, flexible workflow that maps to the specific requirements of the case at hand. This unique workflow is automatically documented by comprehensive reporting tools which can allow legal teams to demonstrate process integrity to the court, opposing counsel, or government regulators. As a result, Transparent Predictive Coding is expected to provide the level of defensibility necessary to enable legal teams to adopt predictive coding as a mainstream technology for eDiscovery review.

Key Transparent Predictive Coding Features:

• Smart Sampling: Offers sophisticated analytics by criteria such as custodian, discussion, concept and participant to ensure the selection of highly relevant training samples. Intelligent training samples result in greater prediction accuracy.
• Smart Tagging: Allows reviewers to highlight the metadata and content relevant to a tagging decision for more granular training, improving the accuracy of predictions and reducing the number of training cycles.
• Prediction Insight: Automatically provides a prediction probability score for the document under review and highlights content and metadata relevant to the prediction. Users can quickly assess how the prediction was generated to improve prediction accuracy and make consistent review decisions, bolstering defensibility.
• Prediction Workflow Management: Provides a review management console with step-by-step guidance that automates the predictive coding workflow, allowing users to begin utilizing predictive coding immediately while achieving optimal and defensible results.
• Prediction Templates: Enables users to import and export prediction models to leverage predictive intelligence across cases. The use of templates streamlines the review of cross-mater issues, resulting in reduced cost and risk.
• Prediction Analytics: Delivers a set of interactive charts and reports that enable reviewers to measure prediction accuracy and analyze documents by probability score, resulting in more informed review decisions.
• Review Quality Control: Provides a comprehensive quality assurance workflow for linear and non-linear review, enabling users to perform statistically valid sampling, identify inconsistent tagging, view disagreements between reviewers, and automatically compare predictions and human decisions to assess and improve review accuracy.