A new generation of e-discovery tools is emerging that promises to revolutionize the process of e-discovery review. This new software relies on a process called “predictive coding” in which a sophisticated statistical algorithm identifies the attributes of relevant and irrelevant documents during the review of a small (but statistically significant) subset of documents. Through an iterative training process, a lawyer with expert knowledge of the case reviews the small subset of documents through a series of document ‘batches’ generated by the program. After each batch, based on the lawyer’s designations, the system refines the attributes of relevant versus irrelevant documents until the system determines it will learn no more from continued training. At that point the system, using the relevance attributes determined through the training process, applies a relevance score to each document in the collection. This intelligent e-discovery strategy is an alternative approach that we believe provides clients with the best possible value by delivering enhanced review quality within a well-controlled cost framework. Intelligent e-discovery combines review by experienced trial lawyers with automated analysis and prioritization of potentially relevant documents to deliver reliable and cost-effective results.

Does It Work?

Recently Squire Sanders conducted internal testing of one predictive coding platform called Equivio>RelevanceTM. Equivio>Relevance leverages ground-breaking technology to introduce a higher level of flexibility, control and accuracy into the e-discovery process.

To conduct the test, our firm leveraged a document production that had previously gone through a traditional lawyer-based review in an actual case. After the training and relevance scoring, we compared the software’s designation of relevant/irrelevant to the relevant/irrelevant designations made by the lawyers during the initial review – the results were impressive.

Consistent with other third-party testing of Equivio>Relevance, our internal test revealed that the software and the human reviewers were about equally effective in identifying relevant and irrelevant documents within the collection. These results are also in line with similar testing of other predictive coding software platforms by The Sedona Conference® and others.

More compelling are the potential cost savings that would have been realized if the technology had been available at the time of the initial review. Our attention to cost-saving strategies and the exercise of reasoned judgment to produce reasonable and proportional discovery responses allow us to deliver high-quality results at a cost that is appropriate to the size and nature of the case.

Utilizing existing case law and reasonableness standards we have already developed a robust process and defensibility package. Moreover, there are numerous benefits to the technology that do not raise defensibility concerns. For example, the application is well-suited to facilitate a cost-effective, expeditious review of incoming document productions. Our recent utilization of Equivio>Relevance for this purpose resulted in a client savings of more than US$200,000 off the cost of a traditional review. By organizing the review set according to relevance scores, Equivio>Relevance is a powerful tool for early case assessment, pre-filtering, prioritization of review and potential automatic culling of non-relevant documents. The system quickly identifies highly relevant documents allowing the lead lawyer, early in the process, to better understand potential strengths and weaknesses of the case. Further, the technology allows lawyers and clients to more accurately determine costs associated with the review. Finally, when used in conjunction with other e-discovery best practices, clients are able to use the technology to better assess the potential risks/benefits related to their decisions.