The quarterly publication of the International Legal Technology Association
Issue link: https://epubs.iltanet.org/i/67910
mix is recommended for optimal results — one employing a combination of technological and human input. In other words, companies would be wise to embrace TAR, but do so with the right mix of technology, human expertise and the mindset that it will be necessary to cooperatively plan TAR usage with the opposition. Know What TAR Is and How It Works Many vendors now offer TAR solutions — whether they're referred to as predictive coding, meaning-based coding, adaptive coding, predictive priority, transparent predictive coding or automated document review — each implements TAR in a different way. Regardless of how each offering is packaged, the focus of the technology hinges on the process or method used by each to enable a quicker legal review process that costs less and operates intelligently. There are three primary methods for carrying out TAR: • Rules-based • Facet-driven • Propagation-based The rules-based approach relies on the creation of a user- defined criteria set (or set of rules) to build the equivalent of a coding manual. The set of rules is then loaded into the review tool where it's applied to the entire document set. In the facet-driven scenario, the user relies on a tool to analyze documents for potentially similar items or items of interest (or facets), and then groups them so reviewers can make decisions about them. Often, visual analytics are employed to provide reviewers with an easy vehicle for getting to the prioritized documents. The propagation-based mechanism entails the dissemination (or propagation) of known traits of a sample set of documents to the remaining body of documents in a set. This approach is often referred to as predictive coding, in that it involves some level of a machine learning to create an overarching mathematical model to subsequently predict what is or is not relevant for the remaining document set. The Right Mix of People and Technology While the use of new technologies is taking on an increasingly dominant role in streamlining the review process, technology alone is not enough. In order to effectively use the software, organizations need to ensure they have trained staff in place who are able to support the technology while managing the TAR workflow. While software is an obviously important component of TAR, having experts in place to program the software with the correct set(s) of rules is critical for the system to be able to intelligently identify and tag documents. Despite some suggestions that TAR can or will remove lawyers from the review process, nothing could be further from the truth. Legal review teams will continue to play a central role as they set the framework for how documents are handled. The quality and consistency of the decisions legal teams make in this regard directly affect the tool's ability to apply that framework to the entire document set. With any defensible process, quality control checks must be in place. Auditors will play a key role in the defensibility and acceptability of TAR, as they will inspect the statistics behind the quality of output from TAR versus the input specifications defined by the legal team(s). Have you used or are you currently using predictive coding? Yes 36.8% No Not sure 0% 4.3% 20% 40% Source: eDJ Technology-Assisted Review/Predictive Coding Survey, March 2012, N=117 60% 80% 100% 59% Peer to Peer 81