Digital White Papers

TECHNOLOGY-ASSISTED REVIEW 2.0 Technology-assisted review (TAR), also known as predictive coding, is being used more commonly to help case teams bypass or modify linear review processes, to prioritize their review by relevance, and to cut costs and resource requirements while still meeting deadlines and production requirements in a defensible process. TAR can be a powerful tool when applied correctly. Surrounding the technology with the right workflows, processes and quality control procedures is essential to ensuring quality output. WATCH A RECORDING OF EPIQ'S "TECHNOLOGYASSISTED REVIEW 102: BEYOND THE BASICS" WEBINAR HOW TAR WORKS Although different vendors conduct technologyassisted review with subtle variations, the steps of TAR are the same. The first step is to get a representative sample of your set of electronic documents. This "seed set" can be created in different ways, but the most common is to use a random sampling of the full set of documents. Search terms can be used to help create the seed set; however, this method should be employed sparingly. The next step involves having an attorney or team of attorneys who are knowledgeable about the subject matter of the case review the seed set and code each document for responsiveness and for any other attributes relevant to the case. Based on this initial coding, the TAR tool will begin to learn how to score documents for relevance. The TAR tool will use this knowledge to create a new smaller set of documents for a senior attorney to review for responsiveness. The results of that review will be put into the tool to allow it to continue perfecting its algorithm. Via an iterative process, the tool will keep feeding additional sets of documents to the attorney(s) for review. In doing this, the attorney(s) will "train" the algorithm by evaluating where their decisions differ from the computer's and making appropriate adjustments. This process will be repeated until the attorney is satisfied the tool is pulling responsive documents efficiently. After each iteration, the tool will provide statistical summaries of how successful the algorithm is in focusing on and ranking the responsive documents. The attorney team will work with the vendor using these statistics and their experience going through the document sets to determine when the algorithm has been perfected. Although TAR is a powerful tool, it does not replace attorney review. TAR prioritizes the documents

• Cover