Technology-assisted review (TAR) -- which is also known as computer-assisted review or predictive coding -- uses software to search and sort through documents that are relevant for the purposes of e-discovery.
Based on a series of algorithms, TAR can sort through millions of documents, sometimes terabytes of data, quickly to determine which documents are considered "responsive" and nonresponsive for a legal case. The goal of technology-assisted review is to eliminate the time and costs of hiring lawyers to sift through documents manually – and potentially increase the accuracy of the process along the way.
TAR incorporates machine learning, whereby the software must be "taught" over time what constitutes relevant documents. Humans thus supplement TAR and train the software to identify documents correctly.