Manual Review is the process of human reviewers painstakingly reviewing documents collected as part of a case, and determining which ones are relevant to the matter at hand. It has long been the most labour-intensive and cost-heavy part of the eDiscovery process.

Managed Review is the service that handles the sourcing, training and monitoring of reviewers and the general managing of a manual review, allowing legal professionals and subject matter experts to spend their limited time on more high-value work. 

As Managed Review is typically so time and cost-heavy, it is often the area that faces the most scrutiny and pressure to adapt and adopt the latest technologies. Here we will be looking at how Managed Review has adapted, over the years, to increasingly more digital data sets, and what emerging technology is in place for the next step in our document review journey. 

A little bit of history:

Prior to the last couple of decades, most managed review projects involved junior members of legal teams sifting through boxes and boxes of hardcopy material, hoping to find that elusive, needle-in-a-haystack document that could make or break a case. The work was complex, repetitive and dull, and a well-known punishment in the legal community. As we slowly began to see a societal shift towards email and faster business communication, it became clear that printing out email chains, storing them in boxes and subjecting the latest unsuspecting paralegal to afternoons of papercuts in windowless file-rooms was no longer the best way to go about things. Hence, eDiscovery was born, and with it, eDiscovery tech!  

At first, there was the introduction of TAR1.0 (Technology Assisted Review 1.0) which allowed users to concept search, use advanced analytics and predictively code their documents. TAR1.0 represented a massive shift in the attitudes of the legal community towards trusting technology to aid us in our work. 

Next came TAR2.0 - Active Learning, the tool most of us think about when we think of "AI in eDiscovery." Active Learning uses a machine learning algorithm to provide reviewers with documents it predicts as 'most likely' to be relevant, based on the reviewers' own coding decisions. Here we see a more complex form of AI, where the algorithm uses human-like reasoning to determine whether or not a document is likely to be relevant based on its content and how other documents with similar content have been coded. For a long time, this has been the industry standard AI tool and has built a tremendous amount of trust with both service providers and legal professionals alike. The reliability of the Active Learning tool, paired with society's shift towards accepting AI in all other areas of life has provided us with the perfect jump-point for incorporating newer, more cutting-edge technology into Managed Review.