For an important matter, the client** and its legal team had a collection of 1.5 million documents that needed to be collected, preserved, reviewed, and produced for deposition preparation. The core challenge was to comprehensively assess this large volume of documents quickly while adhering to a set project budget.
The eDiscovery team jumped right in to develop a Continuous Active Learning (CAL) model configured specifically for the review. This model was then tested on a random set of documents across the entire collection. The findings were used to further refine and train the CAL model for a full run against the entire data set of 1.5 million documents.
The result was a reduction of 1.1 million documents (73%) from the final data set which delivered nearly $800,000 in cost savings compared to traditional document review methods. The overall efficiency of utilizing a CAL model for this discovery project allowed the legal team to meet its tight timeline with greater accuracy at a significantly reduced client cost.