The Challenge: A mid-tier London law firm was managing a matter involving a data set of...
A top-tier Canadian litigation boutique firm needed a scalable, defensible review approach for a complex class action involving 4.9M+ collected documents across 20 custodians and a 750K+ document review population.
CHALLENGE
Complex class action with 4.9M+ collected documents, a 750K+ document review population, and a tight deadline.
SOLUTION
Combined advanced GenAI (Relativity aiR), expert lawyer oversight, and validated defensibility.
OUTCOME
Delivered under budget with validated, defensible results.
By combining senior review lawyer expertise with a custom review workflow incorporating Relativity aiR (GenAI-based review tool), Array delivered scalable review efficiencies while maintaining quality, defensibility, and transparency.
Applied Relativity aiR and advanced analytics to efficiently manage a 750K+ document review population at scale
Achieved 99.5% recall through a defensible validation framework and targeted quality control
Improved AI classification performance through prompt refinement, expert oversight, and statistical sampling
Maintained a 1.1% elusion rate, demonstrating exceptionally low rates of potentially missed relevant documents
Delivered the primary review workflow under budget while maintaining quality and transparency throughout the matter
Array, a Relativity Gold Partner, has long been at the forefront of innovation within the Relativity ecosystem. With a team of over 57 RelativityOne-certified professionals holding more than 113 certifications, we bring deep expertise to every engagement. As an early adopter of Relativity aiR, our team has built custom generative AI workflows to reduce friction and deliver better outcomes for clients.
The Challenge: A mid-tier London law firm was managing a matter involving a data set of...
The Challenge: A national transportation company faced high-volume litigation with thousands of...
The Challenge: To respond to a Competition Bureau Supplementary Information Request (“SIR”)...