The Challenge

A midsize U.S. law firm with a multidisciplinary practice focused on litigation and government relations — sought a cost-effective solution to review and reduce nearly 800,000 email messages collected from a client into a relevant document production. The firm’s goal was to reduce both the time and expense it would have incurred during a linear review of the documents.

They needed an innovative litigation managed services partner with a Relativity environment and supporting infrastructure; robust data protection posture, policies, and procedures; a strong support organization; and the proven ability to pare eDiscovery costs through analytics.

The Solution

Legility worked with several parties to obtain data for this project – including performing collections directly from an international corporation’s O365 environment.

The team started out with 784,000 documents collected from the firm’s client. After eliminating duplicate copies through the Venio software platform, 691,000 documents remained. A date range filter and search terms were applied, and ~99,000 documents were loaded to Relativity.

Legility's Proprietary Technology Solution
Finally, we used Envize™ — Legility’s machine-learning data analytics framework — to sift through the remaining 99,000 documents.

Legility Envize™ played a central role in executing the solution. The framework provided a crucial level of process flexibility that allowed the team to plan a workflow that gave the firm the necessary familiarity with the document collection and defensibly delivered the review and production within the necessary time and budgetary constraints.

Legility accomplished those goals by creating a combined Continuous Active Learning (CAL) and Simple Active Learning (SAL) workflow. Envize™’s ability to construct workflows with this level of flexibility are unparalleled by any other predictive analytics technology.

The workflow:

• Eliminated
  73,000 (73%) out of 99,000 documents as irrelevant. 
• Identified
  26,000 (27%) for potential production.
• Reduced
  the number of documents needing review to only 2,400 (~2.4% of the review population).

Additionally, 1,600 documents were prioritized for review using Envize™’s CAL workflow and another 762 documents using SAL training. Another 900 documents were reviewed in quality control and validation samples.

Overall, our team helped the firm achieve an estimated 94.4% recall and a minimum recall of 90% (with a 96% probability) in the validated and defensible document production.

Results

Thanks to the Legility Envize™ process designed for the project, the client gathered the information it needed from the document review and delivered the document production on time and under budget. Moreover, the solution has helped position the firm to thrive in a competitive litigation landscape.