JND Saves Corporate Clients Hundreds of Review Hours and Cuts Costs by Nearly 60% with Relativity aiR for Review
Customer Since
2011
How did they do it?
- Established a repeatable process to confidently use aiR for Review across hundreds of thousands of documents.
- Saved over 750 hours and $85,000 on a single corporate client matter with aiR for Review.
- Developed tailored prompt criteria for a large antitrust matter, achieving an average recall of over 96% and precision of 71% across multiple analyses.
For years, legal service provider JND eDiscovery has been at the forefront of AI innovation, implementing the latest and greatest technology to better serve their clients.
When given the opportunity to join Relativity aiR for Review’s limited availability program, they jumped on board. Their first order of business: develop new generative-AI processes to deliver fast, accurate results.
Establishing a powerful, repeatable process
aiR for Review uses generative AI to simulate the actions of a human reviewer, finding and describing relevant documents according to the review inputs that the user provides.
To streamline this process, JND developed a repeatable approach for quickly iterating on prompt criteria and ensuring aiR’s accuracy and scalability. Here’s how it goes:
Through this thoughtful approach, JND effectively demonstrates to clients how aiR for Review delivers accurate results in record time and can be defensibly used on projects with hundreds of thousands of documents.
Case 1:
750+
hours saved in an employee dispute
JND’s corporate client needed help responding to a production request related to an employee dispute. They started by using search terms and categorization to reduce the data set from approximately 250,000 documents down to around 38,000 documents.
Next, JND used aiR for Review to identify documents for production and to locate key material for their case. Following the process outlined above, they worked with the client’s attorney to draft prompt criteria and validate document samples before running aiR on the full 38,000-document set.
The entire process, including prompt criteria iteration, sampling review, and the senior attorney’s review of the identified relevant documents, took less than 70 hours.
By leveraging aiR for Review, JND was able to save more than 750 hours compared to linear review, reducing total project time by nearly 90% and cutting their client’s costs by nearly 60%. They completed the entire project in less than two weeks, versus the 15+ weeks it would have taken to look at the documents manually.
Case 2:
Over 96%
average recall achieved in a large antitrust litigation
JND was tasked with sorting through millions of documents in a large class-action case that involved multiple corporate entities. They knew their corporate client would benefit substantially by using aiR for Review’s generative AI to meet their production obligations.
JND worked with the client to develop prompt criteria that responded to the production request, adjusting inputs and validating results for three specific companies involved in the matter. This tailored approach ensured that aiR for Review’s analysis was appropriately targeted to deliver optimal output.
As a result, aiR for Review achieved an average of over 96% recall and 71% precision on each of the document sets analyzed, delivering remarkably accurate results that greatly surpassed traditional review.
JND is just getting started
With these thoughtful processes and demonstratable results under their belt, JND is ready for the new era of AI-powered innovation. By truly embracing aiR for Review’s generative AI technology and exploring how to integrate it across their organization, the team is ready to dive in and deliver incredibly accurate, repeatable results to efficiently solve their client’s most complex document review challenges.
"Through extensive testing and experience on live matters, JND has developed hardened processes for Relativity aiR for Review. We want our clients to feel comfortable and confident using the most innovative technologies on the market. Defensibility is our number one goal."
Ben Sexton, VP of eDiscovery and Analytics, JND eDiscovery