Lighthouse Prism Reduces Review Costs Across Client’s Entire Legal Portfolio

Global pharmaceutical company achieves dramatic cost and time savings by utilizing Prism across their legal portfolio.

Download the case study
34%

Overall Cost Reduction

15K

Reduction in Attorney Review Hours

$1M+

In Review Savings

What They Needed 

A multinational pharmaceutical company sought to substantially decrease their overall ediscovery spend, as well as improve the speed and accuracy of their privilege review. The company was interested in finding ways to reuse prior work product from past legal matters to reduce review costs. 

The company process for identifying potentially privileged documents lacked accuracy, resulting in extensive attorney review of privilege “false positives.” Over 90% of documents flagged for privileged review were ultimately determined to be not privileged. Additionally, the company’s privilege screen, which consisted of a list of search terms and attorney names, was not identifying many privileged documents. 

The company also sought to improve the cross-matter consistency around how the documents were reviewed. They were particularly interested in reducing risk associated with the inadvertent production of sensitive and privileged company data. To solve these challenges, the company was looking to create efficiencies at scale. 

How We Did It 

To reduce cost while increasing review consistency, the Lighthouse analytics team deployed Prism. Prism is a proprietary big data analytics technology that uses AI to aggregate and analyze document data and previous attorney work product from prior legal matters. Prism allows companies to repurpose past work product, including privilege coding, to reduce review time and cost, as well as to improve coding consistency. The Lighthouse team proposed a proof of concept for the company, highlighting how Prism could help them achieve their goals. 

Key data from 22 of the company’s past legal matters were ingested into Prism. This data included duplicate hash values, metadata, document text, production information, and attorney responsiveness and privilege/ redaction coding. Once entered, Prism’s algorithms ‘learned’ from this data to customize its recommendations. The Lighthouse team then applied Prism’s learnings to a separate large review matter to identify possible efficiencies. 

[We] were impressed with the capabilities, particularly with what has already been discovered about our existing data. Having the opportunity to apply these learnings to a client matter would be the ultimate demonstration of value.

Client

The Results 

Prism’s analysis revealed that many documents were being unnecessarily reviewed – in some cases hundreds or thousands of times – across different matters over time. Many of these documents contained keywords commonly used to identify potentially relevant documents, but were themselves non-substantive and never identified as responsive in prior reviews. The analysis also uncovered opportunities to repurpose attorney work product from related cases and reuse coding calls from earlier matters. In total, more than 350K non-substantive and non-relevant documents were removed from review, delivering an estimated $620K in potential document review savings. 

By using multiple inputs and algorithms, Prism was also able to improve the efficiency of privilege review. Prism’s privilege predictive scoring resulted in an additional $412K in privilege review savings, for a total of $1.03 million in savings across the entire review. 

The results of this proof of concept showed dramatic time and cost savings with the additional benefit of improving coding consistency across matters. In all, the Lighthouse team found that by using Prism, the company could have saved more than $1 million in review costs on one matter alone. 

What’s Next 

Following the successful proof of concept, Lighthouse continued to work with the company to develop custom workflows that incorporate Prism’s learnings into the client’s review on new matters. These workflows include the ability to continually validate Prism results via statistical sampling, and continually refine and improve the identification of privileged, responsive, and non-responsive documents in future reviews to further customize the AI for the company. 

With demonstrated cost savings of over $1 million in a single review, along with a reduction of over 15,000 review hours, the company projects exponentially greater savings on future cases, along with the ability to ensure consistent document coding and accountability.