eMerge: Latest News and Developments — May 2024
May 1 2024
Our eMerge team is excited to share the following updates:
- Custom Solutions Work — Class Action Employment Data Analysis
- aiR for Review Case Study — Generative AI in Action!
- Data Collections Update — Client Ticketing System Collections Success
- Annual eDiscovery Updates Webinar — A Client Favorite
Custom Solutions Work
Class Action Employment Data Analysis
Our custom solutions team has enabled our clients and litigators to gain insights into employment data to assist in the defense of class action claims.
In one sample matter, our team
- Standardized leave data from a multitude of diverse and disparate record-keeping systems, each with its own unique format and structure.
- Compiled and rectified a dataset of 12 million timecard entries and three million leave records, a process that required more than 4,000 lines of code.
- Conducted a comprehensive analysis examining specific class members, providing statistical samples, and comparing employee groups.
- Critically evaluated opposing expert testimony and developed customizable dashboards for ongoing analysis by the case team.
Our efforts significantly refined the class action, narrowing its scope to a specific employee type.
“eMerge has been instrumental in the collection, review, analysis, and production of large data points from various client data systems. They have helped us understand the patterns and inconsistencies in the data, which has helped us frame factual and legal issues throughout discovery and briefing.”
–Bianca DiBella, Troutman Pepper Associate
aiR for Review Case Study
Generative AI in Action!
eMerge has continued to expand its use of Relativity aiR for Review as an early access provider of the generative artificial intelligence (AI)-based document review solution. We have utilized aiR to make coding decisions in a variety of projects, including review of opposing document productions and for quality control of review attorney decisions. Additionally, in advance of recent document production in a municipal real estate and construction litigation, eMerge employed aiR for Review to make first-level coding decisions for responsiveness and to identify key documents from a set of 5,000 documents that eMerge narrowed from a larger population via traditional search and culling strategies.
Compared to traditional document review, aiR for Review reduced review costs by 40% and time to completion by more than 50%.
- The entire first-level review (including quality control but excluding privilege review), including manual review of a limited number of documents outside the aiR for Review set totaled roughly 40 billable hours completed over roughly 5 working days. (The aiR-specific prompt generation, testing, and validation totaled just 8.7 billable hours, plus ~3 hours of technical time for aiR to run on the population).
- A manual review of the same population through first-level and quality control coding (and still excluding final privilege review) would have required between 140-170 billable hours and likely would have taken 10-15 working days to complete.
After an iterative process of testing to develop and refine the aiR prompt, accuracy results were perfect on documents identified as very responsive, responsive, and nonresponsive. The “borderline responsive” documents were easily resolved using the substantive rationales generated by aiR supporting each coding decision.
Based on these results on a relatively small review population, we anticipate even greater time and cost savings on larger review sets with similar accuracy performance.
“We were very happy with the results. Use of aiR for Review helped us meet a tight discovery deadline without sacrificing accuracy. We look forward to using it for additional matters.”
–Victoria Alvarez, Troutman Pepper Partner
Data Collections Update
Client Ticketing System Collection Success
Our Data Collection team recently collected structured data from Zendesk, a well-known support ticketing platform. Despite challenges with the format of the mass exports, our team successfully resolved them by creating custom workflows to analyze, parse, and generate documents. The results included all tickets, comments, attachments, and associated metadata. This process allowed us to quickly load and produce the data in our review platform. The project was successful and exemplifies our team’s ability to find innovative solutions to collect from constantly expanding data sources.
Annual eDiscovery Updates Webinar
A Client Favorite
Our team of practicing attorneys and technologists recently presented a webinar about the key decisions and technical innovations impacting discovery in 2023 and highlighted trends for 2024. They discussed the practical considerations of balancing legal requirements with rapidly changing AI and other technologies, how ESI protocols may be both a blessing and a curse for litigants, and the continued emphasis on collaboration sites and short messaging applications by regulators and litigants. To view the webinar recording, click here.