Salient Logo
Salient Logo

12.5TB of data processed and reviewed within 7 days for regulatory collusion case

Salient uses innovative parallel processing techniques, AI and keyword search to rapidly and accurately surface 2500 responsive documents out of an initial dataset of 12.5TB.

Client Sector: Insurance
Technology Used: AI, eDiscovery, Reveal
Data Processed

The Challenge

Salient was recommended to the client (a large insurance company) by their law firm. They approached us to assist in a regulatory collusion case with large volumes of data (12.5TB) resulting from a wide date range.

The client and their legal counsel hoped to optimise the time- and cost-implications of the large data volumes through intelligent culling in order to remain in-budget and one step ahead of the regulator.


  • Large volumes of data
  • Wide date range
  • Short turnaround time


  • 6x priority custodians dealt with in 7 days
  • 15 million documents processed into Reveal
  • Reduced to 7 million documents within a week
  • 2500 responsive documents identified
  • Further investigation is ongoing
Our Solution

Salient ran the data through pre-processing to deduplicate and de-NIST it, applying filters to move it onwards to processing in the eDiscovery platform.

AI was then utilised for early case assessment (ECA), along with keyword search to surface responsive documents.

The investigation is ongoing and Salient continues to assist the client by using Reveal’s AI capabilities to gain insights through the identification of cluster wheels, similar concepts, entities and players, and interactions outside the organisation.

Parallel pre-processing, processing and review

From data collection to review, Salient successfully dealt with six priority custodians in 7 days, reducing 15 million documents to just 7 million documents within a week of ingestion into Reveal.

AI for ECA and keyword search

Keyword search surfaced 2500 responsive documents, enabling the client to achieve their goal of fast, cost-effective preliminary insight into the suspected regulatory collusion matter.