Legal review of 9 million documents completed in 90 days
A powerful combination of AI-driven technology, data analytics advice and targeted human review focused our legal client’s case-build activities on approximately 55,000 relevant documents from a starting point of 9 million, and resulted in the legal review being completed within 90 days.
In the end, our legal client reviewed less than 1% of the overall document count, having reassured themselves that they could have confidence they were not missing out on anything of significance.
This saved considerable expensive review time and expedited the case assessment stage of the process by at least a year.
Transforming legal review with AI-driven CAL
Our experienced eDiscovery experts helped our client to take advantage of the powerful AI-driven, continuous active learning (CAL) capabilities native within Reveal. We assisted them to formulate their cull and search strategy, in order to narrow down the significant volumes of documents actually requiring review.
This case started with over 9 million documents. Based on a list of pertinent keywords, the number of relevant hits narrowed the list to 1.7 million documents. Starting with conventional linear review in Reveal, the legal team began the process of reviewing and tagging relevant documents from the initial documents with keyword hits. As this review process continued, Reveal’s CAL algorithm started to build a machine learned model based on the initial positive and negative responses. Each remaining document was then ranked for likely relevance using the resultant mode. Reviewers then focused their attention on the documents marked highly relevant by the Reveal CAL algorithm, which (as the name suggests) continued to refine itself based on subsequent human tagging. The relevance ranking of all documents was continually updated And the accuracy of the model and relevancy scoring was tested and reviewed throughout using graphical model reporting tools along with spot checks from experienced human reviewers.
Our client was impressed by the speed with which Reveal was able to focus in on and suggest pertinent documents for review and the sophistication and accuracy of its results, to not only identify the initial keyword hits but also to extend the exercise to other documents ranked by learned relevance.
Legal review team performance and reporting
Reveal’s intuitive dashboard reporting helped to demonstrate progress, as well as batching and tracking documents for review and reporting on the performance of the review team, as well as overall case progression.
Significant reduction in legal review time
Our client was left enthused by this use of AI-powered technology which resulted in them only having to review 55,000 documents. It saved them months of review time and significantly reduced the length of the case build.
Throughout the process, they were reassured that the search and ranking by Reveal’s powerful algorithms were successfully focused on identifying highly relevant documents and they had confidence that they had built their case drawing on all the relevant evidence.
Eager to find out how the latest AI-powered technology could transform your next case?
Get in touch to find out about ways we’re using the latest technology, including GenAI to accelerate investigations and improve litigation support.
