Multi-language review facilitated with the help of AI

Epiq uses Reveal’s language identification technology on a multi-language document review matter – enabling them to segregate documents for simultaneous English and Japanese review workstreams.

Challenges

• Documents contained one of two languages and sometimes both; for example, an email in Japanese with English attachments.
• The case team wanted to use separate review teams for each language

Solution

The Epiq team used Reveal’s language identification technology to quickly segregate documents based on language, identifying 25% of the documents as predominantly Japanese. The team matched documents with reviewers by language to expedite the review.

The team then augmented the review process using Reveal’s active learning technology to identify highly relevant documents and, after quality control processes, promote these document families to outside counsel for second-level review.

Results

Using AI engines to group the document collection based on predominant language was an effective method to separate documents for language-specific reviewers. Epiq assigned relevant data to the appropriate teams, and reviewers started the process sooner.

AI-powered review eliminated the need for an eyes-on review of more than 300,000 documents and resulted in savings to Epiq’s client of approximately $560,000.