Multi-Language DSAR Review Across Documents and Audio

A data subject's request required disclosure of both documents and call recordings across Russian, English, and Polish. The project involved an initial dataset of 144,379 documents and 171 audio files totaling 90 hours of calls. The key challenge was optimizing the review and redaction of multi-language audio files and documents at scale.

• Early Case Assessment: Morae conducted an initial analysis of the audio files to assess quality, language, length, and content. The team identified duplicative calls using call metadata, excluding almost 35 hours of audio. Using Morae's scoping workflow, the team provided justifications for the exclusion of 90% of documents — reducing the review population from 144,379 to 13,551 documents following client approval and validation.

• Multi-Language Review: Morae's internal staffing team resourced a multilingual review team within 24 hours of client sign-off. Documents were assigned to the appropriate language resources using Relativity's text-based language detection analytics.

   

• Efficient Audio Review: Leveraging Intelligent Voice™ and Reveal Translation, Morae optimized audio workflows by translating and extracting text from audio files. This allowed limited use of keywords and highlighting to prioritize calls requiring additional assessment by the client's legal team.
• Redaction Utility: Morae's redaction workflow used the latest automation tools available in eDiscovery. A majority of Excel native redactions were performed through automated redactions, freeing the review team to focus on final substantive QC checks. Intelligent Voice™'s integrated audio redaction feature allowed all audio redactions to be performed and approved by stakeholders within the secure Relativity platform.

By applying efficient data culling techniques and optimized workflows, Morae significantly reduced the review population. Of the original dataset, 13,551 documents and 123 audio files (55 hours) were reviewed, with 1,377 documents and 41 audio files ultimately disclosed to the data subject.

The final review cost came in considerably lower than the initial estimate. The project met both the client's financial and regulatory requirements, delivering timely, cost-effective, and thoroughly checked disclosure.