iDS’s client needed to prepare document binders to support a series of interviews that were scheduled to be conducted over the course of one week.
Our client was conducting an internal investigation to determine if employees negligently or knowingly made misrepresentations to regulatory authorities. As part of their investigation, our client scheduled interviews with four employees of interest.
iDS was tasked with:
- Isolating a subset of documents within the larger dataset for targeted review
- Promoting only the most likely to be key documents for eyes-on-review
- Exporting and delivering the key documents to be included in document binders
We leveraged several components of LeanReview™ in order to create a seperate workflow dedicated to identifying key documents that would go into the binders and support the interview process.
The iDS Discovery Services Analytics team used LeanReview™ to deploy a separate workflow that functioned in parellel with the larger review. The workflow was designed to accomplish two major goals:
- Limit the reviewable population through cluster analysis based on exemplar documents.
- Rapidly train a predictive classifier using an existing model and documents selected by a machine learning algorithm.
LeanReview™ isolated a small subset of documents for immediate review enabling the reviewers to focus on only the most relevant content. The work occurred independent of the larger review to ensure minimal interruption to the overall workflow.
LeanReview™ was able to prioritize a set of 2,225 documents for immediate review out of a corpus of over 3MM documents. Within the subset of documents promoted to eyes-on-review, over 80% of the documents were determined to be responsive to the investigation and 55% of those documents were determined to be highly relevant to the specific issues at hand. The final deliverable was a set of 110 key documents that were organized into binders and used by our client to support the employee interviews.