To begin with, we utilised deduplication and email threading as part of the Early Case Assessment (ECA) phase to reduce the number of documents that required analysis.
With many DSARs, the contents of several inboxes usually result in overlapping and duplicate communication. Applying these tools allowed us to drastically reduce the amount of data required for review.
After the amount of data decreased, the remaining documents were subjected to concept clustering. This analytics tool provides a high-level overview of the themes present within a dataset.
The reviewers in this case then reviewed a 30% sample of documents from each cluster, searching for relevant documents. If a relevant document was found within the 30% sample, then the remaining portion of the cluster was reviewed.
The cluster was removed from further review if no relevant documents were found.
Therefore, in principle, there was a 70% reduction in time reviewing certain clusters, meaning fewer hours combing through their data.
By configuring an approach that doesn't depend on the subject matter, the process can be easily replicated, meaning the company can apply it to any DSAR that falls into its lap.