A legal team and counsel were getting access to a rolling set of data. The team were using the default Hot, Warm, and Cold ratings to mark documents of interest. Adio received an additional set of data that was highly responsive to pre-defined search terms.

Action

Adio reviewed the progress of Everlaw’s automated Continuous Active Learning (CAL) prediction models. The models showed that the reviewed documents provided strong signal to predict which unreviewed documents had attributes like the relevant and irrelevant documents. This provided the client options to prioritise the document review and focus on those unreviewed documents with relevant attributes.

Result

The client leveraged the predictions to prioritise their document review. The impressed and time-poor client then decided to use Everlaw’s Rigorous Prediction Training methodology to limit the review of documents based on their assessment of precision and recall. Adio and the client defined a CAL methodology and validated the results to reduce the document set for review by 80% and meet a tight production timeline.

You can read more about using Everlaw’s predictive coding at: Leveraging Predictive Coding for Prioritization and Quality Control – Knowledge Base (everlaw.com)