CategoriX Automated Document Review Technology

Dramatically Improve Document Review
In any matter, finding the critical documents is of utmost importance – but the sheer volume of content associated with large-scale document review often results in a trade-off between review speed and accuracy.

With CategoriX, a proprietary automated document classification technology developed by scientists at Xerox Research Centre Europe, there is no trade-off. Leveraging the expertise of attorneys most knowledgeable about the case, CategoriX applies statistical and machine-learning techniques to prioritize, or rank, documents for review based on how likely they are to be relevant to the issue at hand.

Legal teams rely on CategoriX to conduct a wide range of document review tasks quickly and cost-effectively, while optimizing quality – including automated review prioritization, QC enhancement, first-pass review, issue coding and defensible document reduction. With CategoriX, legal teams can:

Zoom in on the key documents and issues that merit attention, leading to earlier case insights and a faster path to production
Quickly find “smoking gun” documents
De-prioritize documents classified as non-relevant, saving time and reducing expenses
Achieve greater accuracy throughout the review with consistent application of expert assessments
Enhance review quality with built-in algorithms that flag coding discrepancies
Optimize workflow by seamlessly integrating CategoriX-ranked documents into reviewers’ assignments on the OmniX™ platform
Rely on an iterative process and measurable results that stand up in court
How Does CategoriX Work?

CategoriX relies on the assessments made by attorneys most knowledgeable about the matter and leverages XLS technical experts to guide the technology and workflow. In a CategoriX review:

Legal and subject matter experts assess samples of documents from the review population
CategoriX models are built based on these coded documents
CategoriX iteratively incorporates feedback from coded and QC’d samples to progressively improve the accuracy of its relevance scoring
Once optimal accuracy is achieved, CategoriX reviews the rest of the document population and ranks each document according to how likely it is to be relevant