AI is changing how legal teams approach document review, but not every AI tool solves the same problem. While some capabilities help reviewers categorize documents or ask questions of their data, others can take a more targeted approach, extracting specific information from documents based on the needs of a matter.
In Episode 19 of 5 in 5 with CDS and Relativity, CDS Senior Consultant Michelle Broxman and Client Solutions Director Nicole Guyer continue their conversation about AI in legal technology, taking a closer look at Custom Analyses in Relativity aiR for Review. Drawing on their experience working with clients, they explore how this capability helps legal teams uncover information that can be difficult to find through traditional search methods and why context matters when working with complex data.
From Finding Documents to Extracting Information
Traditional searches have their place, but they can struggle when information requires interpretation rather than a straightforward keyword match. Dates, numbers, and details embedded in images or other non-searchable content can be particularly challenging to identify accurately.
Custom Analyses allows teams to define the information they need using plain-language prompts and extract relevant details from documents. In the video, Nicole shares an example from a construction litigation matter involving photos collected from a custodian’s phone. By using Custom Analyses to identify images containing tractors, the team could focus on potentially relevant evidence without unnecessarily reviewing personal photos.
She also describes using Custom Analyses to extract personally identifiable information (PII) and protected health information (PHI) from medical records. The ability to distinguish specific details, such as a patient’s date of birth without the year, demonstrates the value of contextual understanding. Rather than capturing every date or address, the analysis can focus on the information that meets the criteria defined for the matter.
Applying Custom Analyses to Real-World Challenges
The potential applications extend well beyond forensic investigations and medical records. Nicole discusses using Custom Analyses to extract check-in and checkout times from timesheets, calculate shift durations, and identify shifts that may have resulted in overtime.
These examples highlight how legal teams can use AI to address information extraction challenges across investigations, litigation, and incident response. They also reinforce an important point: effective AI workflows begin with understanding the problem and defining what information matters.
Watch the full conversation with Michelle and Nicole below to hear more about their experiences with Custom Analyses and how they see this technology evolving to address increasingly complex legal data challenges.
