Relativity details Data Breach Response for breach review

Relativity’s Data Breach Response is a generative AI product for identifying, extracting, and linking Personal Information and Personal Health Information to relevant individuals or organizations. The product is generally available as of 02/26/2026 and is delivered as a Relativity application for installation in a Relativity aiR Review workspace running in a Relativity aiR production environment.

The workflow starts with documents selected through a Saved Search and ends with a notification report listing potentially impacted individuals and associated PI or PHI. Users configure detectors, run data analysis, review extracted information on project dashboards, resolve conflicts in Entity Analysis, and finalize the output. Each workspace can support up to 300M PI annotations and 150M entities across 1 TB of native data, with quotas tracked in the Ingestion tab.

Data Breach Response handles unstructured documents such as emails and word processing files by using document context, while structured documents such as spreadsheets and CSV files are analyzed through table boundaries, headers, and column values. Relativity warns that outputs require human review because AI may omit information, misclassify data, flag documents incorrectly, or fabricate PI. Azure OpenAI does not retain analyzed document data or use it to train models.