Purpose Legal

A law firm that took over as new counsel on a matter was denied an extension of the production deadline, leaving them to review 300,000+ documents in a single week or face sanctions. There was neither time nor budget to ramp up an associate or contract review team. The production also required RFP correlation and issue-level tagging for every produced document, making a traditional Active Learning / TAR 2.0 workflow—which would have taken multiple weeks and thousands of review hours—unworkable.

The firm engaged Relativity partner Purpose Legal, which used Relativity aiR for Review for generative AI–powered issues analysis across 10 key issues. Working with a single firm partner/SME (and aided by litigation support), Purpose iterated prompts on a stratified sample of fewer than 500 documents, refining the prompt in only three iterations before running aiR on the full population. They validated results with random precision and elusion samples focused on overall review recall rather than issue-by-issue validation.

Purpose Legal completed the 300,000+ document review in seven days with one project manager and one law firm partner for subject-matter expertise. Validation showed recall above 95%, higher than manual review. The engagement reduced review time by 85% (4,000 hours) and delivered cost savings of over $70,000 for the law firm.

• Generative AI issues analysis with tight SME prompt iteration can replace multi-week TAR/contract-review staffing when production deadlines are fixed.
• Small stratified sampling plus rationale-driven prompt refinement (three iterations) can build enough confidence to scale to hundreds of thousands of documents.
• Defensible results still require validation via precision and elusion sampling, with recall as the primary quality gate for production readiness.