Casepoint Makes Recall, Precision and Human Validation Part of Agent Deployment

On September 23, Casepoint released Relevance Determination and Issue Coding agents to early-access users inside its e-discovery platform. The company said both agents run within existing permissions, record decisions in an audit trail and require validation against a sample set, while the Issue Coding agent can calculate recall and precision for each code. For mid-sized litigation teams, the design provides a more useful procurement benchmark than a general accuracy claim because it makes quality control, cost estimates and stop points part of the workflow. The performance claims are vendor-supplied and no independent outcome data were published, so firms should require matter-specific validation thresholds and document who can approve full-corpus runs.