Case Study: Nightfall achieves 2.5x labeling gains and 2.5x fewer false positives with Datasaur

Nightfall Achieves 2.5x Labeling Gains and 2.5x Fewer False Positives with Datasaur

Nightfall, a data security company focused on protecting sensitive information across SaaS, GenAI tools, email and endpoints, struggled with a rigid labeling platform that prevented editing prior labels and forced QA to rely on sheer volume—driving up labor costs and reducing confidence in their data. After evaluating 55 tools, Nightfall chose Datasaur for its labeling and QA tooling to regain flexibility and improve data quality.

Datasaur deployed features like Reviewer Mode, multi-pass labeling, hotkey shortcuts, flexible span editing and QA reports integrated with Nightfall’s AWS pipeline; within a year Nightfall labeled 180,000 documents and found 97,000 sensitive instances. The Datasaur implementation delivered measurable impact: 2.5x labeling gains, 2.5x fewer false positives, a 1.5x improvement in model recall (50% more entities), and a 1.5x increase in models driven by labeled data.

Nightfall