A large online lender was conducting manual risk reviews on nearly all applicants for both consumer and small business risk decisions, and wanted solutions to generate a real-time decision for the most honest applicants, based on a strict superset of current information.
The client processed a carefully constructed set of ~50,000 records against batch manual verification sources and appended unstructured information. The DemystData platform automated data normalization and rule configuration to identify pockets of customers with high conditional likelihood of verification.
- Waived manual verification requirements for verification likelihood above 95%
- 50% of manual verification checks avoided
- Increased conversion rates of up to 5% due to improved applicant experience