By Fintech Global
Click here to read the entire article.
Fingerprint, a leader in device intelligence for fraud prevention, has announced the addition of AI-powered recommendations to its Suspect Score solution, marking a significant step forward in adaptive fraud detection.
Static scoring models have long struggled to keep pace with increasingly dynamic, traffic-specific fraud patterns. Fraud teams frequently lack the time and resources needed to continuously analyze signal interactions and manually tune model weights to suit their unique operational needs. Fingerprint’s latest enhancement directly addresses this gap, enabling fraud teams to eliminate manual tuning, preserve valuable time and resources, and deploy fraud detection that adapts to evolving threats.
Fingerprint provides device intelligence solutions designed to help organizations identify and prevent fraud. Its platform is built around Smart Signals — actionable, real-time device intelligence insights — that deliver powerful fraud indicators to enterprise fraud and security teams. The company’s Suspect Score solution sits at the center of this offering, giving customers a consolidated fraud risk signal drawn from a broad range of device and behavioral data.
The enhanced Suspect Score introduces a production-ready machine learning (ML) system that customers can train on their own labeled fraud data. Enterprise teams can upload this data through the Fingerprint dashboard, enabling the system to intelligently analyze it alongside Smart Signals to generate optimized signal weights tailored to their specific fraud patterns. The updated solution also adjusts signal weights based on patterns observed in a customer’s fraud data to reduce false positives while maintaining accuracy. Before any changes are applied, customers receive a full preview of all recommendations, allowing them to review and approve updates with a single click — preserving complete visibility and control over their scoring configuration.
As threats continue to evolve, organizations can retrain their scoring models with up-to-date data, ensuring detection remains aligned with real-world fraud behavior. Sophisticated AI agents and bots are increasingly capable of bypassing static detection models, and the growing adoption of privacy tools such as VPNs among legitimate users has further complicated traditional signal weighting. Fingerprint’s AI-powered approach is designed to meet these challenges head-on, shifting fraud detection from a static model to a continuously adaptive one.