Fingerprint, a company specializing in device intelligence for fraud prevention, has integrated AI-powered recommendations into its Suspect Score solution. This enhancement provides customers with an adaptive, intelligent fraud score that trains on their own labeled data, aiming to improve detection accuracy while maintaining transparency and user control.\n\nStatic fraud scoring models struggle to keep pace with dynamic and traffic-specific fraud patterns. Fraud teams often lack the time and resources needed to continuously analyze signal interactions and adjust model weights for their unique operational requirements.\n\nFingerprint’s AI-powered recommendations are designed to address this challenge by eliminating the need for manual tuning, thereby saving time and resources for fraud teams and making their fraud detection capabilities more adaptive to evolving threats. Valentin Vasilyev, CTO and co-founder at Fingerprint, stated, “Fraud patterns vary by business and evolve constantly, rendering manual tuning obsolete. Our AI-powered recommendations remove that bottleneck by training on each customer’s labeled data, making Suspect Score customizable, accurate, and easy for customers to use.”\n\nThe enhanced Suspect Score utilizes a production-ready machine learning system built on Fingerprint’s suite of Smart Signals, which provide real-time device intelligence. Enterprise fraud and security teams can now upload their labeled fraud data to train the ML system based on their specific traffic patterns as threats change. The updated Suspect Score intelligently analyzes this customer data alongside Smart Signals to generate optimized signal weights, adjusts these weights to reduce false positives while maintaining accuracy, and offers a preview of all recommendations before changes are applied, giving users full visibility and control over their scoring.\n\nOrganizations can retrain their scoring models with up-to-date data as threats evolve, ensuring detection mechanisms remain aligned with real-world fraud behavior. This capability shifts fraud detection from a static approach to a continuously adaptive one.\n\nThe AI-powered Suspect Score recommendations are currently available to all Fingerprint customers who have access to Smart Signals. Existing customers can initiate the training of customized scoring models via the Fingerprint dashboard. Fingerprint detects the intent of human and agentic visitors, utilizing its device intelligence platform to identify over 1 billion unique devices monthly and process hundreds of signals to assist fraud teams in distinguishing trusted visitors from malicious actors rapidly and at scale. Over 6,000 companies, including Dropbox, Booking.com, and checkout.com, reportedly use Fingerprint to identify high-risk activity and prevent fraud.