Fingerprint, an industry-leading provider of device intelligence to prevent fraud, has launched AI-driven recommendations in its Suspect Score offering. This innovation allows customers to apply machine learning-driven fraud scores, which are based on customer data and can be customized. These scores provide high-quality detection capabilities without compromising transparency and control in fraud scoring.
Static scoring is not always enough to detect constantly changing fraud behavior across different sectors. Fraud teams may be too busy analyzing relationships between signals and adjusting model weightings to be able to react timely. AI-driven recommendations offered by Fingerprint eliminate the need for manual tuning.
“Fraud patterns vary by business and evolve constantly, rendering manual tuning obsolete,” said Valentin Vasilyev, CTO and co-founder at Fingerprint. “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.”
Fraud Detection Adaptation to New Threats
With the advent of sophisticated AI-driven bots and agents, fixed models used for detecting fraud become obsolete, and businesses are open to new kinds of cyber threats. Moreover, users use more VPNs, and therefore it is difficult to apply traditional approaches to weighting fraud signals.
In this regard, fingerprint solves the problem through the advanced Suspect Score based on machine learning technology. The technology leverages Smart Signals to provide up-to-date and actionable insights regarding devices to facilitate effective fraud detection.
Fraud prevention teams within companies may have the option of feeding the machine learning algorithm with the labeled data on fraud to help it understand customer traffic behavior. As such, companies can now train their fraud detection algorithms in accordance with changing scenarios.
A New Standard for Fraud Detection
The improved Suspect Score works by leveraging customer data with Smart Signals to build customized signal weights for specific patterns of fraud. Signal weights are adjusted in accordance with observed patterns to minimize false positives without affecting accuracy.
The user gets a preview of the recommended adjustments prior to implementing them for maximum control over scoring. Companies can train scoring models with new data to keep up with actual fraud behaviors.
Fingerprint transforms fraud detection from a rigid approach to one that is flexible and adapts to the traffic patterns of organizations. The AI-powered recommendations are now available to all customers using Smart Signals through the Fingerprint dashboard.
For related updates on digital trust and cybersecurity, explore our SOC News.
Source: Businesswire
