Sift Unveils ThreatClusters: Revolutionizing Fraud Detection with Industry-Specific Consortium Models

Security
23.08.2024 01:50 pm

Sift, the AI-powered fraud platform securing digital trust for leading global businesses, today announced the launch of ThreatClusters, a groundbreaking data science innovation for fraud detection. ThreatClusters enhances fraud decision accuracy by adding a critical layer of industry-specific model insights, combining the precision of customer-specific risk models with the broad intelligence of a global model to derive risk signals unique to each industry.

Fraud actors are deploying increasingly sophisticated attacks, including AI-powered threats, that can overwhelm and outsmart many fraud prevention tactics. Traditional fraud detection models often fall short, either by too narrowly focusing on a single organization’s data or by applying insights too broadly across diverse industries. ThreatClusters addresses these challenges by clustering companies with similar fraud patterns into cohorts to account for nuances in risk patterns, and driving more accurate fraud decisioning.

By leveraging Sift’s proprietary technology, customers are able to both use a detection model that is fine-tuned to their cluster alongside detection models that could inform on new fraud vectors from other clusters.

Key Features and Benefits of ThreatClusters:

Enhanced Accuracy: ThreatClusters help increase fraud detection accuracy, reducing the risk of false positives/negatives up to 20% by adding the insights of industry-specific fraud patterns.
Faster Time-to-Value: The integration of global and cohort models accelerates model accuracy, providing a faster adoption process and quicker realization of benefits for businesses.
Refined User Friction: Industry-specific fraud patterns better distinguish between legitimate users and fraud actors, invoking step-up friction without compromising the customer experience and conversion rates.

“ThreatClusters represents a significant leap forward in our mission to help businesses stay ahead of fraudsters,” said Raviv Levi, Sift’s Chief Product Officer. “By introducing industry-specific consortium models, we can provide our customers with unprecedented insights into the fraud patterns that are unique to their industry while protecting against emerging ones from other industries. As a result, our customers are better able to assess risk, protect revenue, and grow fearlessly.”

In addition to ThreatClusters, Sift's latest release includes other key innovations that optimize score accuracy and allow fraud and risk teams to more easily detect sophisticated fraud behavior across different use cases, including payment fraud and account takeover.