Sift introduces ThreatClusters for fraud detection

More about ThreatClusters’ anti-fraud capabilities

Fraudsters now use more sophisticated tactics, such as AI-powered attacks, to outsmart many fraud prevention strategies. Conventional fraud detection systems often struggle because they either focus too narrowly on one organisation's data or apply their insights too broadly across different sectors. 

To address these challenges, ThreatClusters groups companies with similar fraud trends into cohorts to accommodate variations in risk patterns, resulting in more precise fraud decisions. Moreover, ThreatClusters helps enterprises by combining industry-specific model insights and blending customer-specific risk models with a global model to produce industry-specific risk signals.

Additionally, by using Sift's technology, cusomers can employ a customised detection model that is specific to their cluster, as well as use detection models that may provide insights into new fraud patterns from other clusters.

Key features and benefits of ThreatClusters:

• More Accurate: ThreatClusters upgrades fraud detection accuracy by incorporating industry-specific fraud patterns, reducing false positives/negatives by up to 20%.

• Faster time-to-value: combining global and cohort models more accurate, leading to a quicker adoption process and more rapid realisation of benefits for companies.

• Refined user friction: industry-specific fraud patterns help differentiate between genuine users and fraudulent entities, introducing necessary security measures without affecting customer experience and conversion rates.

Officials from ThreatClusters stated that a notable advancement in assisting businesses in outsmarting fraudsters is the introduction of industry-tailored consortium models. These models offer clients visibility into the distinct fraud trends within their sector, safeguarding them from new threats from other industries. This helps them evaluate risks, secure revenue, and expand without fear.