AI-Driven Fraud Prevention,
Business-Driven Outcomes

Kount AI changes the way payments fraud prevention is delivered by simulating the decision process of an experienced fraud analyst, yet in a faster, more accurate, and more scalable manner.

AI-Driven Digital Fraud Prevention

Kount AI uses both supervised and unsupervised machine learning along with additional calculations to deliver a near-human decision, allowing companies to quickly and accurately detect existing or emerging, automated, and complex fraud. Over 6,500 leading brands rely on Kount digital fraud prevention solution to control business-driven outcomes such as higher revenue, reduced fraud losses, and lower operational costs.

Kount’s AI emulates an experienced fraud analyst by taking into account both historical fraud patterns as well as anomalies. When fraud analysts consider historic data for known fraud patterns, they look at the company’s data and their own experience to identify whether or not the person or transaction can be trusted. Then, fraud analysts look for anomalies—something in a transaction that doesn’t look right. This is where emerging fraud trends are detected.

Kount’s AI simulates that process, but makes it faster, more accurate and more scalable to fit any business environment, from large ecommerce operations with multiple product lines that require different business policies, to smaller ones with simple operations.

Supervised Machine Learning

Kount’s supervised machine learning technology learns from historical data--decisioned orders and their outcomes. The model looks for patterns and behaviors that have predicted fraud in the past.

When Kount’s AI considers historic data, it uses supervised machine learning that learns from Kount’s universal data network, which is much bigger than human experience alone, because it includes billions of transactions over 12 years, 6,500 customers, 180+ countries and territories, and multiple payment networks.

Kount AI

Universal Identity Network

Comprehensive transaction and identity data that crosses different transaction complexities, different verticals, and different geographies so machine learning models can be properly trained to accurately predict risk. That analytical richness includes data on physical real-world and digital identities, creating an integrated picture of customer behavior.

This provides merchants—regardless of industry, customer base, or geography—insights to protect against fraudulent activities.

Unsupervised Machine Learning

A fraud analyst looks for anomalies—something in a transaction that doesn’t look right. A human can only detect a limited number of anomalies based on their experience.

When Kount’s AI looks for anomalies, it uses unsupervised machine learning and utilizes advanced algorithms and models to detect anomalies much faster, more accurately, and on a more scalable basis than human judgment alone. This significantly enhances the work of a fraud analyst.

Omniscore

Omniscore is the actionable fraud payments score that is produced by Kount’s AI. It is twice as effective than existing models at detecting payments fraud, while maintaining Kount’s 250 millisecond response rate.

It is a highly-predictive transaction safety rating that can be relied upon for decisioning orders, so that there is less reliance on manual review and reactive fraud rules. The result catches more true fraud and allows more good transactions to generate revenue.

Control Center: Business-Driven Outcomes

Kount’s Control Center provides the ability to fine-tune fraud prevention decisions, conduct investigations, and monitor performance.

It enables customers to create rules and policies that meet their unique business needs (from promotions and policy abuse to non-fraud chargebacks) and customize their risk thresholds to address emerging attack methods, new use cases, and issues such as bad marketing affiliates and SKU-specific policies.

Companies can use the control center to set up thresholds based on Omniscore outputs and their business’ risk tolerance and desired business outcomes, whether that is controlling chargeback rates, accept or decline rates, or manual reviews.

Control Center also includes tools for investigation, rescoring, and reporting.