Kount, a digital fraud prevention company, has announced its next-generation, AI-driven solution that changes the way payments-fraud prevention is delivered. This latest advancement creates an extremely close simulation of the decision process of an experienced fraud analyst, but in a faster, more accurate, and more scalable manner. Kount’s AI uses both supervised and unsupervised machine learning along with additional calculations to deliver a near-human decision, enabling companies to control business-driven outcomes such as higher revenue, reduced fraud losses, and lower operational costs.
 
“The ability for brands to have a more detailed understanding of the challenges that fraud can create is more important today than ever before, and this next-generation technology will help brands be more informed and prudent in their efforts,” says Mark Johnson, Loyalty360 CEO.
 
Kount’s AI emulates an experienced fraud analyst by accounting for both historical fraud patterns and 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. When Kount’s AI considers historic data, it turns to supervised machine learning, which is trained on Kount’s universal data network that includes billions of transactions that have occurred over the course of 12 years in over 180 countries and territories.
 
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 uses unsupervised machine learning that employs advanced algorithms and models to detect anomalies much faster, more accurately, and on a more scalable basis than human judgment alone.
 
Johnson says, “There is a rising challenge with fraud and being able to effectively, efficiently, and in a timely manner identify the transactions that are potentially fraudulent and those which are not. Brands must have effective treatments and procedures for those transactions that could be fraudulent so that they can make sure, in a proactive manner, that those transactions do not impact the customer’s experience and potentially impact brand loyalty.”
 
An experienced fraud analyst weighs the risk and safety of the transaction to make decisions based on risk tolerance, whether that means controlling chargeback rates, accept or declines rates, or manual reviews. Kount’s solution enables the analyst to set policies for these thresholds based on a new score: Omniscore. The enhancement is twice as effective as existing models at detecting payments fraud, but it also maintains Kount’s 250 millisecond response rate.
 
“The supervised machine learning aspect of Omniscore reflects the historical experience that seasoned fraud analysts possess, while the unsupervised features simulate the instinct or ‘spidey sense’ of the very best analysts to detect that a new type of attack is underway,” says Tricia Phillips, Kount’s SVP of Product and Strategy. “More than any other model we’ve seen, Omniscore truly behaves as a human would in the risk assessment of a payment transaction, which is the very definition of artificial intelligence.”
 
Steven D’Alfonso, Research Director with IDC Financial Insights, adds, “The next generation of AI in fraud prevention is much more than machine learning—supervised or unsupervised. It is the ability to simulate, augment, and scale the decision process of an experienced fraud analyst to greatly increase the accuracy and effectiveness of fraud prevention and to deliver desired business outcomes. The ability to quickly identify complex and emerging fraud patterns by Kount’s new AI solution, along with customizable controls for the business, will play an important role in allowing businesses to achieve their financial goals without sacrificing customer experience.”
 
Kount’s AI quickly and accurately detects existing, emerging, automated, and complex fraud. Kount provides the customizable control companies need to protect against fraud and confidently achieve specific business objectives, such as balancing chargeback rates, decline rates, and operational costs.
 

Recent Content