How Machine Learning Revitalized Interest in Risk Analytics

David Gaudio, June 18, 2019

In the world of fraud security, we have seen renewed interest in risk analytics in the last year – and for good reason. By including machine learning, a new generation of risk analytics solutions can build on the successes and correct some of the problems of the past.

Will LaSala, Director of Security Solutions at OneSpan, was recently interviewed by Teri Robinson, Executive Editor at SC Media. In the interview, LaSala discusses some of the advantages machine learning brings to risk analytics and why AI is so critical in combating the modern threat landscape.

Risk Analytics With and Without Machine Learning

Previously, risk analytics solutions were static components in a suite of anti-fraud tools. They were useful in analyzing data after a transaction had taken place and providing a financial institution with valuable information to help them more accurately set their rule-based alert and anti-fraud strategy. But, a rule-based system can no longer keep up with fraud attacks evolving in complexity, speed, and automation. Rule libraries keep on expanding, which puts pressure on the system, slows operations, and increases the rate of false positives. The inclusion of machine learning with risk analytics significantly increases the value of an anti-fraud solution. Machine learning provides the ability to collect disparate data, analyze that data at scale and in context, and assign a risk score in real-time. This enables a risk analytics solution to apply the precise level of security, at the right time, through step-up authentication.

  • Assigning a Risk Score in Real Time: Financial institutions collect transaction, device, and user activity data on all of their customers, but such a large pool of data will be unintelligible to human analysts. There’s simply too much information. Machine learning allows risk analytics solutions to identify known and unknown fraud schemes based on this information. Despite the massive collection of information, the risk engine can analyze the data and locate fraud trends that would be otherwise impossible to spot by a human.
  • Real-time Authentication Control: In addition to analyzing such large quantities of data, machine learning makes it possible to do so in real time. This allows financial institutions to determine and apply authentication requirements that match the relative risk of the transaction, as the transaction is taking place.

Frost & Sullivan 2019 Best Practices Award for Risk-based Authentication

Discover why Frost & Sullivan recognized OneSpan with the 2019 Best Practices Award for OneSpan’s Intelligent Adaptive Authentication solution – and how to best protect the digital customer journey with the right authentication at the right time.

Download now

Three Ways Organizations Should Be Leveraging Risk Analytics

In his interview with SC Media, LaSala recommended three ways a financial institution (FI) can benefit from risk analytics solutions:

  1. Analyze Mobile Data: Monitoring the mobile channel is different than web applications. If your organization is not collecting mobile data, you have a blind spot in your fraud analysis. Mobile channel fraud analysis needs to take into account the diversity of devices and operating systems, as well as the fact that the FI has no control over what else is installed on the device. If your anti-fraud solution does not currently recognize the specifics of the mobile channel, then it is not collecting all the data points – and may be blind to mobile fraud attacks.
  2. Simplify the User Journey: While attempting to reduce fraud, financial institutions should be careful to preserve a positive customer experience. FIs should not apply their traditional banking customer journey to new digital channels. This causes a lot of friction created by manual steps and paper-based processes. With a risk analytics engine running behind the scenes, the solution is constantly analyzing customer transactions to identify risk and lower false positives. This preserves the customer experience, because a risk analytics solution will only inject an authentication challenge in the event of a high-risk transaction.
  3. Tie Together All of Your Data: Possessing large quantities of data, even if it comes from all the right sources, will not be useful without a sophisticated risk analytics solution to analyze it. By unifying user data and analyzing it alongside contextual information, financial institutions can better anticipate fraud schemes.

How OneSpan Can Help

OneSpan Risk Analytics helps protect against attacks happening in the online and mobile channels by identifying risk at critical steps, predicting risk levels, and taking instant action when suspicious activities are identified. The solution leverages machine learning to go beyond the static reporting and analytical tools of previous years – and can identify and mitigate fraud by allowing, reviewing, or blocking high-risk transactions. Learn more about OneSpan Risk Analytics.

David Gaudio is the Senior Content Writer for all things security and e-signature at OneSpan with nearly ten years’ experience in digital marketing and content creation. David earned his BA in Publishing and Creative Writing and has since worn almost every hat in the digital marketing closet.