Not surprisingly, machine learning technology was on full display this year at MRC Vegas. Using new self-directing machine-learning models, merchants are increasingly outsourcing significant but noncore operational tasks like writing fraud rules, case management, and even the financial risk itself.
Why? In many cases, retailers are simply not equipped with the rigorous data intelligence, rules, machine learning, and monitoring to effectively discover and combat fraud themselves. Due to this complexity, merchants increasingly want the fraud and the associated chargeback risk to be handled by a third party.
A couple things from the show caught my eye as it relates to fraud prevention:
- In a keynote, Nuno Sebastiao, Feedzai’s co-founder and CEO, talked about how ML technology can’t just produce a fraud score. Rather, humans need to understand how the technology works, and the data output needs to be “explainable” and “controllable,” with clear significance from a business perspective… not just a technology perspective.
- The exhibit hall was full of “new breed” fraud prevention vendors that want to run the technology and “own” the fraud risk. Legacy tools, which are often built in-house, lack the agility, speed, and instrumentation to effectively combat fraud, and old-fashioned, purely rule-based fraud risk scoring models can’t keep up.
In my recent report, Ten Unique Solutions To Five Retail Payment Acceptance Challenges, I talk about some of these trends and vendors.