Acquirers, processors, and payment facilitators (PayFacs) are grappling with a host of challenges involving risk—including sophisticated fraud attempts—that demand innovative solutions. These industry players must harness cutting-edge technologies to strengthen their risk management strategies and ensure the integrity of the payments ecosystem.
In a recent PaymentsJournal podcast, Amyn Dhala, Chief Product Officer at Brighterion, a Mastercard company, and Brian Riley, Director of Credit and Co-Head of Payments at Javelin Strategy & Research, delved into the frustrations many are facing and the technologies they should consider to tackle these obstacles.
The Overarching Challenges in Play
One can argue that the single most important thing within the payments space is the irrefutable ability of the transaction. For acquirers, processors, and PayFacs, that means the payments system has integrity and the controls in place to ensure the transactions are indisputable.
But that’s easier said than done. Many factors are attacking the system, and many players continually seek to disrupt it.
“You have people that might go outside the bounds of their credit lines or their available credit and people gaming the system,” Riley said. “But having the fundamental controls there are really what distinguishes the payment process and an effective transaction through the whole system.”
Dhala agreed, stressing that merchants are ultimately looking to increase their bottom lines.
“It’s basically a very dynamic space with lots of opportunities,” Dhala said. “But at the same point in time, it has its own challenges. It comes back to the core (of it) for acquirers and PayFacs, and that’s how do we actually increase revenue? And how do you minimize fraud risk?”
Harnessing the Power of AI
Artificial Intelligence has become a popular and effective tool for industry players to leverage in detecting and preventing fraud. That’s because AI solutions can analyze an enormous amount of data, which can then detect patterns and anomalies, revealing fraudulent activity. It can also lessen the number of false positives.
“Fraudsters are operating at scale,” Dhala said. “So, there are some quick learnings which you can get by leveraging (AI) insights.”
It’s important to note, however, that an AI solution must be fed an enormous amount of data to be truly effective and accurate in its predictions.
“The importance of AI is to keep learning,” Riley said. “It’s not to have a static model that says these are the exceptions we do. The more transactions that go through give (payments players) the ability to learn more on what’s a good or bad transaction.
“If you do this in a box on your own as an issuer, you’re limited to the information that you have. And your solution really uses a lot of the learnings with consortium data to apply that logic throughout the cycle. That really helps make this more powerful.”
More Access to Global Transaction Data Insights Is Key
An AI solution is only as good as the quality and variation of the data it collects, but amassing data for the sake of it is not the answer. Payments players need to be able to continually learn from every transaction, every approval, and every declination that goes through. Doing so will give them a larger knowledge base, Riley says.
“It’s all about that particular balance, which is so crucial to maintain,” Dhala said. “For the acquirer, for the acceptance ecosystem—and frankly the whole commerce ecosystem—to succeed, that’s the core objective. That’s the basis for some of the market model transaction fraud models because it really leverages the network intelligence, which we have at Mastercard.”
Global Transaction Intelligence Helps Address Ongoing Pain Points
A common roadblock to the full adoption of AI solutions is the complexity of its integration. However, Brighterion’s AI Transaction Fraud Monitoring solutions are not only market-ready but also require 30 data elements to train the model, versus hundreds of types of labeled data elements.
“We’ve honed the technology over the last couple of decades, and our fraud intelligence is enriched every year with over 100 billion transactions,” Dhala said. “The combination of this delivers exceptional accuracy, which we can enable to our customers through these transaction fraud models.”
