Visa and functional space bring fraud fights to new arenas
In today’s war against financial fraud, the battlefield is developing faster than ever, as artificial intelligence (AI) facilitates both sides and opponents becoming increasingly complex.
That's why PYMNTS calls on Visa DPS GM, risk products and solutions Dustin White, and has Space Chief Operating Officer Tim Vanderham to unravel how payment leaders answer through pioneering approaches based on behavioral analysis and collaborative innovation.
“It's not just about stopping the fraud that happened yesterday,” White said. “The real question is, are you ready to stop the fraud that comes to tomorrow?”
Fraud detection has long relied on rules-based systems that flag “bad” activity after it occurs. Although these systems have been historically useful, it is becoming increasingly obvious that they do not align with AI fuel strategies generated by cybercriminals such as social engineering, synthetic identity theft and deep strike scams.
The new goal is to stop fraud before attempting a transaction.
“We’re not only looking at deals,” Plemant’s Vanderham said. “We are looking for login, payee data and behavioral metadata. Fraud events are often the last scene in a longer game.”
Rewrite fraud rules prevention
In response to the AI-generating fuel strategy adopted by cybercriminals, Visa DPS has partnered with Featherpace, a British company specializing in adaptive behavior analysis to refine its fraud detection method.
“Historically, fraud systems have been centralized to the bad,” White said. “But in a world of digital trade, the importance of analyzing the good really cannot be underestimated.”
At the heart of the feature space model is adaptive behavioral analysis (ABA) and paired with a patented automated deep behavioral network (ADBN). These technologies enable AI systems to track normal behaviors of users, including their spending, where to log in, and which merchants they interact with, and bias in milliseconds.
“Visa brings the breadth of global transaction data,” said Premate’s Vanderham. “We bring the depth of modeling and behavioral analysis. In this case, one plus one equals more than two.”
According to Vanderham, the “feature space” model can analyze 500 behavioral variables in less than 50 milliseconds, cross-reference new transactions with users' long-term behavioral history.
“If I make a debit payment at Starbucks every morning while traveling, the system remembers that,” he said. “We do transactions, parse it into its components and compare it to historical behavior at the account and network level.”
This immediacy is more important than ever. As digital payments become ubiquitous, even minor disruptions caused by false positives can get customers packed.
“Fraud is no longer just an indicator of financial exposure,” White said. “It’s an indicator of experience. Missing a deal or mistakenly rejecting a legal deal, the client may walk.”
Fight fraud before you start
The ever-evolving nature of fraud is disturbing. Large language models can now mimic sounds, write trustworthy scripts in any language, and produce AI avatars with real things.
“Our own sales team received a deep effort from me asking for confidential information,” Vanderham said. “This is no longer phishing. It is precise social engineering using voice, video and data manipulation.”
Scams are becoming more humane and emotionally manipulated. Upstream signals, such as login behavior, account origin patterns, and even typing speed, play an enhanced role in enhancing responses across early detection stages.
“Someone builds trust in the victims in a few days or weeks,” White said. “When they start the transaction, it looks effective. Stop stopping at the transaction level, it's too late.”
Worryingly, the rise of dormant M-sub-accounts and cash distribution plans. A customer on the West Coast of the United States used Profem Pass tools to remove fraud rings, capturing 32 arrests and hundreds of thousands of dollars.
“That’s what makes me wake up at night,” Vanderham said. “Triangle fraud, deep smoke and cross-border cash sales, which can lose funds in minutes.”
“The fraudsters have cooperated across borders across borders,” White said. “We have to respond equally uniformly. It's not a visa issue, it's a feature space issue, but an ecosystem issue.”
In addition to technology, cultural changes are also underway. The organization must stop operating as siloed nodes and start thinking like an interconnected network. If every bank or credit union should take a step immediately, it is changing the conversation internally.
“Prefer fraud prevention as a client trust initiative,” White said. “Get rid of the fraud department’s mindset and make it a business-wide priority.”
