How AI Is Changing Fraud Detection in Modern Payments with Guest Brian Keefe
Wrestling Payments Podcast: Season 3 - Episode 23
In this episode of Wrestling Payments, host Joe Casali sits down with Brian Keefe, Manager of Portfolio Pre-Sales at Nice Actimize. They explore how artificial intelligence is shaping fraud detection and compliance in the evolving payments landscape, with a focus on regulatory changes and industry adaptation.
Brian discusses the impact of new NACHA rules and the shift from paper checks to digital payments. He breaks down how fraudsters use advanced technology, including AI, to create synthetic identities and target institutions across multiple channels. Brian explains why real-time monitoring, cross-channel risk assessment, and clear AI guardrails are now essential for banks and credit unions. He highlights the need for explainable AI, where human oversight—what he calls “human in the loop”—ensures that technology supports, rather than replaces, experienced staff.
The conversation closes with practical advice for institutions of all sizes. Brian emphasizes that AI should function as an assistant, helping teams work more efficiently while staying ahead of sophisticated threats in a fast-moving digital world.
Download the Transcript
Guest-at-a-glance
💡 Name: Brian Keefe
💡 What they do: Manager, Portfolio Pre-Sales
💡 Company: Nice Actimize
💡 Noteworthy: Brian specializes in AI-driven fraud prevention and compliance for financial institutions, with expertise in cross-channel risk, monitoring, and regulatory adaptation.
💡 Where to find him: https://www.linkedin.com/in/brianekeefe
Key Insights
Fraudsters Are Using AI—So Institutions Must Keep Pace
Criminals now use artificial intelligence and advanced technology to commit payment fraud at scale. With tools that can produce convincing counterfeit checks and create synthetic identities, fraud groups collaborate online and adapt quickly to new defenses. Traditional fraud monitoring—especially methods focused only on outgoing payments—no longer works. Financial institutions need to evolve their approach, looking at both incoming and outgoing payments and monitoring across all channels. Staying ahead means adopting modern tools, training staff to recognize new fraud patterns, and understanding that fraudsters move fast and share knowledge. If banks and credit unions delay updates to their processes, they risk falling behind and exposing themselves to losses. Recognizing that fraud is now a technology-driven threat is the first step to building a stronger, more resilient payments operation.
Explainable AI and Human Oversight Are Essential
AI can transform fraud prevention, but it brings new risks if not managed well. Financial institutions cannot afford to rely on AI as a “black box” that makes decisions without transparency. Explainable AI—where the system can show why it triggered an alert or flagged a transaction—helps build trust within compliance teams and with regulators. Just as important is “human in the loop” oversight, where experienced staff review AI-generated alerts, adjust recommendations, and ensure the technology works as intended. This partnership between human expertise and machine learning drives better outcomes: fewer false positives, faster investigations, and more effective responses to fraud patterns. Relying on explainable, well-governed AI protects both the institution and its customers while meeting growing regulatory scrutiny.
Cross-Channel Monitoring Delivers Stronger Fraud Defense
Fraud rarely stays in one payment channel. Modern criminals exploit weaknesses across ACH, checks, wires, and digital banking, sometimes moving quickly from one to the next. Cross-channel monitoring—using AI to observe patterns and connections across all payment methods—gives financial institutions a fuller picture of risk. This approach helps spot suspicious behavior that single-channel systems might miss, like a fraudster who initiates a check deposit and then quickly moves funds via ACH or wire. Combining data from every channel reduces noise and sharpens alerts, allowing teams to focus on real threats instead of chasing false alarms. As payment systems become more complex and interconnected, cross-channel monitoring is no longer optional—it is a frontline defense against today’s coordinated fraud strategies.
Episode Highlights
NACHA Rule Changes Drive New Monitoring Demands
Approximate timestamp: 00:01:30 – 00:02:27
Recent changes to NACHA rules have shifted the focus for banks and credit unions. Institutions now must monitor both outgoing and incoming ACH transactions, expanding risk management beyond traditional areas. What was once a manual or overlooked process has become a priority as organizations work to keep up with compliance and protect against new types of fraud. The conversation highlights how this shift has prompted many institutions to upgrade their monitoring systems and seek out new capabilities. By paying closer attention to both sides of the transaction, financial organizations can better catch suspicious activity and meet evolving regulatory demands.
“So with NACHA, everybody knows it's probably forefront of everybody's mind in ensuring that they have the proper protection and monitoring capabilities for the new rules. Back in the day, it was just, we're concerned about money leaving an institution, the ODFI side, but now bringing in the extra ability to monitor incoming, the credit side of it. So we've done a lot and we know that institutions have been reaching out for that particular capability to be able to monitor and compare the credit party to the account, to the SEC codes, and various other things.”
The Real Threat of Synthetic Identities
Approximate timestamp: 00:11:13 – 00:13:32
Synthetic identities, powered by AI, have become a real-world challenge for financial institutions. Fraudsters now use artificial intelligence to combine real and fake personal information, creating entirely new identities that can slip past traditional controls. These identities are used to open accounts, commit fraud, and avoid detection. The rise of deepfakes and AI-driven data mining makes it easier for bad actors to generate convincing profiles, including realistic photos. This reality has forced banks and credit unions to rethink their approach to both account opening and ongoing customer monitoring. Staying ahead means understanding how fraudsters use technology and adapting systems to catch these advanced threats before they cause harm.
“Certainly, from the account opening perspective is getting access to an institution with a synthetic identity, a deep fake, or utilizing AI as well in capturing that intelligence and making a complete profile for an individual. So you're not going to have a fraud of going out and searching the web for particular information. You're going to let AI do that for you. Hey, bring me information regarding a middle-aged male who is in a particular area. It's going to combine all that intelligence—you know how ChatGPT works and how effective in a manner that is. That's certainly what the frauds are doing to create that identity, to get into the institution, to be able to create a new photo.”
AI as an Assistant—Not a Replacement
Approximate timestamp: 00:17:38 – 00:18:22
The role of AI in financial crime prevention is not to replace people but to make their work more efficient. There’s a common misconception that automation will take jobs, but in reality, the most effective systems blend machine learning with human expertise. AI can process large volumes of alerts, highlight unusual activity, and reduce noise in fraud detection. However, it cannot replace the judgment, insight, and oversight that experienced fraud analysts bring. The future of compliance and risk management lies in collaboration: let AI handle the repetitive tasks while people focus on analysis and decision-making. This approach helps institutions stay agile as transaction volumes grow and fraud tactics evolve.
“A good AI solution or a good monitoring solution, AI is going to complement what that individual does, allowing them to do it more efficiently and effectively. AI should never replace what a human does, especially from the risk and compliance perspective, because in this day and age, maybe in ten years, regulators and institutions would feel comfortable with AI making decisions end-to-end, say an alert to a case, a case to a regulatory filing, and then allow that AI to file the regulatory filing. We are not even anywhere close to that.”
Education and Adaptation Are Ongoing
Approximate timestamp: 00:36:18 – 00:37:09
Education is a continuous need as AI and fraud detection tools evolve. Even conservative institutions must keep up, because fraudsters are always adapting and using the latest technology. Staying informed means reading white papers, following industry updates, and learning from peers. The pace of change in fraud tactics, regulatory shifts, and digital banking requires a commitment to ongoing training and adaptation. Leaders who invest in learning—both for themselves and their teams—are better prepared to defend against complex threats and to use new tools effectively.
“There’s no cost to the papers. It’s a white paper. Very, very educational. I’ve learned a lot. When you’re in a position with members where stuff is developing—AI and fraud detection is developing—you have to educate yourself. You guys have been a great resource.”
To hear this episode and many more like it, subscribe to Wrestling Payments on Apple Podcasts, Spotify, or anywhere else you listen to podcasts, or listen above.