Tackling the Complexities of Modern Payment Fraud: Key Insights from Fintech Meetup 2025

As real-time payments (RTP) continue to surge in popularity, new payment processing systems such as FedNow and expanded peer-to-peer (P2P) networks are reshaping the financial landscape, providing unprecedented convenience while raising regulatory and security concerns. The immediacy and irreversibility of such payment systems introduce unique vulnerabilities and put pressure on financial institutions to protect customers while keeping friction out of the user experience while minimizing friction in the digital payments experience.

To address these challenges, industry leaders in financial services and fraud prevention experts gathered at Fintech Meetup in March for a panel discussion titled, “Tackling the Complexities of Modern Payment Fraud.” They explored emerging threats and strategies to mitigate the risk involved with digital payment solutions through AI-driven fraud detection, targeted risk controls and industry collaboration and partnerships. They also discussed the delicate balance between security and user experience.

Among the panelists was co-founder and CEO of DataVisor Yinglian Xie, who emphasized that leveraging artificial intelligence (AI) for fraud detection is crucial for mitigating risks in domestic transactions, and for securing cross-border payments in today’s interconnected financial ecosystem.

In this blog, we take a look at the key insights shared during the panel discussion, use cases for AI-powered fraud detection, and how to prevent fraud losses across the payments ecosystem.

Real-Time Payments and AI-Powered Fraud

One of the most significant shifts in financial crime over the last 12 to 18 months has been the rapid growth of RTP. While the new technology offers immense convenience to consumers and businesses, it also presents a high-risk environment for fraud, as fraudulent transactions settle instantly and are difficult to reverse.

“We’re seeing an increase in both frequency and value of fraudulent transactions,” said Jenna Kaye-Kauderer, Head of AirKey at Capital One. “Stronger risk models and real-time monitoring are crucial to slowing down suspicious transactions and adding extra layers of security.”

Fraudsters have become more patient and coordinated in their approach, as well, often conducting small-scale test attacks before launching full-scale fraudulent campaigns. And, AI plays a dual role: while financial institutions use AI to detect and fight fraud, fraudsters leverage the same technology to enhance their scams.

According to DataVisor’s Xie, AI-driven fraud tactics have reached new levels of complexity. “Just last night, our finance team sent me an email asking me to confirm whether I had an email conversation with a procurement vendor,” she shared. “The email contained a full conversation history: multiple exchanges, a well-crafted storyline, and even detailed apologies for delays. It looked completely legitimate, but it was entirely fabricated. This is the level of sophistication we are dealing with today.”

As AI-generated fraud grows, Xie added, so does the need for advanced, real-time fraud detection models that don’t rely solely on historical patterns. Traditional methods may fail to recognize new attack vectors, especially in high-velocity RTP environments

Social Engineering Scams Target Vulnerable Consumers

A concerning trend discussed is the rise of social engineering scams, particularly those targeting older adults. Fraudsters are shifting from quick-hit scams to long-term engagement tactics, spending weeks or even months building trust with their targets through phishing emails and social media before executing their schemes.

According to the panelists, these scammers are increasingly adept at mimicking trusted individuals and institutions, carefully timing their actions to exploit a victim’s trust. Once victims are convinced, fraudsters can easily manipulate them into approving significant financial transactions and using digital payment methods, causing financial losses.

The human element in fraud remains one of the biggest vulnerabilities, and financial institutions are grappling with how to balance security measures with a seamless customer experience. Trusted contacts and behavioral monitoring were highlighted as key safeguards for vulnerable users.

Improving Security without Adding Friction

One of the biggest challenges for today’s financial institutions is how to implement robust fraud mitigation without introducing excessive friction for legitimate users. Traditionally, fraud detection relied on static rules and manual reviews to identify fraudulent activity, but with the rise of machine learning models and behavioral analytics, financial institutions can now dynamically adjust authentication methods based on transaction risk.

Capital One takes a layered approach to authentication. “You need to have a broad set of authentication tools and dynamically apply them based on transaction risk,” Karder said. “You can’t rely solely on biometrics, government IDs, or SMS one-time passcodes. A layered approach is key.”

The panelists also discussed hardware authentication as a critical fraud prevention tool. Innovations like AirKey and dedicated authentication devices beyond mobile phones are gaining traction, especially for high-risk or high-value transactions. These tools add an extra layer of security without burdening users with complex login processes.

The Role of AI in Fraud Detection: Strengths and Limitations

While AI has become an essential tool in fraud prevention, it’s not without its challenges. Supervised machine learning models, which rely on historical fraud patterns, can falter when fraudsters introduce new tactics that don’t fit past data trends. This is where unsupervised learning, holistic customer profiling and real-time data infrastructure become indispensable.

“Fraud patterns evolve rapidly,” said Xie. “If we rely only on supervised learning, we will always be one step behind. Unsupervised learning allows us to detect anomalies without predefined labels, making it a crucial component of modern fraud detection.”

Xie also highlighted the importance of investing in high-quality data to train fraud detection models effectively, since poor-quality data can lead to false positives, customer frustration, and inefficiencies in fraud prevention systems.

Looking Ahead: Fraud Trends in 2025 and Beyond

The panelists were unanimous on one point: financial institutions must be proactive rather than reactive in their fraud prevention efforts, as fraud will continue to become more complex and sophisticated.

What fraud trends will gain momentum in 2025? The panelists predict:

• More AI-driven social engineering attacks such as deepfake voice, video, and email scams
• RTP abuse to move money instantly and evade detection
• Targeted fraud campaigns based on user demographics
• Increased B2B payment fraud as fraudsters expand beyond consumer channels

In order to strengthen fraud prevention strategies, financial institutions should:

• Invest in real-time fraud detection infrastructure.
  Traditional batch-processing models are no longer sufficient for detecting fraud in RTP environments.
• Adopt a multi-layered authentication approach.
  Combining behavioral analytics, biometrics, hardware authentication, and risk-based decision-making will enhance security.
• Prioritize data quality and AI-driven fraud detection.
  AI algorithms are only as good as the data they’re trained on, so high-quality, real-time data is essential.
• Educate consumers about fraud risks.
  Financial literacy and awareness campaigns can help users recognize and avoid scams.
• Implement trusted contacts and transaction monitoring for vulnerable populations.
  Older adults, in particular, need extra safeguards against exploitation.

“Fraud prevention is an ongoing battle,” Xie said. “The financial community must work together, share intelligence, and continuously innovate to protect consumers and businesses from ever-evolving threats.”

Ready to unleash the power of AI against fraud? Learn more about DataVisor’s AI-powered fraud and risk platform today, and how it can help you improve fraud management and mitigation across the fraud landscape.