Strengthening Digital Trust and Security in Credit Unions with AI

As digital banking continues to evolve, credit unions and financial institutions face a growing challenge: delivering seamless digital experiences while managing increasingly sophisticated fraud risks.

In this interview, Michael Lavia discusses how DataVisor approaches fraud and risk detection, and how the partnership between VeriPark and DataVisor embeds real-time protection directly into digital banking journeys.

Q: DataVisor’s platform is known for its AI-driven fraud and risk detection capabilities. Can you explain how your technology works in simple terms?

At a simple level, you can think of DataVisor as a real-time fraud and risk intelligence layer that sits behind a financial organization’s digital channels and quietly watches for anything that looks risky or out of place.

Every time a member logs in, updates their profile, opens an account, or moves money, our platform analyzes hundreds of signals in the background —things like transaction patterns, device intelligence, behavioral signals, historical activity, and relationships to other accounts. Instead of looking at any one event in isolation, we look at the full context.

Under the hood, we combine flexible business rules with both supervised and unsupervised machine learning models. The supervised models are trained on known fraud patterns, while our patented unsupervised technology is designed to spot new and emerging behaviors. 

Together, these complementary layers allow us to flag suspicious activity in real time, before losses occur, while letting legitimate members move smoothly through their journey.

All of this happens in milliseconds, embedded directly in the digital experience, so members see a fast, seamless interaction.

Just as important, our AI Agents help internal fraud teams do their jobs faster, more accurately, and with less manual effort. 

Q: What differentiates DataVisor’s AI and machine learning approach from traditional fraud prevention systems?

Traditional fraud systems were largely built around static rules and batch processing. They're good at catching yesterday's attacks but often struggle with today's fast-moving, AI-driven fraud. There are a few key differences in how DataVisor approaches the problem:

• Proactive, not just reactive. Our platform combines supervised and patented unsupervised machine learning, which means we can detect well-known fraud patterns and uncover new, coordinated attacks that have never been labeled before.
• Unified, end-to-end decisioning. Instead of separate tools for onboarding, login, payments, and investigations, we provide a single decision engine and case management environment that evaluates risk across the entire member lifecycle —from new account opening to everyday money movement and beyond.
• AI Agents that supercharge operations. Beyond detection, we embed AI "agents" that help create and tune rules, summarize alerts and cases, and reduce manual workload.

Across DataVisor's customers, this modern approach has translated into meaningful impact: up to 60% reduction in fraud losses, 5x earlier detection, and significant efficiency gains in investigations when moving off legacy tools.

Q: How does embedding DataVisor’s technology directly into VeriPark’s digital banking platform enhance real-time fraud protection for credit unions?

Embedding DataVisor directly into VeriPark's digital banking platform means fraud protection is "baked in" wherever members are engaging —rather than bolted on after the fact.

Concretely, VeriPark has integrated DataVisor at key points in the digital journey, including new member onboarding, digital banking login and profile changes, and money movement.

Because this is a native, pre-built integration, financial institutions don't need to engineer their own data flows or screens to benefit from AI-driven protection. Risk decisions are delivered within the same digital workflows VeriPark already uses, with low latency and consistent member experiences across channels.

The result is that financial institutions can turn on real-time fraud and risk defenses quickly, and customers benefit from a safer digital banking experience without extra friction.

Q: What operational benefits do credit unions see with an embedded fraud solution compared to standalone integrations?

From an operations standpoint, an embedded solution changes the equation in a few important ways:

• Faster deployment and time-to-value. Because DataVisor is already integrated with VeriPark's digital banking platform, credit unions avoid long, resource-intensive projects to connect disparate systems.
• Less IT burden, more business control. With pre-built data flows and decision points, IT teams spend less time on custom integration.
• Consistent member experience. Since fraud checks happen directly within VeriPark's digital journeys, there's no need to redirect members to separate tools or introduce extra steps that feel disjointed.

In short, embedded fraud protection lets credit unions focus more of their time on strategy and member outcomes, and less on integration projects and operational overhead.

Q: Can you share real-world use cases or early results from credit unions already using this joint solution (e.g., Beem Credit Union)?

Beem Credit Union is a great example of how this joint solution comes to life. Beem is using VeriPark and DataVisor together as part of a broader digital transformation.

With DataVisor embedded into VeriPark's digital banking platform, Beem has prioritized protection for high-impact digital use cases such as account takeover prevention, safer money movement, and new membership and account opening.

As Beem's CEO Brian Harris has said, integrating advanced fraud protection directly into the digital experience is a key part of how they "enhance how we safeguard members across key touchpoints" while building secure, seamless experiences that evolve with members' needs.
 

Early results from the joint solution show meaningful operational and member-experience improvements, including reducing OTP challenges from nearly 100% of sessions to approximately 10%, and lowering manual review volumes from roughly 3,000 cases per day to fewer than 70.

Beyond Beem, we see similar patterns across our credit union clients: consolidating fraud and AML workflows, reducing manual investigations, and achieving material reductions in fraud losses and false positives when moving off legacy solutions and into a unified, AI-driven platform.

Q: According to your 2025 Fraud and AML report, a large percentage of financial institutions say fraudsters are outpacing defenders with AI. How should organizations respond to this trend?

Our 2025 Fraud and AML Executive Report found that 75% of surveyed financial institutions believe fraudsters are currently using AI more effectively than defenders.

From our perspective, there are several practical steps organizations should take:

• Adopt and operationalize AI thoughtfully. This means moving beyond rules-only systems and embracing a combination of supervised and unsupervised machine learning to catch both known and emerging attack patterns.
• Unify fraud and AML operations. Bringing data, signals, and workflows together in a single platform allows teams to see the full picture, speed investigations, and strengthen both compliance and fraud defenses.
• Use AI to augment people, not replace them. AI Agents that summarize alerts, tune rules can reduce manual work while still maintaining human oversight and governance.
• Build explainability and governance from day one. As AI becomes central to fraud and AML, institutions need transparent models, audit-ready reporting, and clear controls. 

Ultimately, staying ahead of AI-enabled fraud requires institutions to think of fraud and AML not just as a defensive cost center, but as a strategic capability that underpins digital trust and growth.

Thank you Michael.