Future-Proofing Customer Onboarding to Prevent Fraud: Insights from Brian Hughes

Effective customer onboarding in the financial sector is essential for building trust and preventing fraud – but it comes with unique challenges. During a recent DataVisor webinar, Brian Hughes, a seasoned fraud expert with decades of experience, shared his perspectives on how to future-proof customer onboarding.

Hughes, the former Chief Risk Officer at Discover and now a Senior Advisor at DataVisor, has built fraud management programs for digital banks and advised a wide range of financial institutions, payment networks, and fintechs. His experience gives him a comprehensive view of the challenges and opportunities in fraud prevention, particularly during the customer onboarding process.

This blog post dives into insights from the webinar, offering a roadmap for organizations looking to enhance their fraud detection strategies while maintaining a positive customer experience.

The Balancing Act of Customer Onboarding

Customer onboarding is a complex process, and getting it right, according to Hughes, is an optimization problem. “Organizations must balance three key objectives: decline fraudulent applications, approve legitimate customers with minimal friction, and keep costs low,” he said.

“Organizations must balance three key objectives: decline fraudulent applications, approve legitimate customers with minimal friction, and keep costs low.”

The challenge, though, is that fraud teams often focus too heavily on blocking fraud, which can create more friction for good customers. And while rejecting fraudulent applications is essential, it shouldn’t come at the expense of a seamless experience for genuine users.

Hughes highlighted the need to approach customer onboarding with a more strategic mindset – one that takes into account the overlap between fraud and financial crime. “For many organizations, the solution lies in leveraging advanced technologies that can help strike the right balance,” he said. “By utilizing AI-driven models and orchestrating various data signals, companies can make more accurate decisions, reducing false positives while maintaining robust fraud prevention.”

Six Core Capabilities of Fraud-Resistant Customer Onboarding

To manage customer onboarding effectively, Hughes recommends integrating six core capabilities into your fraud prevention strategy. Together, these capabilities form the foundation of a resilient, adaptable onboarding process.

Let’s take a look at each of the capabilities in detail.

Signals

There are 10 essential signals for detecting and mitigating fraud risk. These 10 signals work in concert to create a comprehensive fraud detection framework, with regular updates essential to stay ahead of evolving threats.

The first three – bot detection, behavioral analytics, and device intelligence – focus on distinguishing genuine users from bots and examining devices for any history of fraudulent activity.

• Bot detection typically operates at the firewall level, filtering out suspicious automated actions
• Behavioral analytics assesses user patterns, such as typing speed and device angle.
• Device intelligence dives deeper, analyzing device settings, past associations with fraud, and recurring device use to determine legitimacy.

Following these digital signals are phone and email verifications, which check for burner phones, account name matches, and prior usage in fraudulent activities. Stolen ID and synthetic ID detection use third-party databases to cross-reference applicant information, spotting signs of identity theft or fabricated identities. Together, these signals provide a layered approach to detect discrepancies that may indicate fraud.

First-party fraud – when someone uses their real identity to commit fraud – is becoming more common and harder to detect. Consortium databases and linkage analysis help by connecting applicants with known fraudsters through shared details like addresses or IPs, revealing potential fraud risks.

Document verification serves as a more intensive step, using live video and, sometimes, voice verification to combat AI-driven identity manipulation. Finally, government sources like Social Security Administration matching services provide additional cross-verification of personal information.

Orchestration

Orchestration plays a crucial role in connecting disparate data sources, providing three primary benefits:

1. They connect organizations to a wide range of signal providers without requiring companies to establish individual connections. This simplifies integration, allowing companies to access various data sources through a single orchestrator rather than building and maintaining multiple connections independently.
2. Orchestrators streamline contracting by managing resale agreements, making it easier for businesses to test and identify the best signal providers for their unique needs.
3. Orchestrators like Datavisor can even provide strategic advice based on broad industry experience, guiding companies toward the most effective signals and solutions. Since they understand the performance of different signals across various use cases, they can recommend providers that offer optimal coverage and relevance, and provide valuable insight for updating or expanding singlas and strengthening fraud strategies.

Decisioning

Historically, fraud detection relied heavily on rules-based decisioning, with organizations managing hundreds of rules to determine whether to approve or reject applications. However, today’s environment demands a shift towards AI-powered decisioning.

Machine learning models provide greater scalability, accuracy, and adaptability, making it easier to handle the complex variables involved in onboarding. AI models continuously learn from new data, adjusting to the evolving tactics of fraudsters. While rules still play a role in blocking known threats, they now serve as a secondary line of defense, with AI models taking the lead in real-time decision-making.

Investigation

Even with advanced decisioning, some fraud attempts will succeed, making rapid investigation essential.Speed is critical in identifying vulnerabilities and responding to fraud incidents.

Integrating decisioning insights into the investigation process helps fraud teams quickly piece together related cases, identify patterns, and take corrective action. Linkage analysis, for example, can group related cases, making it easier to uncover fraud rings and respond more effectively. Cross-training staff from related departments, such as financial crimes and customer service, can also improve response times during fraud spikes.

Rapid Response & Continuous Improvement

The fast-paced nature of fraud requires organizations to adopt a rapid-response mentality. Fraud teams should be constantly testing new signal providers, updating thresholds, and optimizing decisioning criteria.

Rapid response also involves having a prepared attack playbook – a set of pre-defined strategies and rules to quickly counteract fraud attacks when they occur. By monitoring early warning signals, such as spikes in fraudulent applications or behavioral anomalies, organizations can respond swiftly before significant losses occur.

Fraud detection is an ongoing process that requires continuous refinement. It’s important to regularly evaluate cut-offs, thresholds, and decisioning models to ensure they are optimized for both accuracy and customer experience.

Fraud red teams, similar to cybersecurity red teams, can help organizations test their defenses proactively. By simulating real fraud attempts, red teams can uncover weaknesses and provide actionable insights for improvement.

Strategy and Governance: Building the Foundation

While Hughes emphasized that the six core capabilities form the operational backbone of fraud prevention, strategy and governance are the glue that holds it all together.

According to Hughes, a common pitfall for many organizations is the tendency to justify fraud prevention investments based on potential loss reductions, rather than benchmarking their capabilities against industry best practices. As a best practice, fraud investments should be focused on maintaining best-in-class capabilities, even when losses are low.

Cross-functional collaboration is another critical aspect of effective strategy and governance. Teams across marketing, fraud tech, and operations should work together to manage the onboarding process, using a shared scorecard with metrics for fraud rates, customer friction, and costs. This collective approach ensures that all departments have a stake in optimizing the onboarding funnel, rather than operating in silos.

Enhancing Fraud Prevention with DataVisor

According to Hughes, a comprehensive, integrated approach to customer onboarding and fraud prevention is crucial today – one that leverages the six core capabilities and aligns strategy with best practices. A winning fraud strategy fosters cross-functional collaboration, so organizations can create a seamless onboarding experience that prevents fraud without negatively impacting the customer experience.

The full webinar can be viewed on-demand here. To learn more about how DataVisor can help you future-proof your onboarding process and enhance fraud prevention, visit our website or contact our team today.