Combat Multichannel Fraud With a Strategic FRAML Approach

Determine the right balance of fraud losses, operational costs, and customer experience

Financial services organizations must strike a delicate balance between mitigating fraud losses, managing project and operational costs, and maintaining a positive customer experience. Overly stringent security measures can frustrate legitimate customers and increase investigation costs, while ineffective or inadequate controls can lead to significant financial losses. The challenge lies in implementing fraud mitigation strategies that are effective and that maintain the customer experience. Newer fraud solutions aim to balance these priorities through adaptive authentication and risk-based transaction monitoring.

Passkeys are a prime example of adaptive authentication. Passkeys can be easier for users to rely on because they use face detection or other biometric methods for authentication. Additionally, they use cryptographic keys instead of traditional passwords, which makes them more resistant to phishing and other cyberattacks. Private keys are stored on the user’s device, but public keys can be shared with banks, which helps enable secure communication. This authentication method eliminates the need for centralized password storage and reduces the risk of data breaches.

Select a flexible solution for an agile response to emerging fraud strategies

When selecting a fraud-combating solution, financial services organizations face a critical choice between robust and agile platforms. Established fraud technology vendors offer products with robust protection and deep analytics, but they might be slower to adapt. In contrast, newer, more agile solutions use AI and machine learning (ML) for rapid response to emerging threats, though model validation can be challenging.

Research tools assisted by generative AI or large language models (LLMs) can crawl through fraud news, the dark web, reports, laws, and publicly available data. Based on this research, synthetic data can be created for testing and sampling. To be effective, a solution should be able to systematically identify, develop, and maintain precise fraud detection algorithms. At the same time, it must not compromise on scalability, regulatory compliance, user experience, or security. Using an AI solution requires appropriate governance, regular model effectiveness reviews, and internal audits to keep pace with new fraud trends.

Source and integrate data from multiple channels

Integrating data across different channels including mobile, online, in-person, and phone banking provides a comprehensive view of customer behavior and helps detect anomalies more effectively. For example, a customer profile update from a new device followed by an unusual wire transfer might indicate account takeover. Hence, identifying, procuring, and validating respective “golden sources” of data are crucial actions for effective fraud control. Established fraud technology vendors offer prebuilt solutions that integrate with various common data sources. Additionally, supplementing internal data with external data sources can provide a comprehensive view for investigators. This process might include credit bureau data, public records, or specialized fraud databases. A few vendors also offer the ability to integrate consortium data from multiple financial services organizations to provide a broader view of fraud patterns and trends.

While obtaining data sources in real time can be challenging, the benefits often outweigh the effort. Real-time detection allows for quick intervention in potentially fraudulent transactions, which can reduce financial losses and increase customer trust. The choice between real-time and batch feeds depends on factors such as transaction type, data volume, latency needs, and infrastructure capabilities. A hybrid approach that uses real-time and batch processing offers a flexible solution for fraud mitigation.

As data constantly evolves, data integration architecture needs to be agile to adapt to new sources, reprocess historical data, and support AI and ML models.

Evaluate on-cloud versus on-premises solutions

Fraud technology often requires real-time data processing to quickly identify and mitigate fraudulent activities. Cloud-based fraud solutions offer several advantages, including scalability, flexibility, and cost efficiency. They can manage large volumes of high-speed data and analysis. Additionally, using cloud-native AI and ML functionalities can enhance the speed and accuracy of fraud detection and provide an advantage over traditional on-premises systems. Also, in an unforeseen circumstance, if a bank’s on-premises infrastructure is compromised, a cloud solution can support business continuity and keep critical fraud controls functional and secure.

On-premises solutions also have their merits. They offer greater control over data security and compliance because sensitive information remains within the organization’s infrastructure. On-premises solutions can be customized to meet specific regulatory requirements and organizational needs. They do not rely on cloud connectivity, so access is uninterrupted for fraud detection systems. Additionally, on-premises systems can be integrated with existing IT infrastructure to provide a cohesive and controlled environment for fraud prevention. While cloud solutions present potential risks related to third-party management and data breaches, on-premises solutions mitigate these concerns by maintaining direct oversight and control over all data and processes.

Streamline fraud and AML solutions

A key benefit of a FRAML approach is that it offers a streamlined suspicious activity report (SAR) filing process. Instead of filing separate SARs for overlapping activities, a FRAML approach consolidates insights from fraud and AML investigations into a unified SAR process, which improves report accuracy and consistency and helps organizations meet regulatory deadlines more efficiently.

Alert prioritization and streamlined investigation workflows are crucial for resolving higher alert volumes. Developing a scoring matrix to prioritize alerts and implementing a case management system that reflects the investigation team’s hierarchy with clear service-level agreement and escalation channels can enhance efficiency. Automated workflows help route alerts to the appropriate teams based on predefined criteria while reducing manual intervention and errors. Additionally, training and an always-available LLM agent trained on the organization’s knowledge base can improve the accuracy and speed of alert resolutions.

Take a multifaceted approach

Navigating the complex landscape of fraud requires a multifaceted approach. By carefully considering implementation factors, such as balanced controls, flexible solutions, multichannel data inputs, hybrid infrastructure, and a FRAML approach, organizations can create a more effective strategy to combat fraud. While challenges are significant, the potential benefits in terms of reduced financial losses and enhanced customer trust make fraud risk management a critical priority for financial services organizations.