Solving Transaction Monitoring Challenges with Entity-Centered Investigations

When it comes to fraud prevention, traditional transaction monitoring methods have primarily focused on individual transactions to identify suspicious or fraudulent activities. However, this approach has presented several challenges for fraud operations teams. 

‍

This article will explore several critical transaction monitoring problems and propose a different technique that centers bank fraud investigations around entities, offering a more effective solution.

‍

Let’s jump in.

‍

‍

‍

Transaction Monitoring Challenges

While transaction monitoring has evolved significantly since its inception in the late 80s, it has encountered its fair share of obstacles. 

‍

‍

Early Days of Transaction Monitoring: Manual Reviews & High Operational Costs

Initially, financial institutions relied on manual reviews conducted by human analysts to identify suspicious activities. In addition, these early systems implemented rule-based approaches, where predefined rules were employed to flag transactions that deviated from standard patterns. However, these systems had scalability, accuracy, and efficiency limitations.

‍

The mass adoption of credit and debit cards, online shopping, and smartphone usage led to a surge in transaction volumes. Unfortunately, the early rule-based engines struggled to detect sophisticated fraud schemes, forcing the policy and rule writers to create extensive sets of rules.

‍

Consequently, the system generated many alerts, overwhelming fraud fighters and causing operational fatigue. Financial institutions faced the dilemma of deploying new rules to address emerging fraud patterns while lacking the capacity to handle the increased alert volumes. 

‍

This resulted in high operational costs and reduced efficacy due to many false positive alerts. And how would you justify more manpower when the agents review 80%-85% of false alerts?

‍

‍

Modern Days of Transaction Monitoring: AI, ML Multichannel Products

Present-day transaction monitoring systems leverage AI and ML algorithms to detect anomalies and patterns indicative of suspicious transactions. These algorithms can learn from historical data, adapt to evolving fraud patterns, and continually improve their detection capabilities. 

‍

They utilize a combination of supervised and unsupervised learning techniques to identify known patterns of fraud as well as previously unseen suspicious behaviors. However, even with the integration of advanced technologies, modern transaction monitoring still operates at the transactional level, posing challenges for fraud analysts.

‍

This is because financial institutions now offer products through multiple channels, creating an extremely complex environment for fraud operations teams to navigate. Challenges include: 

‍

• Team specialization
• Monitoring fraud schemes involving multiple channels 
• Lack of visibility into the decision-making process of AI/ML models 
• The need for more effective communication between teams 
• Channel prioritization 

‍

While balancing the detection of suspicious activities while minimizing false positives remains a critical challenge in transaction monitoring, there is a different approach that can be adopted to help alleviate these challenges.

‍

‍

The Entity-Centered Investigations Approach

Financial institutions must adopt an entity-centered approach to investigations. This will allow them to address the ongoing challenge of detecting suspicious activities while minimizing false positives. 

‍

Rather than reviewing alerts individually, fraud investigators should holistically assess entities, considering their historical alerts and activity across all channels. This approach provides a comprehensive view of entities, their behavior over time, and their interactions with different channels and other entities on the platform.

‍

The entity-centered investigation approach significantly impacts alert volumes by allowing investigators to resolve multiple alerts simultaneously instead of addressing each alert individually. This approach enhances efficiency and reduces fatigue associated with queue work.

‍

Moreover, adopting this approach eliminates the need for internal specializations within the fraud detection teams. Financial institutions can train their agents to work on the same queues and equip them with the skills to detect and handle various types of fraud.

By adopting an entity-centered investigation approach, agents gain the ability to identify patterns of behavior and potential vulnerabilities that might otherwise go unnoticed. As a result, financial institutions can address the root causes of fraud rather than merely treat the symptoms, leading to more effective prevention efforts.

‍

Furthermore, an entity-centered investigation approach unlocks the potential for network analysis investigations. Fraudsters often employ complex schemes involving multiple entities to conceal their tracks. By examining the entire network of entities involved, agents can uncover relationships between these entities and trace the flow of funds through the network.

‍

This approach reveals hidden connections and behavior patterns that would otherwise be overlooked in a transaction-focused analysis.

‍

‍

How Unit21 Helps

With that in mind, the product team at Unit21 set a goal of helping fraud operations agents make the best decisions in the shortest amount of time.

With two new features: “Group Alerts By Entity” and “Entity Page,” We’re making sure that investigating an entity in Unit21 will provide full visibility into historical activity, a clear vision of prior and current risk, and easy access to network analysis and bulk actioning.

‍

‍Schedule a demo here if you'd like to see Unit21's risk and compliance infrastructure in action.

‍

‍

‍