How should we address the growing sophistication of fraud tactics, in an age where the risk of fraud is prevalent amidst the digital transformation for APAC businesses?

Graph techniques can analyze thousands of customer data points – and the crucial relationships between them – to deliver fraud alert scores in real time.

For instance, China Mobile – the world’s largest mobile service provider with more than 900 million subscribers – uses TigerGraph to detect phone-based fraud in real time by analyzing the calling patterns of pre-paid subscribers.

Subscribers are then alerted in real time of a potential fraudster call, with high-probability fraud calls being redirected to China Mobile’s call center for investigation.

CybersecAsia finds out from Chung Ho, General Manager, Asia Pacific, TigerGraph, how graph-based analytics and machine learning provide the performance needed for deep pattern analytics and real-time update capabilities for anti-fraud initiatives.

Do you observe a growing risk of fraud in Asia Pacific? If so, what factors contribute to this trend?

Chung Ho, General Manager, Asia Pacific, TigerGraph

Chung Ho (CH): According to a 2022 Lexis Nexis “True Cost of Fraud” report, across four key markets in APAC – namely Malaysia, Philippines, Singapore and Thailand – the average cost of every fraudulent transaction is 3.99x of the actual lost transaction value. This can be attributed to the prevalent and growing use of mobile apps and wallets to facilitate contactless and remote payments which has become more widespread.

The main challenge contributing to fraud losses is the inability to identify and authenticate synthetic identities through remote payment channels, using only attributes such as phone numbers, email addresses and devices. The lack of real-time risk assessment data and transaction tracking capabilities compound the difficulty of identity verification.

Another key issue contributing to fraud risk is the lack of internal adoption of a proper fraud management system within organizations. Many digital banks and financial institutions have yet to fully assimilate fraud prevention and cybersecurity measures into their core business operations. In fact, according to Lexis Nexis, the percentage of organizations with widespread use of artificial intelligence (AI) and machine learning (ML) models for fraud detection stands at 21%.

What challenges do organizations in the region face in combating fraud? Why is fighting fraud using traditional methods ineffective?

Fraudsters are getting more sophisticated over time, creating a network of synthetic identities combining legitimate information like social security or national identification number, name, phone number, and physical address. Detection becomes much harder with traditional solutions as each individual fraud account looks and behaves much like a legitimate account.

Detecting fraud requires going beyond individual account behavior, analyzing relationships among groups of accounts or entities over time, while often combining information from third party sources. Traditional fraud solutions are built and driven by relational databases.  This platform technology was never designed to address this challenge in addition to the explosion of data volume.

Organizations grapple with fraud detection due to the complexity and volume of data across multiple systems and locations. Because of its increasingly unstructured in nature, traditional analysis through relational databases is no longer adequate or robust enough to help business combat and mitigate fraud risk.

The importance of real-time risk assessment is paramount for stakeholders such as insurers and regulators, which require comprehensive risk materiality mapping that only big data analytics and systems can provide.

Why and how is it important to use graph analytics for real-time fraud detection, in terms of performance and cost?

Fraud detection is time-sensitive: every passing minute, hour, and day that fraud goes undetected results in increasing losses for organizations as well as for their customers. TigerGraph is purpose-built for real-time fraud detection to address this challenge.

Using deep-link analysis, graphs can manage big data, with thousands of customer data points, and analyze crucial relationships between them, to deliver fraud alerts in real-time.  Graph techniques can be used for fighting financial fraud by analyzing the links between people, phones, and bank accounts (among other things) to reveal indicators of fraudulent behavior, not only helping banks pinpoint suspicious activity but also giving them the tools to explain cases and current situations.

Graphs can perform at speed, especially compared to relational database solutions such as SQL. While banks have had fraud detection systems in place for years, graphs add speed and analytical depth to the equation. While SQL depends on bulky table joins, graph is less memory-intensive and able to handle a greater query load.

Banks are challenged with processing terabytes of data to find the needle in the haystack. The use of graph algorithms such as centrality or circle detection, are common queries to detect and identify fraudulent transactions. Using the same hardware settings with a mere data size of 500GB or more, TigerGraph has been proven with independent testing and validation that these important algorithms run at least 100x faster than its closest competitor.  These queries typically just time out on relational databases with the same hardware configurations.    

Additionally, graphs can be coupled with machine learning to further boost depth and accuracy of results. Graph is excellent at generating data for training ML systems because it can produce explainable models of what it has detected. Rather than simply giving something a score based on heuristics, the graph generates data on the links between different objects in the database, which can be fed into ML systems for further analysis.

Graph databases are good at showing results visually, and can be easily understood and explained. This, combined with the ability to explore contextual data, makes it an asset in fraud detection. Today, financial institutions must escalate their fraud detection efforts to stay one step ahead of fraudsters. In doing so, company leaders should work to identify, evaluate and implement specific technology that will help protect their customers — and more specifically, their customers’ hard-earned dollars.

What other benefits does TigerGraph offer to organizations?

CH: Besides real-time fraud detection, TigerGraph’s solutions combine graph analytics with machine learning that allow businesses, namely banks, to uncover data connections between existing “known fraud” credit card applications and new applications. This enables them to identify hard-to-spot patterns, expose fraud rings, and shut down fraudulent cards quickly.

Graph analytics also improves AML compliance: The practice of Know Your Customer (KYC) has become fundamental to banks and their ability to comply with complex anti-money laundering (AML) regulations and governance requirements. Perhaps no other banking use case requires more data-intensive pattern matching than an AML capability. Here, graph must seamlessly collect, analyze, and correlate layers-deep data to reveal complex relationships between individuals, organizations, and transactions. This is how financial services organizations unmask criminal activity and comply with evolving federal regulations.

Lastly, graph allows users to deploy dynamic credit risk assessment: With an estimated 26 million consumers not being tracked by FICO and other credit bureaus, risk assessment and monitoring have only grown more challenging. Determining whether a customer is qualified for a loan, a mortgage, or line of credit presents both risks and opportunities for financial institutions. These organizations must leverage all data at their disposal to make an informed, real-time decision regarding a customer’s creditworthiness in real time or risk losing market share. It also requires the ability to cull data from a variety of disparate third-party sources, normalize the data so it can be quickly analysed, and do so at a scale that doesn’t impede network performance.

The explosive volume and velocity of data along with the need to render real-time decisions has transformed the modern banking industry. Advanced graph analytics enables deeper insights, complementing existing BI technology and powering the next generation of artificial intelligence and machine learning applications. The banks and financial institutions who can secure a data advantage today will be the ones best positioned to thrive tomorrow.