How AI and network analytics transformed fraud detection at DB Insurance

Insurance fraud in Korea was evolving – from being committed by opportunistic individuals to organized collusion networks. Certain clinics, brokers and repair shops were colluding in plain sight across provinces through organized schemes that were hidden among millions of legitimate claims.

Investigators spent hours per case manually tracing fraud through spreadsheets and drawing network diagrams by hand. They caught individuals. However, with 10 million customers generating millions of claims, the patterns existed on a scale too vast for human investigators to discern.

Korean insurance company DB Insurance knew something was amiss; they just couldn't see it. Yet.

Making invisible fraud visible with SAS Viya

DB Insurance partnered with SAS to develop Korea's first AI-powered fraud detection network, enabling investigators to identify invisible fraud connections as visible patterns on their dashboards. Network analytics and AI became the foundation for protecting honest policyholders through responsible and trustworthy data and AI.

The DB T-System, built on SAS Viya, unified operational and informational data from across the business, bringing decades of policy, claims and customer information together on a single platform. What had once been siloed became interconnected. The system began analyzing patterns across millions of claims, identifying hidden relationships between accident participants and revealing where collusion was likely taking place.

By continuously mapping relationships across DB Insurance's entire customer base, every claim, every hospital visit, every repair became a data point in a vast network. The system could now track how and where accidents occurred, which hospitals patients visited, how long they stayed, and which repair shops they used. For the first time, investigators could see an entire ecosystem of behavior – and how seemingly unrelated events were linked together, as well as how people and organizations overlapped. Patterns of deceit emerged from the noise.

The AI behind the system learned from every case it processed. With each investigation, it refined its ability to distinguish coincidence from collusion. Fraud rings that might once have operated undetected for months were now caught in hours. The DB T-System didn’t just detect anomalies; it categorized fraud types, assigned risk scores and prioritized the most urgent cases, giving investigators the clarity to focus where it mattered most.

DB T-System in action

While three small claims in different cities appeared unrelated to local police, the DB T-System revealed connections: All had the same shop owner, the same at-fault parties and the same vehicles routed to the same repair shops. A fraud network that would have taken hours to map manually appeared in two minutes.

"We turned it on, and just like that, dozens of cases lit up revealing invisible fraud connections," explains Kim Chun Hwan, Department Manager of the Life & Fire Claims Team at DB Insurance. "By analyzing comprehensive accident histories and insurance claims data, abnormal signs were detected.”

This wasn't an isolated success. As investigators began processing cases through the new system, the operational transformation became undeniable.

From reactive to proactive fraud prevention

The result? Analysis time dropped from hours to two minutes. Detection accuracy improved by 99%. The team now processes 30 times more cases than before.

But the operational shift runs deeper than efficiency metrics. "This was the turning point," Kim reflects. "We'd stopped reacting to fraud and started preventing its spread."

Today, DB Insurance no longer waits for fraud to surface, it anticipates it. The DB T-System acts as an early warning network, detecting patterns before losses occur. Investigators move from case to case with complete context, supported by a platform that learns, adapts and improves with every claim.

For policyholders, the result is fairer premiums and faster claims resolution. For investigators, it means a clearer mission and better tools to protect families and businesses. And for DB Insurance, it represents a new standard in fraud prevention – an intelligent system built not just to catch wrongdoing, but to protect trust itself.