Amy Hodler, Director, Analytics and AI Program Manager at Neo4j.


Data expert Amy Hodler examines how graph technology is reducing insurance fraud and providing customer insight at one of the world’s largest financial services companies

Financial services firms constantly have to fight financial criminals, but it is getting more demanding for organisations to identify and stop fraudulent activity at the scale it now occurs.

Traditional methods for monitoring fraud, such as setting up rules to examine deviations from normal purchasing patterns, use discrete data. This is useful for catching individual criminals acting alone, but this approach falls short when it comes to detecting fraud rings. Sophisticated criminals continuously alter their strategies to circumvent detection. They utilise synthetic accounts to carry out what appear to be unrelated activities by unconnected individuals. However, these activities are in fact well-coordinated and criminally linked.

The financial services sector needs a better way to follow the trail from one account to another to determine how activities that on the surface appear unrelated are in fact connected. This requires having a 360-degree view of the intricate fraud network to determine how suspicious events are linked.


Fraud detection with graphs

Graph database technology may be an invaluable tool in fighting fraud. In contrast to traditional relational databases, graphs not only interpret individual items of data, but also their relationships with one another. An increasing number of the world’s leading financial institutions are using graph databases to model and monitor data about customers, accounts, devices, locations and other attributes to identify fraudulent activity. Allianz, a multinational financial services company offering insurance products and services to 100 million customers in more than 70 countries, is one such.

As a truly customer-centric insurer, Allianz Benelux takes a zero-tolerance stance on fraud. As the subsidiary’s chief data and analytics officer, Sudaman Thoppan Mohanchandralal, explains, “We need to secure customers from risk – not just today, but into the future. We can only do that by having full insight into the risk environment and with an ability to predict it for our customers.”


Relational data model problems

Mohanchandralal’s colleague, Dr. Jan Doumen, strategic lead for Customer & Broker Information and Insights, agrees. “The best way to understand your customers and the risks they are exposed to on a daily basis is by storing, analysing and visualising them through connected data.

“Graph technology does this at scale, which means we no longer have to rely only on highly demanding, traditional relational technologies.”

Historically, building internal visualisations of suspicious behaviours with relational technology had been too demanding, Doumen confirms. The latest fraud countermeasures, such as network tracking, were too complex to build in a relational database. Sudaman calls this process a ‘2 by 2’ approach, where SQL database-style tables with rows and columns don’t offer the data connections fraud detection and prevention requires.

Working with a relational data model doesn’t allow the Allianz Benelux team to extract useful data on the fly. In contrast, graph technologies spot potentially fraudulent activity in Allianz Benelux’s ecosystem by disclosing concealed illicit connections. Bringing all the customer data into a graph database permits the Allianz Benelux anti-fraud team to reveal the risk exposures in a motor or household context.

“It is the combination of multi-node, multi-connection snapshots of customers and the much more efficient search possibilities coming from graph technology that we believed would revolutionise the way our internal business handles customers’ risks,” Doumen confirms.


Clear business benefit

Equally important for the Allianz Benelux team is having a 360-degree view of the customer. The Belelux operation has gone through a series of mergers and acquisitions and its customer data has become dispersed in separate silos, which has led to a number of operational inefficiencies.

“When we were able to get to a level with graphs to show colleagues this holistic view of a customer, it was so much easier for them to understand rather than through a table with rows and columns. This will enable them to personalise their services towards our customers,” Doumen adds.

Allianz Benelux’s success using the native graph approach has resulted in clear business benefits. Over the course of two years, €2 million of operational profit value was identified. Given the advantages realised with graphs, the Allianz Benelux team plans on offering the solution to other parts of the organisation.

Graph databases can future-proof an organisation’s fraud prevention initiatives by providing insight based on data relationships and connected intelligence. They can also unlock data silos and generate a more unified view of customers – helping you achieve full ‘customer-centricity’, as well as drive more revenue. Sounds well worth investigating.



Explore more