Insurance Fraud Investigation using Linkurious

In the US alone, insurance fraud cost companies around 80 billion dollars each year. Being able to detect fraud schemes before the fraudsters have been able to access the funds they’re trying to steal is a major advantage for insurance companies and their customers. Nevertheless, the detection and investigation processes remain difficult for these institutions. Most of them simply lack the appropriate tools to detect complex fraud patterns that blend with normal user behavior and to investigate complex fraud cases.

Linkurious enables deep and efficient visual investigation of suspicious patterns in your data. Here we’ll take a look at property damage claims and how Linkurious can help analysts investigate fraudulent looking cases.

Using graphs to spot insurance fraud

Insurance fraud is rarely the work of an isolated individual. Fraudsters usually form complex networks that are difficult to detect for insurance and financial institutions. One of the most common techniques used by fraudsters is to forge fake identities, file several claims and cash the insurance checks. Creating fake identities requires to forge or usurp personal information like social security numbers (SSN), addresses, credit cards, etc. These pieces of information are then submitted by the fraudsters to the insurance companies as they become customers. Forging new information for each fake identity they create comes at a high cost for fraudsters. This is why they often recycle this data to create several fake identities.

insurance fraud network example

A graph approach can help us spot suspicious connections

The picture above shows two customers and what they are connected to. Each customer has a unique personal address, phone number and email but somehow they share the same SSN number, which is normally unique for each individual.

A graph approach makes it possible to spot suspicious fraud insurance patterns in large datasets.

Limits in current fraud investigation tools

Typically, insurance companies rely on relational databases (RDBMS) to store their customer data. RDBMS were designed in the 80’s to codify paper forms and tabular structures. They do this task very well and remain one of the best tools for storing and organizing data. Nevertheless, when it comes to querying and visualising important amounts of connected data, RDBMS do not perform well as they were not designed for this purpose. In these cases RDBMS often lack important features, are slow, not flexible in terms of modelisation and are unable to query the data in real time. All of these issues make them difficult to use for fraud, where analysts need to identify and investigate suspicious connections.

Linkurious makes it possible to overcome these issues. It’s easy to use interface simplify the job of analysts and allows data scientists or developers to leverage the power of graph databases like Neo4j. Detection of suspicious patterns at database level is greatly facilitated using the Cypher graph query language. Once these patterns are isolated, being able to visualise them instantly with Linkurious enables deep and efficient investigations.

Investigating suspicious looking patterns in data with Linkurious is fast and intuitive for analysts. Analysts with no particular technical backgrounds can thus carry out complex investigations.

How Linkurious helps analysts investigate insurance fraud cases

Linkurious makes it possible to visualize and understand how different entities are connected. It gives analysts the capability to quickly distinguish between a real insurance fraud case and non-relevant alerts, saving them precious time.

Here is what a  ‘normal’ insurance customer looks like in Linkurious. The customer is connected to  a single claim, SSN, phone number, email address and address. We can see that he doesn’t share any of his personal details with other existing customers or known fraudsters. A lawyer and an evaluator are connected to the claim.

Normal customer linkurious

A normal looking customer in Linkurious.

In a simple glimpse, it’s possible to understand that 8 different entities are connected and assess that the situation seems normal. As an analyst, this picture can be used as a “template” that will make identifying fraud cases easier.

Now, let’s look at two other customers and their claims. The visualization is very different from the normal template we just saw. This situation should thus directly attract the attention of the analysts.

The fact that the two customers share an address and the same last name means that we are probably looking at a couple. The situation seems normal. The investigator can choose to dismiss the case, put a low priority investigation if a doubt remains or chose to keep an eye on it to track interactions with the nodes and react quickly if something suspicious happens.

False Positive Linkurious

Visualising a false-positive that would have otherwise been flagged as suspicious

Linkurious helps save precious resources and avoid customer dissatisfaction.

Let’s look at three new customers and the property damage claims they are connected to. These claims are under the supervision of the same lawyer and the same evaluator. Curiously the two customers that instigated the three property damage claims, John Piggyback and Werner Stiedemann, are both linked to a third existing customer, Paula Smith. Piggyback has the same phone number and Stiedemann the same email address as her. This is an abnormal situation as neither of the two share her name or address.

insurance fraud linkurious

Visualizing a potential insurance fraud case.

This situation is very likely to be an insurance fraud. The fraud investigators can block the transactions for all three cases and launch further investigations and/or legal proceedings.


Linkurious makes it possible to leverage the power of graph databases via a simple interface. Data scientists and developers can design queries to spot potential fraud cases using the pattern-matching capabilities of Neo4j. Analysts focus on the visual investigation of the suspicious cases.

The visualization capabilities of Linkurious means fraud analysts can quickly evaluate if the cases identified by the algorithms are false positives or serious cases. Reviewing these cases with Linkurious before blocking any transaction can be very useful in order not to treat genuine clients like potential criminals and negatively impact their customer experience. Linkurious also makes it easier for investigators to dismantle entire fraud networks at once by tracing their entire ramifications and not forget a key element. Finally, using the collaboration tools to  communicate with other analysts and with the authorities eases the whole investigation and prosecution process, making it one of the most complete graph database visualisation solutions out there.

Want to explore and understand your graph data? Simply try the demo of Linkurious!

Tags: , , , , , , , , ,

No comments yet.

Leave a Reply

Subscribe to our newsletter

Receive monthly updates about Linkurious and graph related news.

Follow us