Unlocking value from data relationships to help insurance industry combat frauds

eFraudy is the Fraud Detection Platform based on data relationships that helps you to detect frauds involving multiple actors and before they occur!

A Revolutionary Approach based on Graph Analysis

With network analysis, we can be effective in detecting some of the most impactful attacks – even from organized rings – by discovering suspicious connections naturally.

On the contrary, doing connected analysis is extremely difficult with traditional technologies other than graph databases.

Gartner’s Layered Fraud Prevention Approach

According to Gartner, the best way to detect organized crime is the analysis of relationships. This goal is easy to achive using a native graph database instead of RDBMS, because it’s designed to be powerfull to find path and patterns.

eFraudy Features


The data are stored into a native graph database (Neo4j), a kind of schemaless database based on relationships.

The data model is very flexible and can be entirely adapted according to your needs and the information you want to store in.

Load your Data

With the eFraudy ETL component you can load your own data with ease.

Put this custom component in your data pipeline and fill the data model.

You can start with the shipped data model and increase or change it incrementally according to your specific needs and favourite information.

Graph Exploration

Using the “graph visualization” you can explore your data like never before, surfing the data like the web.

You don’t have to know exactly how the information are stored, which kind of relationships can be found or read a complicated table: the only thing you have to do is seek a starting point a discover new insights by following the links.


Discover Patterns

Working with patterns is easier and more human friendly than defining thresholds.

Start thinking at your data as a “graph”, a network of connected actors, such as “a Subject who DRIVES a Car  INVOLVED in an Accident”.

You can view the whole situation in a pretty way, easier to understand rather than separated tables.

Custom Rules

A rule is a fraud pattern.

A rule is a pattern of connections that can be detected on the data model.

A set of rules, working on our default model, are shipped with the eFraudy enterprise edition.

When new information come into the database, the rules are automatically applied and alarms occur if a candidate fraud is detected.

“Were people involved in this accident already involved in other ones, with different roles, but all together?” [one of the the simplest rules you can apply in real-time]

Not only thresholds. Solve most complicated cases where someone is connected to others subjects through very complex relationships.

Get new rules to improve the system.

You can buy them from the marketplace, ask for modifying existing ones or build them from scratch with our experts’ team.

Realtime, Fast & Scalable

Enterprise ready!

Find frauds when new data come in.

Apply a new rule to existing data to find, test and validate a new fraud pattern.

Apply a new rule to the entire database

  • As soon as you load it on eFraudy
  • When you change execution parameters
  • When you change the data model adding new information

Ready to manage huge data volumes.

Manage & Monitor

Keep track of fraud trends.

Understand how your company can protect itself using actual metrics.

Get the documentation related to the detected frauds to open a ticket and verify their consistency.