The options offered by the Onwelo solution exceed those of existing anti-fraud systems, which are generally optimised for detecting a set of pre-defined abuse cases, such as data theft, financial fraud, and monitoring employees with special authorisations. Onwelo’s programme provides a far more complete and innovative approach to combatting abuses at companies.

The innovativeness of Onwelo’s platform lies in its capacity to rapidly acquire data for analysis from many sources with minimal input from the IT team and in its use of advanced statistical algorithms to classify the employee and customer behaviours being analysed. The Onwelo solution delivers very high performance levels and can be scaled up to match business requirements as they increase.

Using the Onwelo anti-fraud system a company abuse-detection analyst can:

• analyse abuse cases and abusers from a variety of perspectives (known as the 360˚ view)
• automate the detection of abuse scenarios that are already known about
• effectively seek and define new abuse scenarios
• raise the quality and efficiency of the process of investigating and clarifying identified abuses
• cut short an abuse that is losing the company money
• draw on information from all available sources

The fraud-detection process itself is based on an intuitive graphical interface that visualises information, and on the SQL language, which serves to analyse data. It offers a straightforward way to monitor known abuse scenarios either cyclically or in real time. Having analysed the data, and performed a vulnerability-to-abuse assessment of the business areas, it is possible to define an abuse-detection strategy in the programme and monitor its impact on the discrete business processes for a set time period, explained Dariusz Tarapata, Acting Chief Big Data Specialist at Onwelo.

The following modules make up the Onwelo platform: Data Scientist Sandbox – enables data analysis and the definition and verification of hypotheses; Information Repository – for storing data that is being analysed; Metadata Repository – for storing implemented rules and models; Alerts Repository – to manage identified alerts in the abuse-detection process; Case Management – to organise the detection, investigation and clarification of abuse cases.

Onwelo developed its software on the Apache Hadoop platform, which makes it possible to store and process very large datasets. The use of open source components means that implementing the Onwelo anti-fraud system is extremely cost effective. The system can be actuated either in the customer’s data centre or in the cloud.

“Our anti-fraud solution rests on experience gained in commercial projects conducted for large companies in the finance and telecommunications sectors. We usually begin our implementations by taking small, but precisely directed steps. This immediately gives customers a sense of the project’s added value and allows them to get a firmer grasp of the programme’s potential”, said Dariusz Tarapata.