Fraud Log Investigation




A leading South African bank

Project Overview

FraudLogs was designed to investigate unauthorized logins to bank clients’ accounts. Earlier, a bank officer collected and attached information about the client’s actions manually. Any delay or absence of the relative human resource impacted the investigation timelines.

FraudLogs provides a bot that collects information about recent client’s actions from different systems.  The information is imported in the fraud investigation system as attachments.

Project Execution

IBA Group acted as a subcontractor, the contractor being Diverse I.T. Connections. The development was done from Minsk, Belarus with short planning sessions at the customer’s site. The IBA Group’s team has deep and long-lasting relations with the bank’s teams. This helped us find real-life data that were needed to perform high quality testing of the solutions.  

The project lifecycle included the following stages: business analysis, requirement gathering, application development, creation of business processes (integration), and production. According to Agile methodology, the development stage was broken down to bi-weekly development phases with customer feedback loops.

IBA Group utilized numerous preloaded process templates, pre-built robotic, cognitive, and exception-handling components for the projects. In addition, IBA Group suggested and implemented the idea that instead of processing each record individually, process up to 20 records per bot run, which saved the record processing time manifold.  

The project team developed bots for various application types, including Web, Java Swing, 3270 UI, Windows application, Unix Shell, SOAP/REST services, and other. The following technologies were applied: Java, Groovy, JavaScript, Maven, Git, and XML. The applications combine human and bot tasks and are designed to function on-premise or in the cloud infrastructure.

Project Results

As a result of the RPA project implemented by IBA Group, the processes of fraudulent log investigation were automated. More than 500 cases are processed within three hours and up to 20 cases are handled in parallel, increasing the performance fourfold. The bottlenecks of fraud cases were removed, as the processing time for fraudulent logs dropped from 24 hours to 5 minutes.

Collecting and attaching logs to fraud investigation cases does not cause bottlenecks in the investigation process, and less back and forth communication is required, resulting in better resource utilization and higher customer satisfaction.

The projects solved the problems related to long response times and human errors. 
The solutions did not only automate the repetitive operations. They digitized complex business processes, providing machine learning-powered cognitive automation. In addition, individual operations were unified, providing workforce orchestration. IBA Group was able to re-package complex business processes and transform them into easy-to-use solutions that can be applied to multiple processes and products.

In addition, IBA introduced the best software development practices in the RPA processes. IBA Group built a centralized warehouse to keep common code that is used for different robotic processes.

The future plans include automation of other business processes at the bank using the framework that IBA Group created when implementing the project FraudLogs.


© 1993-2018 IBA Group a.s.