European merchant banker builds fraud detection platform and quadruples its annual cost savings

"The output from these reports has been shared with The Group Board and the results have been very well received and we are now being asked to share this work with other areas of the business. The investigation work from these reports will now become a regular monthly function. These reports are a great asset to this business. I trust you will all agree this is a very positive step forward for the business, thank you for all the hard work in getting this process off the ground so quickly and effectively, and to all those who contributed to the development of these, a great example of one team, one vision."
  • 900+ fraudulent transactions detected in 2 years post implementation
  • 300% increase in annual cost savings post implementation
  • USD 1 million saved in 2 years post implementation
Business Situation

This UK based financier had USD 700+ million on its loan books. The firm was losing substantial revenue annually due to inability to predict and catch fraudulent activities. The company had a complex data management process which was a major deterrent to building a fraud detection solution that can identify, analyze and predict fraudulent claims. This problem was hurting company’s growth margins and there was an immediate need to fix this problem.

Business Challenges

In principle, there were two major challenges that were the bottlenecks in implementing a robust fraud detection and prevention system –

  • Extremely complex process and technology landscape with regards to data management. Building a fraud detection system required completely reimagining the as-is processes and systems which was a big deal.
  • There was no single source of truth – multiple systems each with its own dataset. This environment made it difficult to select the right data for system ingestion and processing which ultimately predicts the fraudulent activities.
Birlasoft Solution

Birlasoft’s fraud detection solution comprised 21 customizable business rules that laid the foundation to predict and catch fraudulent activity. These business rules were developed after in-depth study of the existing processes to ensure their efficacy. In the case of a potentially fraudulent activity, the solution would pick outliers through various statistical analysis (pattern, ad-hoc, broker profiling) and provide detailed analysis to the governance team for further investigation.

Custom in-built reports were made available for tactical, operational and strategic analysis pertaining to stakeholders across the board. In nutshell, this solution helped this merchant banker reduce revenue leakage and improve its margins.

This solution was recognized by IT Europa and received European IT & Software Excellence Award 2014 for 'Best Vertical Solution of the Year'.