Case Study: Robotic Process Automation (RPA) for a Large Financial Client

Radiant helped one of its financial services clients with process automation using industry-leading RPA tools. The factors that lead to effective automation of repetitive business process are:

  • Understand and analyze business processes and their bottlenecks through lean six-sigma principles.
  • Document and map process steps by applying proper queuing methods.
  • Automate steps locally using relevant scripting utilities.
  • Comply with role-based access controls along with system and data controls.
  • Enable users to create their bots through script-less automation tools.

A large financial services organization is transforming its securitization process that impacts the entire secondary mortgage industry.  Three large financial services institutions and a governing body required quality reporting on a program’s progress to have a common understanding, track and monitor the issues and make informed decisions from time to time. It was a three-year-long, half-a-billion-dollar program.

The quality reporting data has to be obtained from all three participating organizations that use five different tools: Two instances of HP ALM, Rally Dev, IBM Clear Quest, and Service Now. These tools support their custom processes and configure very differently, which resulted in the following challenges:

  • How to deliver meaningful reports to all three organizations and the governing body?
  • How to generate consistent, reliable, and timely reports?
  • How to secure a way of reporting with proper controls in place so that underlying data is not exposed to the outside world? It can have a devastating impact on markets.
  • How to absorb the changes and generate new reports at more frequent intervals?

Our Approach:

Radiant solved the problem in the following systematic steps:

Process / Data Normalization – We studied and analyzed the existing business processes of 3 different organizations and how the tools are configured in support of them. Since the business processes are tied to their other internal supporting processes, they cannot be changed without impacting the business. Hence, we developed a mechanism to fetch data from feeding organizations into a shared repository, which involved writing VBA macros, SQL scripts, and Python Programs that made seamless connections to all the tools in different organizations.

Six-Sigma Analysis to understand process bottlenecks – We analyzed and timed the 32-step process that took 4.5 hours to generate one weekly report to eliminate bottlenecks and reduce error-prone areas.

Automated Manual processes – Error-prone tasks were further automated to improve the accuracy of reporting and reduce the time taken to generate the report by 70%

Implemented Blue Prism to support more processes, implement controls to processes, support more processes, and support more intervals.

Summary

The Quality reporting is consumed by 300 stakeholders, including executives and decision-makers, and is carried out every day since February 2017 without a break, including holidays and furloughs of the client. Reporting errors are a rare scenario. New reports can be developed, implemented, and stabilized within a week. The necessity to support the late-night and weekend reporting was significantly reduced. This mechanism was humming along with process optimization and the implementation of Blue Prism.

The critical factors to success come with an in-depth understanding of business processes, applying engineering methods to optimize the processes, using the right tools and technologies such as python, SQL scripts, VBA scripts to automate small steps in the processes, documenting business rules for manual steps, tuning to the culture of an organization, securing infrastructure, establishing controls, and the ability to configure the RPA tool so that new processes can be easily supported.

Benefits / Facts

  • Manual errors in reporting were eliminated. However, controls were introduced to avoid data errors, but some manual checks are still required for foolproof reporting.
  • The Manual Steps involved in the process were reduced to 3 from 32.
  • The number of people required to support the reporting was reduced by 80%, from 15 resources to 3.
  • The number of reports increased to 20 with the same bandwidth of 3 resources.
  • The need for off-peak support was eliminated.

We learned how to leverage the culture, technology stack, and infrastructure of an organization to implement Robotic Process Automation. Please contact us to discuss our RPA expertise.