Introduction

Irregularities and exceptions are an inevitable outcome of the reconciliation process. The research and remediation of these exceptions can be a time consuming and laborious process, with unresolved exceptions adding to an organizations financial and operational risks. In a world of high transaction volumes if organizations cannot automate this process, their operations will never scale to meet increasing demands.

The Client

The client, a tier one global bank needed help automating the research and remediation of reconciliation exceptions arising from their credit card business. Processing an ever-increasing number of transactions per day meant teams struggled to scale their operations efficiently and effectively.

The Need

Within the reconciliation data chain, the client needed a solution that could automate the manual break remediation processes:

• Exception Classification: The client needed to understand the root cause of each of the exceptions they were seeing.
  • What was the reason for the exception?
  • Who or what was at fault?
  • What do they need to do to resolve the issue?
• Automated Resolution: Once the root cause of the exceptions was understood, a targeted resolution for each was sought.
• Scalable & Adaptable: The process needed to be scalable to grow with the client’s business, whilst also adapting to the changing business landscape.
• Observable: Internal and external controls required look through into the how these exceptions were being classified and why specific downstream actions were taken.

Selecting EZOPS ARO®

EZOPS ARO was selected to address these challenges due to its track record in the reconciliation space, in combination with its advanced AI enabled data control capabilities. The platform tackled this problem using:

• Curie Predict: Supervised Machine Learning models were trained on historic reconciliation exceptions. Learning from this data, it was able to predict the root cause and required resolution of new exceptions generated during the reconciliation process with confidence intervals assigned to each prediction.  
• Intelligent Process Automation (IPA):  Based on predictions and confidence intervals provided by Curie, ARO was able to automatically resolve large numbers of these exceptions. Utilizing the IPA Framework, a number of steps within the resolution process were automated.Target data enrichment was performed, transactions were booked and amended, and additional matches were identified and transmitted back to source systems.
• Predication Explainers & Model Cards: Provided the client with the ability to stand over the process, satisfying audit, control and governance requirements.  
• AI Control Hub:  The integrated AI control hub allowed the client to monitor the accuracy of models over time with the ability to continuously retrain models to keep up to date with drift and change within the exception dataset.

The Outcome

By deploying EZOPS ARO the client:

• ‍Reduced Number of Exceptions: Over half of the Exceptions were automatically resolved with no human intervention required.
• Reduced Median time to Resolution of Exceptions: For the remainder of exceptions, the median time to resolution with reduced by 45%. This efficiency was gained due to a combination of:
  • Automated classification and data enrichment; putting additional data at the fingertips of users.
  • Increased amount of focus from operations team members due to the general decrease in exception numbers requiring their attention each day.
• Reduced Outstanding Break Value: The reduction in number of exceptions and median time to resolution resulted in an 80% reduction in the value of outstanding exceptions present at any one time.
• Reduced Error Count: Owing to high model accuracy, decreased exception volumes and increased attention allotted to outstanding breaks, errors in break processing reduced by 50%.
• Increased Transaction Throughput: Driven by the above, the client was able to scale their operations. Processing more transactions, more accurately, with reduced operational risk and a smaller footprint.

Conclusion

By using EZOPS ARO, the client was able to completely rethink how they managed their reconciliation exception management process. By harnessing the power of AI they were able to automate manual bottlenecks in their reconciliation process, allowing them to scale their business and reduce their operational risk. We invite you to explore how using AI EZOPS can harness your organizations data to provide a higher level of service, with reduced risks and lower cost.

Please contact us today to start the conversation.