Skip to Content

Transforming anti-money laundering through Robotic process automation


Robotic process automation (RPA) is gaining popularity by leaps and bounds. This is confirmed by the findings of the World Quality Report 2018 (WQR 2018) where 54% of respondents have stated that this is an important automation technology that is important in 2018. We have seen a proliferation of robotic process automation techniques in the AML space as well as where there is a dire need to comply with these regulations.

Regulatory compliance is expensive, with average annual spending amounting to $48 million, according to a Forbes survey. Technology and automation play a key role in AML compliance for key processes such as risk assessment automation, Know your Customer (KYC), and customer and transaction screening.

AML investigators screening customer and transactions have an important, but time-consuming job. They perform periodic reviews on high-risk accounts and conduct investigations using transaction histories via several parameters. Looming deadlines and volumes of information make addressing all factors for each investigation quite challenging. On the other hand, failure to consider key factors could result in incomplete analysis.

One of the banks we work with faced such a challenge. It had an extremely time-consuming process for monitoring and evaluating transactions for suspicious activities. The bank’s AML investigation team performed these investigations manually on domestic and international transactions through monitoring alerts. Once an alert was generated, AML investigators had to analyze the transactions on the incoming checks by manually retrieving checks from systems and then uploading the checks to audit systems to support their investigation.

This process was a key candidate for robotic process automation for two reasons: to achieve operational efficiency and to enable investigators to work with a larger data set to improve the quality of the investigation. Our solution involved designing assisted and unassisted robots that were able to automate the following activities:

1. Every time an alert was generated, three months of transactions check images were automatically downloaded for investigation.
2. On an average, this meant 38,000–40,000 checks monthly – a huge increase from the approximately 8,000 checks that were previously investigated manually.
3. This automated process runs for 24 hours to retrieve images in advance so that investigators can start their investigation immediately.
4. The process automatically uploads images to the audit compliance system, eliminating repetitive download of checks for the client.

This solution has reduced effort spent by each investigator by 20%. More importantly, it has improved the quality of the validation process by enabling the investigators to work with a larger data set of checks.

It was extremely useful during a major investigation, where it was necessary to evaluate months of data for targeted accounts that had processed several thousand checks. It would have been impossible to manually retrieve these checks. With the help of the RPA retrieval, the investigators were able to download twenty months of data, leading to the discovery of suspicious patterns that were at the heart of this investigation in just ten days.

RPA presents a great opportunity for banks to automate their AML compliance, processes, reducing speed to market and improving the quality of investigations. Discover the power of robotic process automation by intelligently selecting the right business processes and identifying the appropriate use case.

Co-authored by Avik Chatterjee and Venkatesh Ballani from Capgemini Financial Services