Fraud investigations require knowledge, skill, and timeliness to be successful. Every investigator will tell you that they must access many data sources to begin a fraud investigation. All the facts of the case must be understood — the who, what, when, how, and where are basic requirements to understand each specific case. Investigators gather this information from various internal systems, public websites, social media websites, etc., and load it into a case management system. Any site or system that may contain data that will help determine what actually happened and how the fraud occurred is necessary to the investigation, and the data must be retained for future reference.
While accessing all these data sources is necessary to obtain the full picture, the process is very cumbersome and time-consuming, as is entering all the data into case management. An executive from a very large financial institution recently stated that his staff spend 80% of their time aggregating data to review and only 20% of their time analyzing that data. Robotic process automation (RPA) can help reverse those percentages.
As the rate of fraud attacks accelerates, investigators must often juggle heavier caseloads. Time is of the essence in fighting fraud, so any delays that occur before working on new cases can be detrimental to the goal of recouping stolen funds. With heavy caseloads and the time it takes to accumulate all the necessary data, even experienced investigators may not be able to gain an understanding of newly-assigned cases until it is too late to recover funds.
RPA Heightens Investigation Speed
Using RPA to automate many of the investigator's manual tasks can dramatically improve investigative results. Since time often equates to money in fighting fraud, RPA processes can assemble the facts of the case so an investigator isn't spending precious minutes looking for data. Once RPA has the data and prefills it into the appropriate file in the case management system, the investigator can quickly review the case to understand the events that unfolded. Hopefully, he or she can also identify the location of stolen funds fast enough to request the funds be returned before they are transferred again or withdrawn. This speed is particularly beneficial when money is moving quickly and could leave the country (e.g., wire transfer fraud).
RPA can be particularly helpful with an investigation's customer and internal communications. When funds are stolen from a customer, the customer is anxious and wants updates; this often leads to customers calling for information and preventing the investigator from working all his or her assigned cases. Typically, the line of business that will suffer the fraud loss is also anxious to know what is happening and whether funds can be recovered, particularly in large-dollar cases.
Using RPA, updates can be sent to the appropriate internal stakeholders as well as to the customer, sharing what has happened, the current state of affairs, and what steps are planned in the case. Communication templates can be created, and extracted data from the case management system can be used to create these communications (which likely will be very different internally and externally), and instead of writing and sending a communication, all the investigator has to do is quickly review it and send. (Once a level of comfort is reached, the communications may be sent automatically without review.) Proactive communication will stop the bulk of incoming queries so the investigator can spend time doing his or her job; at the same time, proactive customer communication improves the satisfaction level of bank customers concerned about fraud on their accounts, and internal stakeholders are apprised of the status of cases promptly.
Addressing operational efficiency
The above are two examples of using RPA to improve fraud investigations, but many other opportunities exist; RPA can be used to automate almost any repetitive, time-consuming, manual tasks. Improvements in fraud investigations from using RPA to automate repetitive tasks can dramatically improve operational efficiency and the job satisfaction of investigators, potentially leading to less turnover. Retaining highly skilled investigators is essential to achieving fraud loss goals.
RPA is often discussed in the context of eliminating jobs, and as such, may carry some negative connotations. In the case of fraud investigations, many financial institutions have had to set fairly high dollar thresholds below which they do not work cases due to limited manpower. These cases are automatically charged off, and the bank incurs the resultant fraud losses. In this situation, RPA can initially be used to enable investigators to more effectively work existing heavy caseloads, allowing investigators to spend more of their time using their unique skill set to resolve cases and liaise with law enforcement to prosecute criminals.
This scenario will certainly lead to higher job satisfaction than is possible when investigators are overwhelmed with more cases than can be worked and have to spend so much time on repetitive, manual processes. Once caseloads become manageable thanks to RPA, financial institutions can consider lowering the dollar threshold for case investigations. Doing so should result in lower fraud losses, since many of these unexamined cases may lead to loss avoidance.