Implementing RPA in financial services organizations takes careful consideration and diligent planning. These 10 tips will help guide you.
Most success stories around Robotic Process Automation (RPA) are larger than life: cost savings of 60 to 80 percent, projected ROI in three digits, turnaround times that go from days to minutes and so on.
The financial industry is understandably excited. According to a recent global study, 33 percent of major enterprises, and 44 percent of banking and financial services companies, have committed serious investments in RPA over the next two years.
But there is a flip side that no one’s talking about, which is that 30 to 50 percent of initial RPA implementations fail. A little probing reveals that business leaders are unhappy about the time it takes to reduce cost, operations teams feel that the productivity gains are less than impressive, and with bots failing frequently, the technology and IT staff are clearly overworked.
Why RPA fails sometimes
In the financial services industry and in other industries, the reasons for the underwhelming performance of robotic process automation range from lack of scalability and excessive reliance on IT to inadequate preparation to unrealistic expectations and short-term thinking. Also, a big problem is that enterprises often choose the wrong processes for automation.
Implementing RPA in large financial services organizations takes careful consideration and diligent planning. Below is a list of top 10 tips for selecting processes for robotic automation to help guide financial services organizations.
Assess processes consistently across the enterprise: Instead of shopping for use cases at random, enterprises should employ a standardized, top-down process hierarchy to assess the right candidates for automation.
Estimate effort accurately: Enterprises should base their business case for RPA on current and accurate data on the effort, number of steps, and manpower needed to complete different processes. At present, this assessment relies significantly on inputs from process subject matter experts (SMEs). It would also benefit by factoring the results of time and motion studies, activity based costing, and average handling time/ effort analysis.
Minimize process variation: A number of RPA implementations overshoot budgets and timelines because the processes vary so much. Financial services organizations should address that by mitigating the usual causes, such as multiplicity of systems and interfaces, variation in the extent of process centralization across different country operations, proliferation of country-specific processes shaped by legal or local requirements, and lack of standardization in the formats and processes employed by different vendors.
Build a solid business case: Broad brushstrokes and gut feel have no place in process selection. Rather, the decision should be grounded in a comprehensive business case, which takes all costs, benefits, complexities (e.g. number of systems, data types etc.) and risks (e.g. legal, regulatory etc.) into account.
Factor in other strategic technology and business programs: Large organizations have ongoing business and technology initiatives, such as cloud adoption or outsourcing, which might impact some of the processes selected for robotic automation. It is vital to take that into consideration before going ahead with RPA.
Think end-to-end: In RPA, the sum is truly greater than the parts. When assessment is conducted in silos, a process that occurs in or draws data from two or more functions is in effect treated as two (or more) separate processes. This is sub-optimal from an automation standpoint.
Balance sheet control
Here an illustration may be useful: A key process called balance sheet control is part of the financial controller’s domain. However, to make up the balance sheet account totals, the process needs transaction data from other functions, which is sent to the finance team for reconciliation. An organization trying to automate ledger “proofing” without an end-to-end view will automate only the activities of the finance function, leaving all the feeder activities in the other functions to be performed manually or automated in a separate exercise.
Measure complexity, so you can manage it: Since no two RPA use cases are of the same complexity, enterprises cannot base their calculations on general assumptions. They should create a continuum of activities that are graded by complexity to assess feasibility and decide priorities.
Include solution analysis in process assessment: After figuring out the complexity and nature of a use case, the organization should analyze the various solutions available to identify an appropriate toolset. This has a direct bearing on the business case since implementation costs and timelines vary by the tool.
Review and redesign: Walking through a process will help to identify its pain points. The enterprise should then try to determine the “dominant root cause” of each (for instance, is it policy and procedure or organization design, or poor error detection, or something else?) before proceeding to redesign the post RPA-process.
Perform assessment in “sprint” mode: Often, organizations try to complete the assessment and identification of use cases in a single iteration, which creates a long gestation period until implementation by which time the opportunity might have staled or the process itself might have changed. There is also a risk of misinterpreting or dropping information required for the implementation. The ideal approach is to take up RPA in “agile mode” – maintaining a process backlog that allows the development team to plan for the next few sprints as well as take into account any feedback from the previous iteration.
Following as many of these tips as possible, financial services organizations should identify the right automation partner to see them to their goals.
Here’s an example of a large financial services company, which turned around an RPA implementation that was headed nowhere. The company’s shared services unit had run into serious problems with its robotics-led initiative, which had not operationalized even a single process yet. The reasons for this included wrong selection of processes, unrealistic cost saving expectations and a team that was totally unprepared for the job. The company then with some help from external partner set out to define the right RPA demand generation model, compute the business case, design the post-implementation operating model and flows, and execute the project in agile mode.
Thanks to this exercise, the unit now has a clear line of sight into RPA backlog and a sound business case. Having found the right RPA solution, it is now implementing several RPA projects and is on the way to meeting its cost saving targets.