Responsive Menu

Announe IconHyper Automation using MuleSoft Automation Suite | June 21th


A $1.5MM Lesson in Planning for Scalability

The Mistake of Thinking Small with RPA Projects

One of the biggest roadblocks to a successful RPA implementation is a lack of planning for RPA scalability RPA scalability within the business. Executives who see automation as optional tend to have a “start small” attitude towards RPA adoption, taking a piecemeal approach that often results in a series of failed pilot projects. 2 It’s important to remember that just as there is no “starting small” when implementing ERP software or any other big software transformations, RPA and AI technologies require big thinking about collective incremental gains attainable from these technologies and the value they can deliver well into the future. Executives must take a holistic view of their automation program while they’re still in the beginning phases and start planning for scalability early on. This paper will discuss an enhancement Accelirate made to one of our Financial Client’s existing automations to help the solution scale to match the client’s rate of business growth.

Business Scenario: Enhancing an Existing KYC Automation for a Large Financial Client

The client is a full-service bank based in North America with more than 150 branches and over $25 billion in assets. The company’s services include personal and business banking solutions, cash management services, business loans, and commercial real estate financing solutions. Accelirate’s engagement with the client began in 2021. The client was already using RPA to automate its KYC Evaluations process, but was seeking to improve the functionality of the bot to handle increased workloads. The automation had been implemented by a previous vendor, and the team had not taken scalability into account, causing the bot to run less and less efficiently as workload volumes increased.

Process Overview

The client performs Customer Evaluations on a quarterly basis. This KYC process begins with the bank receiving a file containing customer financial data including financial status, transaction history, credit history, and bankruptcy. Each record contains about 600 total rows of data per customer. The bot goes through these records, gathers the information, and places it into a single Excel spreadsheet to display all of the data in a consolidated view, allowing an analyst to manipulate the data and finalize their review of the customer. When this process was originally automated in 2019, the average volume the Bank received would range from 50 to 100 records per file. As the business saw rapid growth over the next two years, the number increased to 1,000 records per file on average in 2020, then to nearly 1,500 in 2021. With such a large increase in volume, the Bank was seeking to improve the bot’s processing power and overall functionality. Since the solution was not developed to be able to handle a growing workload in its original iteration, the client was experiencing slow processing times, errors, and pauses while running it.

Process Issues and Challenges

  • Poor Workload Handling – The bot is considerably large and was originally developed to run end to end, causing problems when attempting to process larger workloads. If the bot were to encounter an issue with an individual customer record at any point in the process, it would not be able to skip the record and move on to the others. Instead, the bot would need to restart the entire file from the beginning, creating significant productivity losses and major disruptions to the analysts who need to perform their reviews on time. Furthermore, because this bot was created to run end to end, it could only be run during work hours when developers were available to handle bot errors and system issues.
  • Low Efficiency – Since the Bank performs routine system updates in the evenings, the bot could only run specific applications during work hours.
  • Poor Functionality – The pivot tables that the bot would create were clunky and difficult for the KYC analysts to review.
  • Lack of Documentation – The previous vendor, when initially automating this process, had not provided the client with any documentation. By the time Accelirate began our engagement with the client, no one at the company had performed the process manually in over a year, leaving a huge gap in process knowledge.

Solution for Enhancing the Automation

We began enhancing this automation by having numerous in-depth discussions with the business teams to create thorough process documentation on the current process and the enhancement. This was needed to establish clear expectations and a deeper understanding of the process, given the lack of documentation available when we began the engagement. This included the PDD and SDD, as well as a detailed operational handbook mapping out potential errors and exceptions, runtimes, and contact details. Creating centralized and up-to-date documentation helped ensure knowledge continuity and process clarity.
The next step we took to improve this process was implementing the new KYC UiPath template. This template would allow the bot to create a richer and more visual Excel sheet for the analysts to review, replacing the previous clunky pivot tables with seasonality dashboards and expected behaviors aggregated on a customer level. To improve speed and processing capacity, we split the bot into two parts: a dispatcher, which extracts all data necessary for the template; and a performer, which inputs and filters all the data that is appropriate on the template. This two-part solution provided far greater control over the process while solving the issue of limited workload capacity. The dispatcher runs during the workday when systems aren’t being updated, and the performer is able to run at all hours and retry individual records without delaying the rest of the file from processing.

Enhancement Results

With the enhancement in place, the Evaluations process now runs monthly instead of quarterly. The business was able to multiply the volume of Evaluations processed from 150 per month on average, to 1,000 reviews per month – increasing capacity by 566%. Now able to handle the exponential increase in volume, the client saw an ROI of $1.5 million within 8 months of the enhancement going into production. The client also saved over 200 hours per month previously spent on restarting or reconfiguring the bot and doing rework associated with bot errors. This enhancement also streamlined the analyst review process, with KYC analysts now spending 17 minutes on average to finalize their review of a single customer, which previously took 25 minutes on average – a 32% decrease in cycle time, due to the clear and concise way the bot now presents the customer data.

The $1.5 Million Question: Are You Taking a Broad View of RPA Scalability?

Using RPA to automate business processes is becoming a priority for enterprises across nearly every industry to streamline work and save time and resources, but failing to plan how you’ll scale the technology can limit your return on investment. When getting started with RPA, businesses must develop a strategic and tactical plan that covers automation sustainability from the pilot stage all the way to enterprise-wide adoption. This includes thinking about how the technology could evolve in your business – whether by handling increased workload, expanding usage to other departments, or integrating new applications and technologies; taking these factors into consideration early on is essential to RPA’s long-term success.
The enhancement discussed in this case study was necessary in order to help the client meet the increased need for Customer Evaluations they were experiencing, but this roadblock could have been avoided if the original vendor had built the solution with scalability in mind. Because they never anticipated scaling this process, the bot was not developed with the functionality to handle increased volume. This lack of foresight nearly cost the business the success of their entire program. Organizations can avoid this critical misstep in their own RPA journeys by taking a broader view of RPA scalability and always keeping an enterprise-wide scope in mind.
In other words, think big from the start. In the age of hyperautomation, your company’s survival may depend on it.