When looking for opportunities to automate business processes, Credigy strove to understand how automating a specific business process could also connect to another business process. “As automations are deployed in an intelligent manner and are ‘connectable,’” says Adams, “they will gradually grow to become fully automated processes end-to-end.”
“Along this RPA journey, our strategic goal has been to determine which business processes could benefit from automation,” says Rebecca Sims, Senior Business Systems Analyst at Credigy. “We started with the quick wins so we could expand our automation experience while determining where automation will aid us the most in achieving a strategic advantage using RPA.”
Working with RPA consultants, Credigy prioritized which business processes would benefit most from automation. “The purpose of the consultants was to teach our developers best practices and set us up for long term-success,” says Sims. “They did code reviews and helped show us the best ways to automate.”
“The consultants helped us discover the most viable processes for automation,” says Adams. “For example, if we do something once a month and it only takes five minutes to do it, it makes zero sense to put development effort into that process.”
Each year, Credigy processes thousands of data files received through email or an SFTP site. Because of the large volume of data coming in for analysis and evaluation, one of the first focus areas was to manage how data and files are moved—a laborious, time-intensive task. To move files, employees manually identified and opened files to validate the information. Next, they renamed documents according to uniform labeling conventions. The last step was to navigate to the correct repository to save the documents.
“Instead of having employees monitor data sources, we have a robot periodically check and name them according to a defined naming convention and place the files in locations where our system can pick them up and process them into our internal systems,” says Sims. “This helps us have an accurate understanding of account status so we can make informed decisions.”
Another opportunity for automation related to how Credigy processed vendor invoices. “We analyzed the set of vendors who submitted regular invoices and decided to automate the process into our invoice processing system,” says Sims. “The finance team can forward invoices attached via email to the RPA robot and the invoice is created automatically.”
Currently, Credigy processes 22 vendor invoices automatically. The end goal is to completely automate invoice processing when the automation framework is fully developed.
Due diligence—reviewing the related documents in connection with buying and selling financial assets—was another labor-intensive task that was ripe for automation. In many cases, documents were downloaded from third-party sites and included records for thousands of accounts at a time. Credigy deployed automated processes to scan account-based documents and scrape data into a Microsoft Excel file for reporting and decision making.
“For example, if we are potentially buying a portfolio of 10,000 consumer loans, part of due diligence is reviewing the loan files to make sure the original loan-related documentation is there and conforms to the applicable requirements,” says Adams.
“Previously, human beings would do that work, which is often time intensive and under a short deadline. But now, with RPA, we are doing 100% automated due diligence reviews for some deals, and it is a big win for us. Being able to complete the diligence process faster allows us to determine which issues need to be addressed and whether the deal that really works for us and our potential counterparty. It minimizes the time we spend on deals that are not going to work, helping us and our business partners focus on opportunities that make sense for all involved.”
To make sure Credigy derives maximum long-term value from its investment in RPA, the company is developing the framework for an ongoing RPA center of excellence. “Some companies start running RPA software and, a year later, might have 20-30 robots,” says Adams. “But then somebody asks, ‘What does this robot do?’ and nobody knows the rules. We will ensure that our processes are well documented and make metadata available so we can think about how each process fits into larger processes going forward.”