Processing financial data
Every working day, Singapore Management University's (SMU) Accounts Receivable team received financial data from its Treasury Management System (TMS). Before the information could be used for financial tracking and reporting, staff had to complete a series of manual processing tasks.
The process began with two separate emails: one containing a password and another containing a password-protected archive of financial records. Staff had to retrieve the password, extract the files, convert them into Excel format and fill in missing information.
The records also had to be matched against a separate journal entry report to obtain the corresponding accounting document types and reference numbers. Once completed, the processed information was consolidated into a master spreadsheet maintained for financial reporting.
Although the work followed an established procedure, it required approximately 60 minutes each working day. The number of files could vary, and staff had to ensure that all relevant records were processed correctly, with no missing information or overlooked files.
As a recurring task, the process consumed a considerable amount of time that could otherwise be spent reviewing financial information and addressing matters requiring further attention.
Making financial information available sooner
To improve the workflow, SMU developed an internal robotic process automation (RPA) solution that handled the repetitive preparation and consolidation of its daily financial records.
Staff could initiate the process by placing the received files into a designated folder. The automation would then retrieve the password, extract and convert the relevant files, populate missing data fields and match accounting references against the journal entry report.
The processed information was consolidated into the master spreadsheet, with the resulting records sent to the process owner for verification. The workflow also incorporated backups, organised storage of processed source files and notifications when required information was missing or could not be processed.
SMU's documented a reduction in processing time from 60 minutes to approximately 8 seconds. Including the remaining manual involvement, the workflow saved approximately 59 minutes per working day, equivalent to nearly five hours over a typical five-day working week.
During subsequent user acceptance testing in 2025, the automation completed its processing in approximately six seconds using sample real data. Tests also verified its ability to handle missing password files, identify ambiguous input files and correctly prepare the financial records.
For SMU's finance team, this meant that information which previously required an hour of preparation could be made available for review almost immediately.
