The Finance Director of a Financial Services company was struggling to keep Graduate Entrants. Even though Graduates were joining a growing business with a good starting salary and career prospects (CIMA sponsorship and later progression), the daily monotony of Accounts Payable was putting Graduates off and triggering resignations. Accordingly, it became clear that this mundane but important accounting activity was a good candidate for automation.
Like many other businesses, the accounts payable process for this company involved many manual steps, including:
- Opening supplier emails with PDF invoice attachments
- Opening the accounting application, Sage50, and accompanying spreadsheets
- Saving PDF documents to a supplier folder by month and supplier
- Copy/paste data from the PDF and enter data into Sage50
- Gain authority to pay from the internal departmental owner
All in all, the process was repetitive, mundane, and also prone to error.
Although Accounts Payable is a common application of RPA, it was only important to prove the robot could accurately read the PDF Supplier Invoice and select the correct fields. Accordingly, a Proof-Of-Concept (POC) for OCR was developed to prove the Bot could read the relevant data from supplier invoices such as Supplier name, Invoice Date, Invoice Number and Subtotal.
The capability was proven by integrating AWS Textract into the RPA workflow.
- Microsoft Outlook
- Sage 50
- Microsoft Excel
- AWS Textract (OCR)
Process Flow Map
RPA Bot Steps
Triggered manually from the Desktop
- Automatically login into Sage50
- Open Supplier PDF Invoices attached to email
- Read the PDF’s data accurately including Supplier, Date, Invoice# and Subtotal
- Saved PDF invoices to appropriate folders by month and supplier
- And input invoice data into Sage50 on the batch invoice screen
Process as % of Employees Time
Estimated Payback (RAAS)
Redeployment of Finance Clerks, and future Graduates, to less mundane, more demanding and valuable roles for the business.