Case Studies

OCR Invoice Processing

The Accounts Payable Department of a major manufacturing company is responsible for locating and uploading all incoming invoice information into the company’s SAP system. This process is currently extremely manual and consumes up to 75% of the AP team’s day during peak times. The Invoice Processing function is essential to the company’s operations as fulfilling orders quickly not only effects the company’s overall revenue but their customer satisfaction as well. The Accounts Payable Department is held responsible for the accuracy and efficiency of locating invoice emails and uploading more than 200 received invoices into SAP daily. At this volume, the AP team was struggling to keep up with the fluctuating number of invoices, especially as vendors order more items during certain months of the year and often place orders that need to be completed quickly. To automate this process in its entirety and fully optimize the results, Optical Character Recognition (OCR) was integrated within the automation to read the PDF invoices and enter the data accurately into SAP. Although this process was originally low-hanging fruit in terms of possible processes that were fit for automation, once automated, this task became one of the biggest wins for the company’s RPA program.

Automated Solution:

  1. Bot Checks if Email is Read: This process has two bots operating simultaneously to complete the process in a timely manner. To ensure that items are not completed twice, the first step the bot takes is checking if the other bot is currently reading an email with an attached invoice
  2. Bot Clears Workspace: Bot eliminates junk emails and unrelated items from invoice email inbox
  3. Bot Sorts Emails: Bot logs if an email has a PDF or other applicable file that can be converted to PDF attached and separates them accordingly
  4. Bot Downloads Non-PDFs: Bot downloads all applicable files such as XLS, TIF, and DOCX and converts them to PDF format
  5. Bot Attaches Relevant Info: Bot logs its actions as it attaches vendor name, email body message, email sender, and subject to the appropriate PDF file
  1. Bot Reads Native PDF Files: The Bot sorts the PDF’s into four categories: Native PDF with PO number, Native PDF without PO number, Non-Native PDF with PO, and Non-Native PDF without PO. The bot opens all the Native PDFs, looks for the PO number in the file, and separates the files into a matching folder
  2. Bot Uses OCR: Bot uses OCR to read scanned/hand-written Invoices (Non-Native PDFs) to locate the vendor PO number and sorts them into their respective Non-Native PO or Non-PO folders
  3. Bot Opens SAP: Bot uploads the sorted file folders into the matching category folders within SAP
  4. Bot Ensures File Accuracy: Bot uses SAP to determine if all files uploaded fully and correctly, if not, it creates an email and sends it to the AP Department notifying them to re-upload the file
  5. Bot Creates Log: Bot makes a log of all actions simultaneously so AP Department can track invoice progress as its processing and look back at the files on a later date to validate the process was complete and audit the invoice processing task in its entirety

This process eliminates the need for the Accounts Payable Department to dedicate 75% of their day to sorting and uploading invoices into SAP. With this new freedom, the AP Team is able to shift focus to receiving the matching invoice payments, following up with vendors who have not paid, assisting vendors to ensure smooth orders, and much more.