How OCR technology improves spend management.
Optical character recognition (ORC) technology allows us to connect physical evidence to the digital world. The multiple benefits of accounting automation have long been limited by the need to manage paper receipt and invoice collection, submission, and reconciliation. This remaining holdout for digitization is typical in any movement towards automation, according to consulting firm McKinsey & Company, who says that the last 20% of processes remaining to be automated are typically the most difficult. OCR technology reduces that number for AP teams, particularly when state-of-the-art OCR software is implemented.
Benefits of OCR.
OCR software extracts data from printed or written text in a scanned document or image, such as a PDF, TIFF, or JPEG, then converts it to machine-readable data. Once digitized, the data is searchable, editable, easily stored, and able to be repurposed. When thoughtfully implemented to AP processes, OCR software can:
- Eliminate the need for manual data entry — one of the most time-consuming manual processes for finance departments.
- Save money, since well-trained finance professionals won’t have to spend their valuable time entering data.
- Reduce the risk of errors.
- Free up physical storage space since all necessary documents can be stored digitally.
Historical limitations of OCR.
The origins of OCR can be traced back to WW1, when early devices scanned printed text and converted it to telegraph code. For decades, however, OCR’s practical usefulness was limited for finance teams. One problem was that earlier technology was often restricted to extracting text images only in certain fonts. (You might recognize the font OCR-A, which was created specifically to be machine readable.)
Another was that although the technology could convert an image to data, it couldn’t discern any meaning at all. Without this ability, the text to be extracted had to follow a set format, which rarely happens in AP departments. Consider an invoice, for example. The price on one invoice might be displayed where the due date is found in another. Because this creates the possibility of confusion, or even errors, the data extracted by OCR isn’t very useful on its own unless it can be properly categorized, stored in a consistent format, and placed in the correct field. In other words, a strong foundation, centered around machine learning, must exist for the ingested data from a receipt or invoice to be truly useful.
Advances in OCR technology.
Today, the best OCR software uses machine learning to constantly improve itself while "learning" the context of extracted data to place it in the correct field. It iterates with each transaction, essentially getting smarter every time.
However, not all OCR software is created equal. Not all make effective use of machine learning. Some struggle with fairly common fonts, leading to inaccuracies, while others have extremely slow processing speeds. For these reasons, Airbase performed a thorough vendor evaluation before arriving at our own OCR provider.
Optimizing OCR in AP processes: Best practices.
Let’s take a look at how OCR can be implemented in a spend management platform to streamline processes, save time, and improve accuracy.
Reimbursements: Submitting receipts for employee expense reimbursements is typically a nuisance task for employees, who may forget to keep the receipt (or even ask for it). Plus, submitting the receipt with an expense report, then waiting for reimbursement slows down the whole process. When OCR is built into a mobile app, an employee can simply take a photo of a receipt and submit it. A well-designed system will even remind the purchaser to take the photo. Alternatively, they can submit a receipt from their phone’s photo library, or find it in the system’s Receipt Inbox. Once submitted, OCR technology scans the receipt and extracts the required information.
Bill payments: Speed up the bill payment process with OCR. When an invoice is submitted, OCR technology can recognize it is an invoice, and place it in the correct place in the system. It can even electronically file supporting documents correctly, including email chains, so the context of a bill is instantly available. OCR can also recognize submitted documents like W-9s, and file them accordingly in the system for easy reference when it’s time to prepare 1099s.
When creating a bit from an invoice, ORC can extract the relevant information quickly. At Airbase, for example, the following fields are auto-filled:
- Invoice Date
- Due Date
- Invoice Amount
- Invoice Number
Plus, if the vendor has had bills in the past, we auto-fill the fields in the GL.
OCR technology continues to improve — make sure you take full advantage of its ability to make everyone’s life easier. Schedule a demo to find out how Airbase can help.
Airbase offers a one platform solution to manage all non-payroll spend. It provides oversight and control over spending with real-time reporting and automatic syncing directly to your general ledger. Control all payments – physical cards, virtual cards, ACH, and checks – from one place. Close faster. Empower employees. Control spend.
Finance & Accounting Slack Group.
Join to connect with other finance professionals building great companies. Ask questions, provide your perspective, join the conversation, find resources.