Optical Character Recognition or OCR is transforming accounting by reducing data entry for accountants. This tech tool can take on time-consuming tasks, allowing professionals to focus on more value-add activities. But first things first:
What is Optical Character Recognition (OCR)?
Optical Character Recognition is a technology designed to extract text from an image and convert it into editable and searchable data. You can think of it as a scanner that has evolved far beyond photocopying. OCR can read and store printed or handwritten text, providing a huge benefit to accountants who spend time manually coding documents including receipts, checks, and invoices. Hallelujah!
While OCR has only recently started to disrupt accounting, it is by no means a new technology. Some of the first computerized applications of OCR date back to the 1950s. In 1951 a Department of Defense engineer created an elementary scanner that could interpret Morse Code and read messages aloud. The US post office has used OCR for over thirty years to scan address labels. As digital cameras and processing software modernized over the past 10 years, the speed and accuracy of OCR have improved. As a result, the applications of OCR have expanded across industries.
While the benefits of OCR in accounting are transformative, enhancements can still be made to how accurate OCR converts data from both structured and unstructured documents.
Optical Character Recognition and Structured Documents
Structured documents, like checks, follow a static layout. The key data is always found in the same place. This standard format is ideal for Optical Character Recognition technology. You’re likely to be familiar with one particular use of OCR and structured documents: depositing checks using your mobile phone.
The efficient point-and-shoot method of depositing checks is a great first step for OCR. When you snap a picture of your check, OCR knows precisely where to look and what to look for. Checks even use the same block-like font, based on magnetic reader technology, for routing and account numbers.
Having this kind of structure in place dramatically improves the accuracy of OCR. However, even despite the uniformity of structured checks, banks still require the user to enter the total check amount, leaving part of the data entry on you.
Optical Character Recognition and Unstructured Documents
Documents that vary in formattings such as bills and receipts are considered unstructured documents. The variation makes these documents difficult for Optical Character Recognition to interpret. Fortunately, we are witnessing a breakthrough moment for unstructured OCR.
Tech titans such as Google and Microsoft are investing heavily in OCR applications and we are beginning to see the gains. For example, Google Translate allows users to take a picture of foreign text, such as a traffic sign or a menu, and receive a translation on the spot.
To share these OCR developments with the broader tech community Google released Cloud Vision: a smart image analysis tool that gives startups access to OCR technology without having to build the software in-house. A win-win for consumers and businesses alike.
Even Oxford University has joined in on the fun, releasing a series of publications on text spotting. The renowned institution has developed a technology that can read words in non-standard fonts and handwritten text, and interpret the information much like our brains do.
Optical Character Recognition isn’t perfect
As a result of the ongoing investment, computers are beginning to assume more and more data entry responsibilities for accountants. Automating this process will take a significant burden off of accountants, however, the primary obstacle for adoption in our profession is accuracy.
Today, we see OCR accuracy in the 80% to 90% range for unstructured documents and greater than 90% for structured documents. Although imperfect, accountants are beginning to leverage OCR to shift their time from data entry to data review. This phenomenon will have a lasting and transformative effect on the daily responsibilities of accounting teams.
After OCR, what’s next? In our next post, we’ll explore data science and its role from data review to an external audit.
lainey 