OCR in Finance speeds up invoice processing
Optical Character Recognition (OCR) is a popular method of digitizing printed invoices. It allows a computer to understand the contents on the invoice so that it can be booked and archived quickly and easy. There are two flavors of OCR that differ in the degree of automation and responsibility. We call them OCR-on-Premise and OCR-as-a-Service.
Regardless of your preferred choice, OCR is an incredibly practical solution for companies that digest thousands of invoices annually. OCR makes it possible to have the computer retrieve any type of data from the invoice, such as supplier information, headers, line items, different amounts et cetera. Dealing with recurring invoices from regular suppliers saves time and minimizes (human) errors when the data is entered into the finance system automatically.
We measure OCR success in recognizability and reliability of data. Recognizability is understanding the contents of the invoice, while reliability is digitizing the right data. An OCR success rate can be 92 recognized/97% reliable.
Working with the on premise variant of OCR, your organization itself is responsible for tuning the invoice recognition. On its own, OCR will recognize much of the invoice contents, but by tuning the output – which is assisting the software by telling what information is where and how it is notated – the accuracy improves. Organizations that perform this task routinely can enjoy very high OCR success rates.
However, very few organizations structurally tune the invoice recognition. Especially when supplier turnover is high, consistent recognition tuning becomes a poor exercise over time. Once the OCR success rate drops to – say – 50%/35%, there is a dwindling motivation to keep the data in the system properly updated.
The solution is to assign this responsibility to an employee or team. But only if they have the skill and efficiency by routinely exercising this task will it save time and resources. Recognition tuning is not something most accountants payable employees do on a day to day basis. As a result, it will take longer and results may be less than optimal.
With OCR-as-a-Service, recognition of new supplier’s invoices is tuned by Onventis. There is no manual labor from the buyer’s side as this solution is really plug & play. The system automatically recognizes an unknown supplier and Onventis makes sure the supplier data is added, for future reference and automation.
New suppliers undergo a 3x Tuning process. Which means that Onventis employees will manually check and tune the first three invoices from this supplier, to make sure that recognizability and reliability are optimal. This is also the case if certain data changes; Imagine a bank account that is suddenly different, Onventis checks its validity before the invoice is confirmed.
This time-saving process makes sure that the OCR success rate is always high. Even years later, with no employee time lost on OCR tuning, the success rate will never dwindle due to this ongoing service.
Your department can now busy itself with other, qualitative activities.
Modern invoice processing goes much further than OCR. Robotic Accounting automates the entire process, from scanning and booking invoices to making intelligent booking proposals for new suppliers. Read more about in our free Robotic Accounting white paper.
Weitere BlogsMore BlogsMeer blogs