Customer inquiries, invoices and reminders, contracts, orders – scanned documents (images) or correspondence are often part of business processes. To automate them, you need an RPA product with high-quality image recognition and character recognition (OCR). In combination with Artificial Intelligence, you can even fully automate selected processes (end-to-end).

For many RPA products, the recognition of images and text or characters (Optical Character Recognition, OCR) is a challenge. Therefore, companies should check how well an RPA product can handle these functions. Otherwise, the possibilities, success and scalability of their RPA activities are significantly limited.

High-quality image recognition and OCR technology is based on human vision and functions exclusively via the user interface or the image.

Intelligent “fuzzy logic”: Hide unimportant information and deal with blurriness

People can recognize familiar patterns, even if they’re blurred. Few recognition features are enough for us. This also applies for Servicetrace’s patented image and character recognition: it reduces to the essential recognition features and hides unimportant information. This is made possible by an intelligent “fuzzy logic”: The flexible handling of blurring increases the stability of the automation, as there are fewer mishaps and malfunctions.

Thus, companies can automate processes that integrate images or text in image format just as quickly and in the same high quality as “normal” business processes without media discontinuities – without additional effort.

Practical example: End-to-end automation in call centers with NLP, OCR and RPA

A medium-sized company with a growing online business struggled with a steadily increasing number of customer inquiries. Due to long response times in some cases, there was an increase in poor customer ratings and negative social media feedback. With the aim of processing inquiries faster and around the clock, the company examined possibilities for automation.

The solution: Customer inquiries received via chat or e-mail are evaluated thematically with the help of an intelligent chatbot. Incoming letters, for example, are scanned and then the topics are identified with the help of OCR competence. Depending on the issue, the solution then triggers a predefined and automated process. The intelligent solution can handle questions about delivery dates and rebooking to express deliveries completely automatically. For more complex inquiries, the RPA robot brings a clerk on board at certain decision points. Here the bot supports the service employee (“attended bot”).

The result: an 85 percent faster response time in customer service led to an 80 percent increase in positive customer ratings within just a few weeks.