​In an era of “digital first”, still a large fraction of the processed documents exist in printed format. A train ticket, a signed contract, an invoice coming with shipped goods… To cut down on manual handling and automatically processing such documents, the first step is often Optical Character Recognition – the process to extract text from printed pages. Though many technologies to solve this problem have been proposed over decades, they are often unsatisfying in accuracy and runtime. In this podcast we discuss Chargrid-OCR, a solution developed by SAP AI Business Services based on cutting-edge deep learning technology. Thanks to recent advances in algorithms and making use of the most recent hardware, Chargrid-OCR allows to boost accuracy and efficiency of automatic information extraction.