Artificial Intelligence in Automated Document Processing Using the Example of Health Insurance Companies
摘要
The fully automated processing of documents in insurance claims is referred as dark processing by German insurers. Dark processing is a form of digitalization in document processing. For cost reasons, insurance companies have been working for years to increase the share of dark processing. Invoices in private health insurance in Germany offer good conditions for automated processing due to the fee code system based on the German fee schedule. Both classic artificial intelligence (AI) techniques such as rule systems and newer connectionist approaches such as convolutional neural networks (CNN) with deep learning are used here.