Explanation of Results

In this chapter, we specify the business requirements and propose the solution concept for explainability. To build trust between human and machine, it’s important to explain the results provided by artificial intelligence models. Transparency and traceability of artificial intelligence models are also needed for statutory reasons. However, depending on the underlying artificial intelligence techniques, this can be very challenging, for example, neuronal networks are hard to explain. In the context of ERP systems, additionally, the explanation must be transferred into a business language. Thus, user interface designers must investigate long time, for each use case translates the statistical numbers into the end user’s business domain.