An Explanation User Interface for a Knowledge Graph-Based XAI Approach to Process Analysis

In consulting practice, effective use of AI technologies presupposes consultant’s and client’s ability to understand generated results. Our knowledge graph-based approach to explainable process analyses represents a hybrid AI approach that integrates symbolic approaches of structured knowledge and interactive machine learning methods for making algorithmic procedures traceable and representing analysis results in a human-readable form. In order to display identified weaknesses and suitable improvement measures of analyzed business processes in a user-understandable way and to enable human-in-the-loop interactions, an explainable, user-friendly interface is required. While much attention is paid to the computational aspects of generating explanations, there is a need for further research into the design of explanation user interfaces. A systematic literature review was conducted to derive a design catalog, which was demonstrated and evaluated by developing a suitable XUI for our knowledge graph-based explainable process analysis.