We have successfully completed our research project on data-driven chatbot optimization in collaboration with the Smart Data Innovation Lab (SDIL). In the project, we analyzed chatbot conversation data from hsag Heidelberger Services AG (hsag) on the SDIL infrastructure to understand trends and patterns in customer behavior. Focus of the analyses was to identify when customer-bot conversations fail and how these situations can be automatically derived from data using machine learning. The results of the project provide chatbot researchers and practitioners with a framework to analyze their own operational chat data. Learn more about our research on hybrid human-AI systems here.