Using a Chatbot to Improve the Usability of Process Discovery for Non-Technical Experts

Abstract

Increasing amounts of event data are collected by a wide variety of information systems. Such data provide accurate information about the history of business processes implemented in companies, organisations, agencies and industries. These companies and organisations try to improve their business processes to stay competitive in fast-changing environments. Process mining can be a useful way to design, control and support business processes by providing the functions of process discovery, process conformance checking and process enhancement. Standard open-source tools such as ProM or programming libraries such as PM4Py provide process mining functions for experts where deep knowledge is required either to execute process mining or to understand the outcome. This thesis examines whether chatbot technologies can be used to simplify process mining in terms of usability and understandability. Chatbots are an increasingly common technology and are used in many industries such as healthcare and education. In this study, the analysis of the literature on the user-friendly characteristics of chatbot conversations and the feedback from survey participants led to a straightforward conversation design. A suitable chatbot provider was selected during the development phase, and the chatbot was built according to the developed blueprint.Having a usable chatbot allowed feedback to be collected from the participants and the research questions to be answered. The final chapters explain the findings of this thesis and its contribution to the literature. Additionally, this thesis shows what further study can provide to academia.