KIT | KIT-Bibliothek | Impressum | Datenschutz

Towards Designing a Conversation Mining System for Customer Service Chatbots

Schloß, Daniel ORCID iD icon 1; Gnewuch, Ulrich 1; Mädche, Alexander 1
1 Institut für Wirtschaftsinformatik und Marketing (IISM), Karlsruher Institut für Technologie (KIT)


Chatbots are increasingly used to provide customer service. However, despite technological advances, customer service chatbots frequently reach their limits in customer interactions. This is not immediately apparent to both chatbot operators (e.g., customer service managers) and chatbot developers because analyzing conversational data is difficult and labor-intensive. To address this problem, our ongoing design science research project aims to develop a conversation mining system for the automated analysis of customer-chatbot conversations. Based on the exploration of large dataset (N= 91,678 conversations) and six interviews with industry experts, we developed the backend of the system. Specifically, we identified and operationalized important criteria for evalu-ating conversations. Our next step will be the evaluation with industry experts. Ultimately, we aim to contribute to research and practice by providing design knowledge for conversation mining systems that leverage the treasure trove of data from customer-chatbot conversations to generate valuable insights for managers and developers.

Verlagsausgabe §
DOI: 10.5445/IR/1000156160
Veröffentlicht am 21.02.2023
Cover der Publikation
Zugehörige Institution(en) am KIT Institut für Wirtschaftsinformatik und Marketing (IISM)
Publikationstyp Proceedingsbeitrag
Publikationsdatum 12.12.2022
Sprache Englisch
Identifikator KITopen-ID: 1000156160
Erschienen in ICIS 2022 Proceedings. Bd.: 1
Veranstaltung 43rd International Conference on Information Systems (ICIS 2022), Kopenhagen, Dänemark, 09.12.2022 – 14.12.2022
Verlag AIS eLibrary (AISeL)
Schlagwörter Chatbot, Customer service, Conversation Mining, Design Science Research
KIT – Die Forschungsuniversität in der Helmholtz-Gemeinschaft
KITopen Landing Page