Machines are becoming more versed at adapting to environmental impulses and their operational contexts, changing their roles in service systems. These machines can autonomously fulfill goals within defined boundaries set by legal actors and thereby exhibit agency. As interactions are at the core of services, the integration of such non-human agents into value co-creation has the potential to heavily impact the innovation of services. Following a four-step type construction approach based on empirical open-source data on 130 services, we develop a multi-dimensional characterization of six roles that non-human agents fulfill in service systems. Taking a systemic perspective, we identify service system interactions involving non-human agents and how their contributions impact value propositions. Our findings forward the understanding of service processes involving non-human agents and their impact on value co-creation, benefiting both theory and practice through knowledge on the engineering of service systems and value-driven, user-centered human-machine interactions.

Chatbot technology has rapidly spread, especially in digital customer service. However, the automation potential of chatbots can only be realized if customers are satisfied with their service. Collecting explicit feedback is a promising technique for assessing customer satisfaction and identifying issues with the chatbot. It enables chatbot managers and developers to enhance performance and design of operational chatbots on an informed basis. The evident significance of explicit customer feedback comes with a multitude of design options available. However, there is a lack of research on chatbot feedback mechanisms and practical as well as theoretical clarity. In this paper, we adress this gap by introducing a chatbot feedback taxonomy derived from existing research and a sample of N = 72 real world customer service chatbots. Furthermore, based on a cluster analysis, we identify four archetypes of feedback mechanisms and provide strategic guidelines for the informed use of each of those feedback design variants.