Design Thinking with a Bot - Supporting the Design Thinking Process through a Conversational Agent

Abstract

Motivation Companies need to create innovation to stay competitive. One way to innovate is Design Thinking (DT), an analytical approach to create new solutions for a previously defined problem in a five to six step process with an interdisciplinary team. It focuses strongly on user centricity with a neutral facilitator having diverse tasks such as organizational tasks, coaching the DT team and providing technical support. DT can be conducted both in digital and analog way. Conducting DT can lead to cognitive load and, in case of digital facilitation, also to technostress. Cognitive load is the mental load resulting from extensively processing information in various ways such as facilitation and problem solving. Technostress is experiencing stress due to the use of collaboration systems such as Microsoft teams which results in lower satisfaction and productivity. One way to conquer these challenges is to use a conversational agent (CA). These are interactive software systems using natural language processing and understanding for various purposes such as acting as a co-facilitator, conduct research, support creative work and be able to affect group cohesion. Informational elicitation as well as providing recommendations is another advantage of CAs. However, current research focuses mostly on information elicitation and merely covers CA as DT co facilitator. This field lacks systematic and user-centered design knowledge on how a CA can support the DT process and, thus, relief the facilitator in the DT process. Research Design This research follows the Design Science Research (DSR) approach focusing on the iterative creation of an artifact to address business needs from the environment. The artifacts in this dissertation are constructs in the form of design guidelines (DG) derived from the application domain and a design artifact instantiated in the collaboration tool Microsoft Teams. To support this research approach, several additional research methods were applied across the publications in this dissertation. This includes a systematic literature review, semi-structured qualitative interviews, development of a research agenda as well as quantitative and qualitative data collection and content analysis. Contribution This dissertation contributes to the knowledge base covering the conceptualization of creative work and group facilitation with a CA, a deep analysis of business needs for CA supported DT facilitation as well as nascent design knowledge for an instantiated CA supporting DT facilitation. These were then applied to the DSR Knowledge Contribution Matrix focusing on Improvement and Exaptation. The conceptualization of creative work and group facilitation shows five main concepts: fostering creativity, showing emotional intelligence towards humans, support for facilitation and group dynamics as well as elicitating and processing information and giving recommendations. At last, users´ expectations and needs are derived. A thorough analysis of business needs in the application domain shows four main problem areas which are individial and group related challenges as well as difficulties in working in new work modes and workshop situations in general. Focusing on digital DT facilitation, two main problem areas arise which are lack of nonverbal communication and workshop conduction in a digital environment. All problem areas were derived from both the facilitator´s and the participant´s perspective. Both contributions expand known solutions to new problems and, thus, lies in the Exaptation quadrant. Nascent design knowledge was then developed focusing on providing a neutral position, giving various kinds of input, conducting administrative tasks as well as providing network opportunities for participants. Besides concrete design guidelines (DGs), the CA can be assigned dual roles as in an assistant and the main provider of facilitation due to the fact that CAs are supposed to provide isolated tasks that are exactly the tasks of the facilitator. Thus, the main capability of a CA is derived as recognizing the respective challenge and need within a DT workshop and react respectively. Another main capability of a CA is its reciprocity of the CA´s benefits due to the fact that the facilitator is reactive towards the participants in its nature and, thus, beneficial to both parties. Besides these theoretical contributions, this dissertation contributes to two audiences in practice, namely the DT team and software developers. The DT team benefits from a consolidated overview of current literature showing possibilities and current state of the art research. Besides, a thorough analysis of business needs is provided which may help to conquer difficulties in both analog and digital DT workshops. Lastly, an instantiated CA for DT facilitation is provided for a first user testing. Software developers are provided with design knowledge which can be used to create a CA and further developed and tested. Limitations This dissertation faces limitations regarding the scope. In this dissertation, analog and digital conduction of DT were analyzed seperately. However, hybrid work is considered more and more in practice. Moreover, although a holistic approach was aimed in this research, not the whole DT process was considered but mainly general practices as well as the three phases Define, Ideate and Prototype. Furthermore, this research contains a small sample size which leads to restricted generalization and does not represent general DT workshop conduction. Considering the instantiated artifacts, they were built in a non functional way solely focusing on socio-technical aspects which may have an impact on the perception towards the artifact and its functionalities. Finally, mainly qualitative studies were conducted due to the novelty of the topic and lack of knowledge in existing literature. Future Research This dissertation gives implications for future research. First, further investigation of DT facilitation should be conducted in a cross industry approach as this research focuses on one specific domain in one industry. Besides, group facilitation as a whole should be considered in future research. Furthermore, hybrid conduction of DT should be examined. Role and task allocation between a CA and facilitators as well as the duality of a CA should be further analyzed. Lastly, the instantiated artifact should be further developed following the design knowledge derived in this research.