Dialogue-based knowledge transfer
Dialogue-based knowledge transfer: User acceptance of internal corporate chatbots
Master's thesis by Paul Heß
3rd semester, 2024
Supervisor: Prof. Dr Sebastian Pranz
Field of research
“We need our own chatbot!” This is a statement that people in various office roles may have heard from their line manager since the launch of ChatGPT in November 2022 and the hype that followed (Angwin, 2024). But are corporate chatbots and their conversational format suitable for every application? This research explores this topic, with a particular focus on internal bots.
Alongside robots and automation – which have been a long-standing aspiration of humankind, dating back to well before the advent of chatbots, as evidenced by Leonardo da Vinci’s plans for robotic constructions (Mainzer, 2019, p. 7), knowledge management has been relevant to humans at least since the time of Egyptian hieroglyphs (Spender, 2015, p. 4). To examine the combination of these two topics in relation to internal corporate chatbots, a number of more recent fields of research are relevant.
Organisational theories by Max Weber and Herbert A. Simon provide insight into organisations, with Weber describing organisations as goal-oriented systems (Weber, 1947, p. 95) and Simon highlighting the need for communication and the technologies that facilitate it, to provide a basis for decision-making (Simon, 2000, p. 208). In the field of knowledge management, the utilisation of knowledge to achieve corporate goals is examined (Greiner et al., 2007, p. 4). Another relevant field is that of internal communication, which involves communicating with all employees (Welch & Jackson, 2007, p. 184) within an organisation as part of knowledge management (O’Leary, 1998, p. 59). For this research, the channels, knowledge elements, objectives and metrics, and challenges of internal communication were examined. Finally, research into the technology itself—namely chatbots—can provide insight into the research topic. From the first chatbot, ELIZA (Bruns & Kowald, 2023, p. 6), through to exploring a linguistic perspective on conversing with computers (Lotze, 2018) and existing use cases of corporate chatbots in HR (Kohne et al., 2020, p. 37) or IT (Fiore et al., 2020, p. 86). This research focused on rule-based chatbots (Hussain et al., 2019, p. 947), examining artificial intelligence to varying degrees of relevance.
Method
To provide a productive perspective, based on a literature review, expert interviews were conducted with internal communication specialists. Furthermore, users were presented with internal corporate chatbot technology in comparison to the widely used (Hudcova, 2014, p. 56) Intranet. This was carried out as a laboratory quasi-experiment followed by a panel discussion.
The tasks for the quasi-experiment were carried out using prototypes, accessible at https://intranet.paulhess.de/.
Subsequently, the transcripts of the expert interviews and the laboratory sessions were transcribed and coded qualitatively.
As shown in Figure 2, the expert interviews, together with the literature, formed the basis for the laboratory tasks and questions, thereby also addressing research question 2, whilst the laboratory experiment aimed to address research question 1.
Results
The expert perspective offered insights into internal communication channels, knowledge elements, goals and challenges, and the potential and challenges for a chatbot. With regard to the channels, no clear clustering was identified, but approaches were suggested for clustering into static/simple/lean versus interactive/advanced/media-rich channels. With regard to knowledge elements, no clear clustering was found either, highlighting the need to assess each element individually. The objectives were mostly seen as informing employees or improving the company’s reputation, whilst the challenges included diversity, informing employees, managing topics and consulting stakeholders. Potential for chatbots was identified in knowledge management, fostering engagement and motivation. No potential was identified in relation to networking. Challenges associated with chatbots were seen in terms of output, efficiency and adoption, but there was no consensus regarding GDPR or ethical issues.
Users also considered knowledge elements, as well as the user experience, objectives, fears or concerns, and future potential. Users viewed documents and multimedia elements as knowledge elements suitable for bots. Opinions on booking tools were mixed, whilst news and social elements were seen as better suited to the Intranet. With regard to the user experience, the potential of bots was seen in providing the right information and speed. The Intranet was considered to offer a better overview, and humour was not considered desirable. In terms of achieving internal communication goals, users felt that the aim of informing employees was well suited to bots. Motivation was seen to work better on the Intranet. Users’ main concerns centred on functionality, whilst attitudes towards privacy and the GDPR were mixed. No ethical concerns were raised. Future potential was identified, particularly with AI-powered chatbots.
Conclusion
As a basis for the conclusion, it is evident that, due to the distinction between static and complex / one-way and two-way channels, chatbots are suitable for comparison with the Intranet, as both are complex, two-way channels. However, knowledge elements must be assessed on their own merits, whilst goals and challenges can serve as a yardstick for evaluating a chatbot’s usefulness.
When assessing the potential for dialogue-based knowledge transfer using the example of internal corporate chatbots, it is evident that documents are considered most suitable for bots, alongside multimedia elements, people and external links. There is also a consensus that news and social elements work better on other channels. Bots receive a positive assessment for their handling of knowledge and information, as well as for their speed. The prerequisite for accepting a bot is that it must function effectively, whilst there is no clear stance on ethical or privacy concerns. Limitations are observed when it comes to providing an overview or incorporating humour. Furthermore, an onboarding process is required for the bots. A connection with artificial intelligence is also considered appropriate.
From this, the following hypothesis can be derived:
Employees feel better informed when internal communication uses chatbots compared to when it does not.
Internal corporate chatbots improve companies’ productivity compared to when internal communication does not use chatbots.
Users expect artificial intelligence to be incorporated when interacting with chatbots.
These areas could be explored further in the future.
Overall, there is potential for dialogue-based knowledge transfer via internal corporate chatbots.
Sources
Angwin, J. (15 May 2024). Press Pause on the Silicon Valley Hype Machine. New York Times. https://www.nytimes.com/2024/05/15/opinion/artificial-intelligence-ai openai-chatgpt-overrated-hype.html
Bruns, B., & Kowald, C. (2023). A Practical Guide to Chatbots. Springer Fachmedien Wiesbaden. https://doi.org/10.1007/978-3-658-39645-9
Fiore, D., Thiel, C., & Baldauf, M. (2020). The Potential of Chatbots for In-House IT Support. HMD Praxis Der Wirtschaftsinformatik, 57(1), 77–88. https://doi.org/10.1365/s40702-019-00578-7
Greiner, M. E., Böhmann, T., & Krcmar, H. (2007). A strategy for knowledge management. Journal of Knowledge Management, 11(6), 3–15. https://doi.org/10.1108/13673270710832127
Hudcova, S. (2014). Tools of Internal Communication from a Knowledge Transfer Perspective. Journal of Competitiveness, 6(4), 50–62. https://doi.org/10.7441/joc.2014.04.04
Hussain, S., Ameri Sianaki, O., & Ababneh, N. (2019). A Survey on Conversational Agents/Chatbots Classification and Design Techniques. In L. Barolli, M. Takizawa, F. Xhafa, & T. Enokido (Eds.), Advances in Intelligent Systems and Computing. Web, Artificial Intelligence and Network Applications: Proceedings of the Workshops of the33rd international Conference on Advanced Information Networking and Applications (WAINA-2019) (Vol. 927, pp. 946–956). Springer Cham. https://doi.org/10.1007/978-3-030-15035-8_93
Kohne, A., Kleinmanns, P., Rolf, C., & Beck, M. (2020). Chatbots. Springer Fachmedien Wiesbaden. https://doi.org/10.1007/978-3-658-28849-5
Lotze, N. (2018). On linguistic interaction with chatbots – A linguistic perspective. Innsbruck University Press. https://doi.org/10.25969/MEDIAREP/19874
Mainzer, K. (2019). Artificial Intelligence – When will the machines take over? Springer Berlin Heidelberg. https://doi.org/10.1007/978-3-662-58046-2
O’Leary, D. E. (1998). Enterprise knowledge management. Computer, 31(3), 54–61. https://doi.org/10.1109/2.660190
Simon, H. A. (2000). Administrative behaviour: A study of decision-making processes in administrative organisations (4th ed.). The Free Press.
Spender, J. C. (2015). Knowledge Management: Origins, History, and Development. In E. Bolisani & M. Handzic (Eds.), Knowledge Management and Organisational Learning. Advances in Knowledge Management (Vol. 1, pp. 3–23). Springer International Publishing. doi.org/10.1007/978-3-319-09501-1_1
Weber, M. (1947). The Theory of Social and Economic Organisation. The Free Press of Glencoe. 68 https://search.alexanderstreet.com/view/work/bibliographic_entity%7Cbibliographic_details%7C4708644
Welch, M., & Jackson, P. R. (2007). Rethinking internal communication: a stakeholder approach. Corporate Communications: An International Journal, 12(2), 177–198. https://doi.org/10.1108/13563280710744847