Masterarbeit
Categorization of Chatbot Interaction Patterns based on an Analysis of German
Municipal Chatbots
Completion
2024/12
Research Area
Students
Sonia Abhyankar
Advisers
Description
Even though chatbots have seen a recent rise due to LLMs, they were already previously used as a form of content from companies, organisations or government institutions. While some chatbots are text-based, they can also contain graphical elements which are forming interaction design patterns. However, there is no in-depth analysis of chatbots in the wild available yet. Moreover, a holistic categorization or even a dataset of chatbot interaction patterns is needed for future research on their usability, user experience and user satisfaction
This thesis aims to comprehensively categorize graphical chatbot interaction patterns. The basis for this will be a collection of German municipal chatbots as they offer several services for which a wide range of chatbot interaction patterns can potentially be used. The interaction design of municipal chatbots was already researched on Dutch chatbots, even though it was not on the level of individual interaction patterns. The methodology for the categorization can be based on a similar publication which mapped chatbot functionalities to known interaction design patterns. Based on this analysis and categorization, a chatbot interface can show exemplary use cases with chatbot interaction patterns, which can be used for future research on chatbot interaction patterns.
The objective of this thesis is the creation of a solution or the combination of existing approaches to solve the problem of not yet categorized chatbot interaction patterns of German municipal chatbots as described above. This comprises the analysis of the state of the art of interaction design both for chatbots and graphical user interfaces, methodologies for creating datasets, chatbot technologies, and other relevant literature. From this analysis, a dataset of municipal chatbots has to be to created and analysed, before a prototypical implementation should demonstrate the feasibility of the categorized chatbot interaction patterns in meaningful use cases. A suitable evaluation has to be conducted based on the usability of this prototype and its compliance with requirements which were extracted through the literature research.
