Implementace chatbota pro Fakultu informatiky a statistiky VŠE

The thesis focuses on the implementation of a chatbot for the Faculty of Informatics and Statistics of the Prague University of Economics and Business. The result is a prototype version suitable for further research and implementation. The purpose of the chatbot is to provide undergraduate students in the Applied Informatics program with the information they need to pursue their studies, thus saving them from unnecessary stress. The thesis briefly introduces the history of chatbots along with the origins of artificial intelligence in the 1950s, while also describing state-of-the-art language models such as GPT-4. The thesis builds upon a previous chatbot implementation and attempts to make it more user-friendly, showing how to appropriately design a chatbot's communication style. Part of the thesis is also dedicated to the current channels of communication by the Faculty. The chatbot's knowledge base is created by employing both the method of observation within a student group and a questionnaire survey method. This data is further analyzed in the thesis and based on this analysis, a conversation including factual answers to questions related to the studies is designed. The chatbot is implemented in the open-source tool Botpress and it demonstrates its flexibility for future extensions. The resulting implementation is repeatedly iterated over based on respondents' feedback from in-depth interviews, and finally, an experiment is conducted to determine whether the current state of advanced language models allows for the implementation of a similar chatbot using prompt engineering.