Chatovací aplikace s integrací jazykových modelů

Abstract

Customer support automation, particularly outside working hours, has long been a widely discussed topic. With the growing popularity and adoption of large language models, this area has gained new significance and practical feasibility. This bachelor's thesis focuses on the design and implementation of a platform for managing chat widgets that serve as customer support tools for websites, integrating language models to automate responses.

The thesis combines theoretical knowledge from the field of chatbot development and their application in customer support with a practical implementation in the context of modern customer service requirements. The system employs modern implementation approaches, including the Software as a Service (SaaS) model and microservices architecture. Special attention is given to the Retrieval-Augmented Generation (RAG) principle, which, in combination with a knowledge base, ensures a relevant and reliable source of information for the chatbot, thereby reducing the risk of providing inaccurate or misleading responses.

The result of this thesis is a functional prototype application that demonstrates the potential of artificial intelligence in the field of automated customer support.