AI-Powered Virtual Assistants in the Metaverse: Leveraging Retrieval-Augmented Generation for Smarter Interactions

The metaverse has evolved at an unprecedented pace, creating new demands for on the spot user support and more engaging digital encounters. This paper describes a chatbot system built for the NT metaverse that combines retrieval-based search with advanced generative AI methods to provide accurate, context-driven responses. At the core of our approach is Retrieval Augmented Generation (RAG), which adeptly interprets diverse user queries while sustaining high performance under concurrent usage, as evidenced by a cosine similarity score of 0.79. In addition to maintaining efficiency during load testing, the system manages compound queries with ease, enhancing user satisfaction in complex virtual environments. Although these results are promising, future upgrades such as integrating voice-based interactions, multilingual support, and adaptive learning could further expand the chatbot’s utility. Overall, this study demonstrates the tangible benefits of AI-driven conversational agents in digital realms, laying the groundwork for richer, more intelligent user experiences in emerging metaverse platforms.