This paper presents AI-powered Metaverse for Speaking (AIMS)—a platform integrating AI and virtual reality (VR) metaverse for English presentation training. AIMS features an immersive environment for accessing learning materials with an AI tutor to provide personalised assistance in preparing for presentations. A preliminary study on learner experience and perceptions of the platform was conducted. Interview results indicate that the immersive nature of the platform was effective in engaging learners in training activities. The AI tutor was viewed as advantageous for its immediate recommendations, though participants noted its limitations in providing nuanced, context-specific advice compared to human feedback. Usability challenges, including technical glitches and headset discomfort, were identified as barriers to adoption. The findings suggest the platform’s potential to enhance presentation training by addressing limitations of traditional approaches, while facilitating conditions such as hardware infrastructure and user support also play a key role in ensuring success.