MASSTer: an AI-assisted MUET speaking simulation trainer for enhancing task fulfilment and interactive communication

MASSTer (MUET AI SpeakSim Trainer) is an AI-assisted pedagogical innovation designed to enhance students’ performance in the Malaysian University English Test (MUET) speaking component. Although MUET is aligned with the Common European Framework of Reference (CEFR), many pre-university students face challenges in fulfilling task requirements, generating relevant ideas, and participating effectively in group discussions. These issues are largely due to limited opportunities for structured speaking practice, lack of immediate feedback, and absence of realistic simulation environments. This innovation aims to develop a structured, rubric-aligned AI training system that supports both individual speaking tasks and group discussion simulations. MASSTer is developed using a design-based approach and operates through two modes: Practice Mode, which provides guided prompts and iterative feedback to support idea development, and Full Simulation Mode, which replicates real MUET test conditions without intervention. The system utilises AI-driven analysis aligned with MUET assessment criteria. A pilot implementation involving 180 students indicated positive outcomes, with 73.6% reporting improved confidence and preparedness. The system also recorded high levels of user acceptance, demonstrating its effectiveness in enhancing task fulfilment, discourse management, and interactive communication skills. In conclusion, MASSTer provides a scalable and cost-effective solution for MUET speaking preparation. It demonstrates strong potential for commercialisation through application development, institutional integration, and wider adoption in English language education.