A Voice-based AI Oral Exam Coach to Empower Students’ Learning and Exam Preparation
We develop a voice-based AI coach that helps students practise oral exams with course materials and feedback, and test its effects on learning.
Duration: January 2027 – December 2027
Education level: Tertiary Level
Topics: AI – Artificial Intelligence; Digital Tools
Keywords: oral exam preparation; voice-based AI; formative feedback; learning by explaining; self-regulated learning
Background
Oral exams ask students to explain ideas, make connections and defend their reasoning in conversation. They are gaining attention as generative AI makes it harder to judge students’ own understanding from written take-home assignments. Yet students often prepare alone, with few opportunities to practise answering questions aloud or receive useful feedback. Existing AI tools can help with speaking or simulate exam questions, but we know less about how to combine realistic spoken interaction with coaching that supports subject understanding. We develop an AI Oral Exam Coach that helps students prepare using their own course materials. We investigate how students use it, how they act on its feedback, and whether practice improves learning and performance in a subsequent oral exam.
Objectives
We design and evaluate a voice-based AI coach for university students preparing for oral exams. Students will set preparation goals, practise explaining course content aloud and receive feedback on understanding, reasoning, explanation structure and oral communication. We aim to produce an open-source prototype, design guidance and evidence on the coach’s educational value and limitations.
Method
We will adapt an existing voice-based AI platform to course materials, refine it through interviews and pilot studies, and evaluate in a multi-session study whether coaching improves students’ oral exam performance, preparedness, confidence and anxiety.
Planned Translation
We will release an open-source web application with technical documentation and guidance for instructors. The guidance will explain how to adapt the coach to other courses and where further evaluation is needed.
Students could gain more opportunities to practise explaining course content and use feedback to guide their preparation. Instructors will gain evidence on when voice-based AI coaching is useful and where it falls short. By examining performance in real oral exams alongside confidence and preparation behaviour, we will assess whether the coach supports learning beyond the practice session. Findings will inform its use in other university courses.
