English as a Lingua Franca (ELF) communication prioritizes intelligibility over native-like accuracy among speakers with diverse linguistic backgrounds. This mixed-methods study examines whether real-time automatic speech recognition (ASR) transcription enhances the intelligibility of Indonesian-accented English (IAE) in ELF contexts. Data were collected from Universitas PGRI Kanjuruhan Malang students through pre- and post-intelligibility transcription tasks, completed with and without Google Live Transcribe, as well as questionnaires and semi-structured interviews. Quantitative results revealed a statistically significant improvement in listener intelligibility when ASR support was available, particularly for low-frequency vocabulary, technical terms, and sentence-final elements, alongside reduced performance variability. Qualitative findings indicated positive user perceptions of real-time transcription as accessible, user-friendly, and supportive of comprehension and communicative confidence, despite occasional transcription errors. Overall, the findings suggest that real-time ASR transcription functions as cognitive scaffolding that mitigates accent-related processing challenges in ELF communication. This study contributes to ELF and CALL literature by providing empirical evidence that ASR-mediated interaction functions as cognitive scaffolding, supporting listener intelligibility in multilingual English use, particularly in the processing of low-frequency vocabulary, technical terms, and sentence-final elements.