AI-generated content and business English speaking: A mixed-methods study of SpeakGuru's impact on performance, motivation, and engagement

The integration of Artificial Intelligence-Generated Content (AIGC) tools in language education, particularly for Business English learners, has gained increasing attention; however, experimental evidence on their pedagogical impact remains limited. This study addresses this gap by evaluating the effectiveness of SpeakGuru, an AIGC-supported learning platform, in enhancing speaking performance, motivation, and engagement among Business English learners. A quasi-experimental mixed-methods design was employed over eight weeks with 120 participants, using intact classes with non-random group assignment to compare AI-assisted instruction with traditional classroom practice. Quantitative results revealed that learners using SpeakGuru demonstrated significant improvements in speaking performance, motivation, and engagement, accompanied by large to very large effect sizes. Qualitative insights further indicated that participants reported increased confidence, reduced speaking anxiety, and greater enjoyment during AI-mediated speaking tasks. Grounded in Self-Determination Theory, the study shows that SpeakGuru supports learners' autonomy, competence, and relatedness, thereby strengthening intrinsic motivation and engagement. The findings underscore the pedagogical potential of integrating AIGC tools into Business English instruction through scaffolded practice, authentic simulations, and real-time feedback. Despite limitations related to non-random group assignment, the sample size, and the short intervention period, this study provides evidence for SpeakGuru's capacity to enhance oral proficiency and learner engagement. Future research should examine long-term effects, cross-skill integration, and comparisons with alternative AI-based speaking platforms.