This study investigates the effect of integrating SmallTalk2Me AI within a flipped classroom model on students’ English speaking skills. A pre-experimental design with a one-group pretest-posttest structure was employed at SMA Negeri 1 Purwoasri, East Java, Indonesia. The sample consisted of 22 eleventh-grade students from class XI SAINKES 6, selected through cluster sampling. Students’ speaking skills were assessed across five components grammar, vocabulary, comprehension, fluency, and pronunciation using Brown’s (2004) speaking rubric. The intervention followed a three-stage flipped classroom procedure: pre-class independent study via instructional video, in-class communicative practice, and post-class self-study using SmallTalk2Me AI for personalized feedback. Data were collected through speaking tests and analyzed using a paired-sample t-test. The results showed a substantial improvement in students’ speaking performance, with the mean score increasing from 39.82 (pre-test) to 79.82 (post-test). The paired-sample t-test indicated a statistically significant difference (p = .000 < .05). These findings demonstrate that integrating SmallTalk2Me AI into a flipped classroom is an effective strategy for improving EFL speaking skills among senior high school students.
Copyrights © 2026