This mixed-methods study investigates the efficacy of Grammarly, an AI-powered writing assistant, in enhancing L2 English writing proficiency among 60 intermediate-to-advanced learners at Ilam University. Over a 12-week intervention, an experimental group (n=30) used Grammarly for essay drafting and revision, while a control group (n=30) received traditional teacher and peer feedback. Quantitative analysis of pre/post writing assessments revealed significantly greater improvements in the experimental group for linguistic accuracy, coherence, and syntactic complexity (p<0.001; large effect sizes: d=1.40–1.75), validated via rubric scoring and Coh-Metrix text analysis. Behavioral analytics showed revision frequency (mean=18.3/essay) and time-on-task predicted 54% of writing gains (p<0.001). Qualitative themes from semi-structured interviews (n=15) highlighted enhanced self-efficacy, motivation, and trust in AI feedback for grammar/style, though learners emphasized the need for complementary human guidance on higher-order concerns (e.g., argumentation). Results demonstrate Grammarly’s effectiveness in promoting writing proficiency and engagement, advocating for an integrated AI-human feedback approach in L2 pedagogy.
