VietTESOL International Convention 2026

Written corrective feedback has long been central to L2 writing instruction, yet a persistent gap exists between teacher investment in feedback and learners’ meaningful engagement, particularly concerning affective engagement. At the same time, the growing presence of AI-mediated feedback tools in writing classrooms is opening new possibilities for how feedback can be delivered. Grounded in a multidimensional view of learner engagement, this study addresses two questions in the Vietnamese L2 Essay Writing context: what intermediate learners need from written corrective feedback, and how their affective engagement shifts when an AI preparatory stage precedes teacher feedback. A sequential mixed-methods design was employed, beginning with a quantitative questionnaire (n=221) to identify learner preferences, which subsequently informed a qualitative intervention phase involving five participants across three writing tasks. Preference findings suggest students consistently favoured timely, stage-by-stage and personalized feedback, suggesting the value of AI stage within a feedback sequence, where AI could deliver immediate feedback, address lower-order issues before teacher feedback and reduce teacher workload to enable more personalised higher-order comments. Findings for affective engagement indicate that the AI preparatory stage appeared to reduce emotional overload and shift how students arrived at teacher feedback, though outcomes varied across individuals. Teacher and AI relationship was viewed as complementary, with AI handling lower-order errors, while the teacher handling higher-order ones. The findings contribute to understanding how integrating AI into the feedback process may shift students' affective readiness without displacing teacher authority, offering practical implications for teachers seeking to design more effective feedback workflows.