Task Based Datasets

According to the EF English Proficiency Index (2024), the Kingdom of Saudi Arabia (KSA) remains in the "low proficiency" band, posing persistent challenges for EFLlearners required to master academic and workplace communication. As Vision 2030 of KSA advances a shift toward a knowledge-based economy, innovative pedagogical strategies become essential to improving English instruction in technical educational institutes. This study compares two instructional models: Task-Based Instruction (TBI) using <i>Technical</i> <i>English</i> <i>Book</i> <i>1</i> and TBI integrated with Cambridge's Write &amp; Improve AI feedback tool. The study’s aim is to assess their impact on collocation mastery and sentence-level writing fluency among Foundation Year students at Yanbu Technical Institute. A mixed-methods design was employed with 58 participants assigned to two instructional sections. Data were collected through five graded writing assignments, a 17-item Can-Do self-evaluation survey, and semi-structured interviews with instructors and program administrators. Quantitative analysis revealed that students using Cambridge Write &amp; Improve AI feedback tool achieved significantly higher writing scores and demonstrated greater improvement across assignments. Qualitative data supported these findings, emphasizing increased learner motivation, autonomy, and engagement. The study offers practical implications for curriculum designers and policymakers aiming to modernize English for Specific Purposes (ESP) instruction through blended models. It contributes to applied linguistics by embedding AI-driven feedback into task-based frameworks. The proposed model is replicable and adaptable to broader ESP and multilingual contexts.