Peer-assisted learning (PAL) has been widely adopted in higher education to promote engagement and develop instructional competence. However, traditional models, such as one-to-one peer tutoring or unstructured collaborative learning, often face challenges including uneven participation, cognitive overload, and limited scaffolding. This study introduces MultiPAL, a structured multi-tutor peer-assisted learning model designed to enhance instructional quality and collaborative preparation in online settings. Implemented in a postgraduate course on learning design, MultiPAL involved a four-phase cycle: individual preparation, collaborative planning, synchronous multi-tutor sessions, and structured reflection. Thirteen students participated over seven weeks, with data collected through instructional artifacts, video recordings, and written reflections. Findings indicate that MultiPAL fostered meaningful peer feedback, instructional clarity, and collaborative engagement. Teams with consistent planning and shared responsibility produced higher-quality instructional outputs. Video analysis showed frequent use of scaffolding and peer support strategies, while reflections highlighted pedagogical growth through teaching and collaborative problem-solving. Challenges such as time constraints, cognitive load, and uneven tutor preparedness were also identified, pointing to areas for design refinement. The study contributes to structured PAL literature by illustrating how peer-led instruction can be effectively scaffolded in digital environments. Implications include the need for clearer role allocation, inclusive planning tools, and technological supports to enhance coordination and accessibility in future iterations.