Curriculum mapping plays a critical role in education to ensure alignment between outcomes, content, graduate skills and assessment. Program requirements are typically informed by industry needs and embedded within the curricula and assessment tasks. Curriculum Analytics (CA) has introduced a level of automation to the process of curriculum mapping primarily through machine learning (ML) models. While such CA approaches have served to reduce workload pressures, they continue to face challenges in capturing the nuanced extent to which graduate skills are developed across a program. This study introduces a novel approach using Large Language Models (LLMs) to act as co-curriculum reviewers. Using data from an undergraduate program, we evaluate the effectiveness of LLMs in generating weighted mappings of graduate skills across assessments and compare them to those produced by ML based CA methods. The findings suggest that LLM-generated mappings more closely align with expert judgements (Krippendorff’s alpha of 0.76) than those produced by ML based CA models (0.65). The results demonstrate the potential for LLM-driven approaches to curriculum mapping to enhance quality assurance processes linked to curricular alignment, accreditation, and personalized learning pathways.