Requirements in AI systems are rarely fixed at the start of a project, as changes are often necessary—and sometimes unavoidable—due to traditional factors and AI-specific challenges such as evolving data patterns and model retraining needs. Managing these changes effectively is essential to ensure system reliability and project success. However, current AI development practices typically separate the development pipelines for ML and non-ML components, creating a lack of systematic coordination when changes occur. This fragmentation often leads to inconsistencies, overlooked impacts, and error propagation across pipelines. To address these challenges, this paper presents an approach to streamline requirements change management using process modeling techniques. By providing a cohesive framework, our approach enables stakeholders to identify and address critical changes early in the development lifecycle, minimizing risks and preventing errors from reaching production.