This paper proposes a structured and reusable function generation method by encapsulating code generated by large language models (LLMs) using a Helper Object. The method aims to streamline the integration of the domain-specific language AJAN and the general-purpose language Python by adopting a unified interface design that facilitates seamless data exchange between the two languages. Specifically, the Helper Object absorbs the differences in data structures between AJAN and Python, providing a mechanism to easily generate function groups that are consistent and reusable. Furthermore, a prototype system, the AI Code Capsule (AICC), was developed based on the Helper Object. The system demonstrated that automated and reusable function encapsulation significantly improved development efficiency, reducing average development time by up to 82% and achieving more uniform task completion times across developers. This finding highlights the potential of the proposed method to simplify complex workflows and promote standardization, making it highly effective in industrial applications.