Enhancing Computational Cognitive Architectures with LLMs: A Case Study