Facilitating Holistic Evaluations with LLMs: Insights from Scenario-Based Experiments