RAG to Riches: Supercharging LLMs for Superior AI Testing