Evaluating the Output of Your LLM (Large Language Models): Insights from Microsoft & LangChain