In this step-by-step tutorial, we demonstrate how to leverage large language models such as Google’s PaLM 2 and OpenAI’s GPT-4 to analyze concepts, with a focus on considering their respective training cut-off dates. We discuss the process of tiering these models, how to set up fallback mechanisms using data from the internet, and how to design an adaptive approach that ensures the system understands the concept at hand.