Models like GPT are excellent at answering general questions from public data sources but aren't perfect. Accuracy takes a nosedive when you need to access domain expertise, recent data, or proprietary data sources.

This is where retrieval augmented generation (RAG) comes in – by enhancing your LLM with custom data sources, you can build a chatbot with only Python that combines the smarts of GPT (or your model of choice) with the specificity of your own data.