Build a chatbot to query your documentation using Langchain and Azure OpenAI

First of all - thanks for a great blog, easy to follow and understand for newbies to Langchain like myself.

Question: what is, in your opinion, the benefit of using this Langchain model as opposed to just using the same document(s) directly with Azure AI Services? I just made a comparison by implementing your solution and - in parallel - using the Azure AI Services with the same document. Out of the box, Azure AI with own data added provides similar insights and conversational flow. Granted, some questions will be answered differently, presumably based on how the indexing works in both models, but on par, the user experience is very similar.

I understand Langchain has a wide range of other options, but for relatively limited models with a fixed set of existing documentation, Azure AI Services equipped with private data might be an easier way. Any comments on this?