Hi Team, Need your expertise in solving my problem.

I'm building a RAG application which uses AOAI as a chat model, Azure AI search as a vector database and Langchain as an orchestrator.

Features of the application

1. Q&A over Internal Documents
2. Generation of resolution for New Tickets based on previous similar Ticket
3. Data Analytics on previously reported tickets

Data Format:

Tickets are stored in CSV format and contain around 10K rows

Sample schema of CSV:

Question:

For both 1 & 2 feature naive RAG architecture gives an appropriate response
For 3rd feature,
I'm retrieving all documents from Index for every user question and performing pandas queries using an agent.
I feel this is not an effective method.

Is there any way I can tackle this problem using Azure AI search?

Thanks in Advance for response