Langchain for SQL Data

[soumyabeebolt]

Hello,

I am trying to build a chatbot for the platform I am developing, that provides user support for how to use our platform and facilitates the user to ask questions about their data.
I am storing the support document in the pincone database and the user data is in CloudSQl postgres.
In my current implementation, I am using a Langchain agent to decide which tool to use:- if it should search in the pinecone or CloudSQL database. My database is quite large and it has 25 tables.
The agents seems to take a lot of time in returning the final response when asked questions about user data. ie deciding which tool to use, detecting schema of the DB, creating sql query, processing and return the output.
The user experience wouldn't the great as it takes up to 60seconds.
Any recommendations on how to approach the problem?