SQL agents may need both MCP tools and searched knowledge
A company data analyst SQL bot needs the right information before it can write useful SQL queries. One option is to connect a RAG-based made from Markdown files in a , with settings for document format, chunk size, and overlap before the files are embedded. Another option is to connect directly to systems such as GitLab through MCP.
The current BigQuery MCP setup only has a tool for running a query, so metadata and metric information still need to be provided from outside. The main design problem is how to split knowledge, such as table and metric definitions, from action tools, such as running the final query.
Key points
- The goal is to build a data analyst SQL bot that can write SQL queries.
- One option is a RAG built from Markdown files in a .
- Another option is direct MCP access to systems such as GitLab or BigQuery.
- The current BigQuery MCP only runs queries, so metadata and metric definitions must come from elsewhere.
- A hybrid setup may be needed: searched knowledge for context, MCP tools for actions.