Prompt-to-SQL Agent (T&D Workflows)

Purpose

Prompt-to-SQL Agent (T&D Workflows) translates natural-language questions into optimized, secure SQL queries using business context, data models, and permissions. It validates query logic, explains how the query works, and proposes reusable templates for Transmission & Distribution workflows.

Primary users

The provided primary user field is “nan”, so no specific end-user role is specified. The agent is associated with the T&D team or business unit and is owned by louis.thibaut@sia-partners.com.

Where it fits (process/stage/trigger)

This agent fits into Energy Transmission & Distribution data workflows when a user needs to query structured datasets through natural-language questions. It is triggered when textual metadata, database schemas, examples, or relevant table and document context are available to support SQL generation.

Key capabilities / workflow

The agent analyzes business context, textual metadata, database schemas, examples, and RAG context over tables and documents to generate optimized and secure SQL queries. It validates the query logic, checks whether results are valid, explains the query, extracts datasets, provides insights, and can propose reusable SQL templates.

Inputs

Typical inputs include textual metadata, database schemas, examples, and RAG over tables and documents. The provided dataset sources include fake internal procedures and web search.

Outputs / Deliverables

The agent produces validated SQL queries, extracted datasets, and insights. It also explains query logic and proposes reusable templates when relevant to the provided natural-language question and available context.

Value

The agent helps T&D users convert business questions into secure, validated SQL faster while improving consistency, transparency, and reuse. It supports more reliable access to data-driven insights by combining business context, schemas, examples, permissions, and validation in one workflow.