Data Copilot Reference Implementations

Overview

Data Copilot is a set of patterns and reference implementations for building AI agents capable of interacting with structured data, primarily through Text-to-SQL and automated data analysis. This directory contains detailed schemas, guides, and templates for implementing these patterns.

Components

Implementation	Description
Skeleton Guide	Standardized project structure for Data Copilot implementations.
Answer Synthesis Schema	JSON schemas for structuring natural language answers from data results.

Related Architecture & Patterns

• Text-to-SQL Architecture
• Agentic RAG
• MCP Tooling
• SQL Validation Playbook

Typical Workflow

1. Query Intent Recognition: Understanding what the user is asking for in terms of data.
2. Schema Retrieval: Fetching relevant database schema information.
3. SQL Generation: Generating a valid SQL query based on the intent and schema.
4. Execution & Validation: Running the query and validating results.
5. Answer Synthesis: Converting raw data into a natural language response.

Sources / References

• Data Copilot Project Case Study
• Verification Report

Contribution Metadata

• Last reviewed: 2026-06-05
• Confidence: high