AI-Powered Database Querying Assistant with MCP

We have developed a PoC for an AI-powered database assistant that allows users without a corresponding technical expertise to query a PostgreSQL database using natural language questions in English. The solution executes queries against the Northwind sample dataset and returns human-readable answers.

Our specialists used Angular and Tailwind CSS to develop the solution’s interface that combines a tabbed database browser with a conversational AI chat panel. A .NET-based back end connects to Claude via the Anthropic SDK. Npgsql handles direct data access for the browsing interface. We used the Model Context Protocol (MCP) integration to grant the AI assistant read-only access to the database via the official MCP PostgreSQL server.

The framework demonstrated in this project can be used for connecting any relational database to an AI assistant for simplified data exploration across various business domains.

Diagram of an AI querying a database through an MCP server. A user asks "What are the top 5 best-selling products?"; the question goes to the AI, which writes the SQL and interprets the result; the AI sends a tool call to the MCP Server (database gateway), which passes a SQL query to the SQL Database that runs it; data rows return to the MCP Server, the tool result returns to the AI, and the assistant replies with a ranked list

POC FEATURES

Natural language icon
Database Querying with Requests in Natural Language
mcp integration icon
Model Context Protocol Integration
Data summaries icon
AI-Generated Data Summaries
Tool discovery icon
Dynamic MCP Tool Discovery at Runtime
Read only icon
Read-Only Database Access Enforcement
Schema context icon
Schema-Aware AI Context
Case study icon

Discover Case study

Get the complete story behind this solution — from the business reasoning to a detailed project breakdown and an in-depth overview of the solution’s key capabilities.

VIEW CASE STUDY