About IntentSQL

Understand the question.
Get the data.

IntentSQL brings natural language understanding to relational data, allowing users to ask business questions without writing SQL manually.

Natural Language to SQL

What is IntentSQL?

IntentSQL is an AI-powered natural-language-to-SQL system that allows users to interact with business data using everyday language.

Instead of writing SQL queries manually, users describe what they want to know. IntentSQL interprets the question, understands the underlying intent, uses the available database structure, generates an appropriate query, executes it, and presents the resulting data.

The Idea

Ask questions, not queries.

Business users often know what information they need, but they should not have to understand database schemas, joins, aggregation functions, or SQL syntax to obtain it.

IntentSQL explores how natural language can become a more intuitive interface for interacting with structured business data.

The Workflow

How IntentSQL works

1
Ask a question
Describe the business information you need using natural language.
2
Understand the intent
The system interprets what the question is asking for.
3
Understand the data structure
Available tables, columns, and relationships provide the context required to construct the query.
4
Generate SQL
The natural-language request is translated into an SQL query.
5
Execute and return results
The query is executed against the underlying data and the results are presented to the user.
Artificial Intelligence

Large Language Model

IntentSQL uses a Large Language Model (LLM) to understand natural-language business questions and translate them into SQL queries.

The LLM works together with database schema information so that the generated query can be aligned with the available tables, columns, and relationships.

💬
Natural Language

Users communicate using ordinary business questions rather than SQL syntax.

🧠
Intent Understanding

The system interprets the meaning and analytical requirements behind the question.

🗄️
SQL Generation

The interpreted request is translated into a query that can be executed against the data.

✓
Read-only Analytics

Designed for safe exploration

The current implementation focuses on read-only analytics. Generated queries are restricted to SELECT statements and are intended to retrieve information without modifying the underlying data.

Transparency

See what happened behind the answer.

IntentSQL keeps execution information for each question so that generated queries and system behavior can be inspected rather than hidden behind a simple answer.

  • ✓ Generated SQL
  • ✓ Query results
  • ✓ Token usage
  • ✓ Processing time
  • ✓ SQL generation attempts
  • ✓ Execution status
Current Scope

A working proof of concept

IntentSQL is currently a proof of concept exploring natural-language interaction with structured business data.

The focus is on understanding how effectively an AI-powered system can translate real business questions into database queries while keeping the generated SQL, execution process, and results visible for inspection.