THOR: Transformer Heuristics for On-Demand Retrieval

TL;DR

THOR is a secure, scalable Text-to-SQL system that transforms natural language questions into verified SQL queries with fault tolerance and schema awareness, enabling non-technical enterprise data access.

Contribution

It introduces a decoupled architecture with integrated self-correction and schema-aware query generation for reliable enterprise data retrieval.

Findings

Demonstrates reliable ad-hoc querying in finance, sales, and operations.

Ensures enterprise-grade safety with read-only and compliance guardrails.

Enables non-technical users to access live data with zero-SQL complexity.

Abstract

We introduce the THOR (Transformer Heuristics for On-Demand Retrieval) Module, designed and implemented by eSapiens, a secure, scalable engine that transforms natural-language questions into verified, read-only SQL analytics for enterprise databases. The Text-to-SQL module follows a decoupled orchestration/execution architecture: a Supervisor Agent routes queries, Schema Retrieval dynamically injects table and column metadata, and a SQL Generation Agent emits single-statement SELECT queries protected by a read-only guardrail. An integrated Self-Correction & Rating loop captures empty results, execution errors, or low-quality outputs and triggers up to five LLM-driven regeneration attempts. Finally, a Result Interpretation Agent produces concise, human-readable insights and hands raw rows to the Insight & Intelligence engine for visualization or forecasting. Smoke tests across finance, sales, and operations scenarios demonstrate reliable ad-hoc querying and automated periodic reporting. By embedding schema awareness, fault-tolerant execution, and compliance guardrails, the THOR Module empowers non-technical users to access live data with zero-SQL simplicity and enterprise-grade safety.