Making Databases Searchable with Deep Context

TL;DR

This paper introduces Tursio, a platform that leverages large language models and semantic knowledge graphs to enable natural language search over enterprise databases, improving accessibility for non-expert users.

Contribution

The paper presents a novel architecture integrating LLMs into database query processing, creating an accessible natural language interface for complex enterprise data systems.

Findings

High accuracy in natural language query processing

Efficient and scalable performance on real workloads

Effective data modeling and query rewriting techniques

Abstract

Databases are the most critical assets for enterprises, and yet they remain largely inaccessible to people who make the most important decisions. In this paper, we describe the Tursio search platform that builds an abstraction layer, aka semantic knowledge graph, over the underlying databases to make them searchable in natural language. Tursio infuses large language models (LLMs) into every part of the query processing stack, including data modeling, query compilation, query planning, and result reasoning. This allows Tursio to process natural language queries systematically using techniques from traditional query planning and rewriting, rather than black-box memorization. We describe the architecture of Tursio in detail and present a comprehensive evaluation on production workloads, and synthetic and realistic benchmarks. Our results show that Tursio achieves high accuracy while being efficient and scalable, making databases truly searchable for non-expert users.