Routing End User Queries to Enterprise Databases

2026-01-27

AI Summary

This work addresses the challenge of accurately routing natural language queries in multi-database enterprise environments, where overlapping domains and semantic ambiguity often hinder precise query interpretation. To tackle this issue, the study introduces a structured reasoning mechanism into the query routing task for the first time, proposing a modular, reasoning-driven reranking strategy. This approach explicitly models schema coverage, structural connectivity, and fine-grained semantic alignment, effectively integrating the strengths of retrieval-based methods and logical inference. Evaluated on a benchmark dataset constructed from real-world scenarios, the proposed method significantly outperforms baseline approaches—including pure embedding-based models and direct prompting of large language models—across all metrics, demonstrating marked improvements in both routing accuracy and robustness.

Technology Category

• Data Mining & Knowledge Management: Intelligent Query Processing
• Planning, Routing, and Scheduling: Planning with Language Models
• Natural Language Processing: Question Answering

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Graph Algorithms and Modeling for the Web: Querying, indexing, and retrieval in Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

We address the task of routing natural language queries in multi-database enterprise environments. We construct realistic benchmarks by extending existing NL-to-SQL datasets. Our study shows that routing becomes increasingly challenging with larger, domain-overlapping DB repositories and ambiguous queries, motivating the need for more structured and robust reasoning-based solutions. By explicitly modelling schema coverage, structural connectivity, and fine-grained semantic alignment, the proposed modular, reasoning-driven reranking strategy consistently outperforms embedding-only and direct LLM-prompting baselines across all the metrics.

Problem

• query routing
• natural language queries
• enterprise databases
• NL-to-SQL
• database selection

Innovation

• query routing
• NL-to-SQL
• reasoning-based reranking
• schema alignment
• multi-database systems