Hy-Search: A Hybrid Retrieval-Augmented Framework for Factual and Context-Aware Enterprise Knowledge Discovery

DOI

10.2991/978-94-6463-978-0_37How to use a DOI?

Keywords

Enterprise Search; Hybrid Search; Information Retrieval; Large Language Models; Retrieval-Augmented Generation

Abstract

Conventional search systems used in enterprise settings often find it difficult to overcome the semantic gap between intention and document content on one hand within their information seek, hence making the process of information seeking to be inefficient and productivity low. At the same time, the use of Large Language Models (LLMs) per se is impeded by their factual error concerns, or hallucinations, which makes them unreliable when it comes to tasks that require knowledge-intensive efforts within companies. The present paper presents Hy-Search, a new smart search model that uses a Hybrid Retrieval-Augmented Generation (RAG) model to help overcome these two challenges. This system is a combination of a dense semantic retriever and sparse lexical retriever, ingested with a tutorialized context into a generative LLM. Our full empirical assessment is based on a corpus of 15,000 technical documents of a giant company. Its results indicate that Hy-Search can perform at a state-of-the-art, extremely surpassing standard baselines of the search process based on keywords and pure dense retrieval. More importantly, the RAG architecture has a 73 per cent lower incidence of hallucinations per an isolated LLM, and user experiments have a 94 per cent task completion rate and 87 per cent preference over existing search systems. This publication confirms the fact that a hybrid RAG solution pro- is a scaleable, factuallyfounded, and user-friendly solution to next-generation 3rd generation enterprise knowledge discovery.

Copyright

© 2025 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Saisuman Singamsetty
AU  - Sudheer Singamsetty
PY  - 2025
DA  - 2025/12/31
TI  - Hy-Search: A Hybrid Retrieval-Augmented Framework for Factual and Context-Aware Enterprise Knowledge Discovery
BT  - Proceedings of the 1st Engineering Data Analytics and Management Conference (EAMCON 2025)
PB  - Atlantis Press
SP  - 431
EP  - 439
SN  - 2352-5401
UR  - https://doi.org/10.2991/978-94-6463-978-0_37
DO  - 10.2991/978-94-6463-978-0_37
ID  - Singamsetty2025
ER  -