Hypergraph Enterprise Agentic Reasoner over Heterogeneous Business Systems

arXiv cs.AI Papers

Summary

HEAR is an enterprise agentic reasoner using a Stratified Hypergraph Ontology to perform multi-hop reasoning over heterogeneous business systems, achieving up to 94.7% accuracy on supply-chain tasks.

arXiv:2605.14259v1 Announce Type: new Abstract: Applying Large Language Models (LLMs) to heterogeneous enterprise systems is hindered by hallucinations and failures in multi-hop, n-ary reasoning. Existing paradigms (e.g., GraphRAG, NL2SQL) lack the semantic grounding and auditable execution required for these complex environments. We introduce HEAR, an enterprise agentic reasoner built on a Stratified Hypergraph Ontology. Its base Graph Layer virtualizes provenance-aware data interfaces, while the Hyperedge Layer encodes n-ary business rules and procedural protocols. Operating an evidence-driven reasoning loop, HEAR dynamically orchestrates ontology tools for structured multi-hop analysis without requiring LLM retraining. Evaluations on supply-chain tasks, including order fulfillment blockage root cause analysis (RCA), show HEAR achieves up to 94.7% accuracy. Crucially, HEAR demonstrates adaptive efficiency: utilizing procedural hyperedges to minimize token costs, while leveraging topological exploration for rigorous correctness on complex queries. By matching proprietary model performance with open-weight backbones and automating manual diagnostics, HEAR establishes a scalable, auditable foundation for enterprise intelligence.
Original Article
View Cached Full Text

Cached at: 05/15/26, 06:22 AM

# Hypergraph Enterprise Agentic Reasoner over Heterogeneous Business Systems
Source: [https://arxiv.org/abs/2605.14259](https://arxiv.org/abs/2605.14259)
Authors:[Ling Wang](https://arxiv.org/search/cs?searchtype=author&query=Wang,+L),[Songnan Liu](https://arxiv.org/search/cs?searchtype=author&query=Liu,+S),[Jianan Wang](https://arxiv.org/search/cs?searchtype=author&query=Wang,+J),[Cheng Cheng](https://arxiv.org/search/cs?searchtype=author&query=Cheng,+C),[Xin Liu](https://arxiv.org/search/cs?searchtype=author&query=Liu,+X),[Yihan Zhu](https://arxiv.org/search/cs?searchtype=author&query=Zhu,+Y),[Enyu Li](https://arxiv.org/search/cs?searchtype=author&query=Li,+E),[Yu Xiao](https://arxiv.org/search/cs?searchtype=author&query=Xiao,+Y),[Jiangyong Xie](https://arxiv.org/search/cs?searchtype=author&query=Xie,+J),[Duogong Yan](https://arxiv.org/search/cs?searchtype=author&query=Yan,+D),[Jiangyi Chen](https://arxiv.org/search/cs?searchtype=author&query=Chen,+J)

[View PDF](https://arxiv.org/pdf/2605.14259)

> Abstract:Applying Large Language Models \(LLMs\) to heterogeneous enterprise systems is hindered by hallucinations and failures in multi\-hop, n\-ary reasoning\. Existing paradigms \(e\.g\., GraphRAG, NL2SQL\) lack the semantic grounding and auditable execution required for these complex environments\. We introduce HEAR, an enterprise agentic reasoner built on a Stratified Hypergraph Ontology\. Its base Graph Layer virtualizes provenance\-aware data interfaces, while the Hyperedge Layer encodes n\-ary business rules and procedural protocols\. Operating an evidence\-driven reasoning loop, HEAR dynamically orchestrates ontology tools for structured multi\-hop analysis without requiring LLM retraining\. Evaluations on supply\-chain tasks, including order fulfillment blockage root cause analysis \(RCA\), show HEAR achieves up to 94\.7% accuracy\. Crucially, HEAR demonstrates adaptive efficiency: utilizing procedural hyperedges to minimize token costs, while leveraging topological exploration for rigorous correctness on complex queries\. By matching proprietary model performance with open\-weight backbones and automating manual diagnostics, HEAR establishes a scalable, auditable foundation for enterprise intelligence\.

## Submission history

From: Xin Liu \[[view email](https://arxiv.org/show-email/37917084/2605.14259)\] **\[v1\]**Thu, 14 May 2026 01:57:59 UTC \(352 KB\)

Similar Articles

HAGE: Harnessing Agentic Memory via RL-Driven Weighted Graph Evolution

Hugging Face Daily Papers

HAGE introduces a weighted multi-relational memory framework that enables query-conditioned traversal over unified relational memory graphs, improving long-horizon reasoning accuracy through adaptive memory retrieval and reinforcement learning-based optimization.

GraphReAct: Reasoning and Acting for Multi-step Graph Inference

arXiv cs.AI

This paper introduces GraphReAct, a framework that extends reasoning-acting paradigms to graph-structured data for multi-step inference. It combines topological and semantic retrieval with context refinement to improve performance on graph learning benchmarks.

AgenticRAG: Agentic Retrieval for Enterprise Knowledge Bases

arXiv cs.AI

This paper introduces AgenticRAG, a framework from Microsoft that enhances enterprise knowledge base retrieval by equipping LLMs with tools for iterative search, document navigation, and analysis. It demonstrates significant improvements in recall and factuality over standard RAG pipelines on multiple benchmarks.