Agentic GraphRAG: Navigating Unstructured Financial Data with Collaborative AI

2026-04-15

arXiv.org

AI Summary

This work addresses the challenge of efficiently and accurately analyzing public commercial registry data, which—despite being publicly accessible—contains critical facts scattered across vast volumes of multilingual, time-stamped records riddled with entity aliases. To tackle this, the authors propose an auditable, graph-augmented generative architecture that uniquely integrates deterministic structured ingestion, LLM-assisted unstructured information extraction, and deterministic entity disambiguation to construct a Neo4j knowledge graph comprising over five million nodes. The system employs intent routing, bounded reflection mechanisms, and state machine–guided response synthesis to support reliable complex querying. Human evaluation demonstrates that graph-based retrieval boosts factual accuracy from 0.26 to 0.83, substantially enhancing relevance and completeness; ablation studies confirm the contribution of each component, while also enabling auditable visualization of evidence chains and execution traces.

Abstract

We present a collaborative agentic GraphRAG framework for expert analysis of commercial registry data. Public registries are often formally accessible, yet difficult to use in practice because they combine structured records with large volumes of unstructured legal text. This limits conventional keyword and vector-only retrieval, especially for multi-hop, temporal, and entity-centric investigations. Our approach builds a Neo4j knowledge graph through a three-phase pipeline: (i) deterministic ingestion of strong nodes from verified structured fields, (ii) LLM-based extraction of weak nodes from unstructured notices, and (iii) deterministic identity resolution and deduplication. On top of this graph, we introduce an analytical modular agent that integrates zero-shot intent routing, a bounded reflection loop, secure tool-mediated graph access, and state-aware response synthesis. A human-in-the-loop dashboard exposes evidence and execution traces to support transparency and auditability. We evaluate the framework on the Swiss Official Gazette of Commerce, a multilingual corpus of more than seven million publications over seven years. We further contribute a multi-tier evaluation protocol covering entity-resolution precision, tool-routing behavior, answer quality, and multi-turn conversational performance. Across automated, human-curated, and conversational benchmarks, the proposed agentic GraphRAG system consistently outperforms a standard agentic vector-RAG baseline, with strong gains in correctness, answer relevance, information recall, turn success rate, and context carryover accuracy. The architecture is modular, reproducible, and transferable to other commercial gazettes and public-sector registry systems.

Problem

• commercial registry
• information fragmentation
• entity resolution
• multilingual legal notices
• auditable analysis

Innovation

• Agentic GraphRAG
• Knowledge Graph Construction
• Entity Resolution
• Retrieval-Augmented Generation
• Auditable AI