Agentic GraphRAG: Navigating Unstructured Financial Data with Collaborative AI
Arthur Capozzi, Dirk Helbing

TL;DR
This paper introduces a collaborative agentic GraphRAG framework that constructs a knowledge graph from unstructured legal texts and structured data to improve expert analysis of commercial registry information.
Contribution
The paper presents a novel multi-phase pipeline for building a knowledge graph and an agentic system that enhances retrieval, reasoning, and transparency in analyzing unstructured financial data.
Findings
Outperforms standard vector-RAG baseline in correctness and relevance.
Achieves high accuracy in entity resolution and answer quality.
Demonstrates effectiveness on Swiss commercial registry data.
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…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
