VERAFI: Verified Agentic Financial Intelligence through Neurosymbolic Policy Generation

2025-12-12

AI Summary

Financial AI systems face dual challenges: while retrieval-augmented generation (RAG) enables document retrieval, large language models (LLMs) frequently commit computational errors and regulatory violations in financial reasoning. This paper proposes a trustworthy AI framework for finance, integrating dense retrieval, cross-encoder re-ranking, domain-specific tool-calling agents, and a novel neural-symbolic policy generation layer. The latter embeds GAAP/SEC regulations and mathematical verification as executable, logic-consistent reasoning strategies, enabling verifiable end-to-end decision-making. Evaluated on FinanceBench, the system achieves 94.7% factual accuracy—a relative improvement of 81% over baselines. Ablation analysis shows the neural-symbolic policy layer alone contributes a +4.3 percentage-point accuracy gain, significantly enhancing regulatory compliance and operational reliability.

Technology Category

• Philosophy and Ethics of AI: Safety, Robustness & Trustworthiness
• Natural Language Processing: Safety and Robustness
• Reasoning under Uncertainty: Other Foundations of Reasoning under Uncertainty

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Economics, Online Markets and Human Computation: Trust and reliance of crowd workers and data experts on GenAI
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Financial AI systems suffer from a critical blind spot: while Retrieval-Augmented Generation (RAG) excels at finding relevant documents, language models still generate calculation errors and regulatory violations during reasoning, even with perfect retrieval. This paper introduces VERAFI (Verified Agentic Financial Intelligence), an agentic framework with neurosymbolic policy generation for verified financial intelligence. VERAFI combines state-of-the-art dense retrieval and cross-encoder reranking with financial tool-enabled agents and automated reasoning policies covering GAAP compliance, SEC requirements, and mathematical validation. Our comprehensive evaluation on FinanceBench demonstrates remarkable improvements: while traditional dense retrieval with reranking achieves only 52.4% factual correctness, VERAFI's integrated approach reaches 94.7%, an 81% relative improvement. The neurosymbolic policy layer alone contributes a 4.3 percentage point gain over pure agentic processing, specifically targeting persistent mathematical and logical errors. By integrating financial domain expertise directly into the reasoning process, VERAFI offers a practical pathway toward trustworthy financial AI that meets the stringent accuracy demands of regulatory compliance, investment decisions, and risk management.

Problem

• Addresses calculation errors and regulatory violations in financial AI reasoning
• Integrates neurosymbolic policy generation for verified financial intelligence
• Enhances factual correctness in financial document analysis and compliance

Innovation

• Neurosymbolic policy generation for financial compliance
• Integrated dense retrieval with cross-encoder reranking
• Financial tool-enabled agents for automated reasoning validation