Bypassing Document Ingestion: An MCP Approach to Financial Q&A

Sasan Mansouri; Edoardo Pilla; Mark Wahrenburg; Fabian Woebbeking

cs.IRMar 19, 2026

Bypassing Document Ingestion: An MCP Approach to Financial Q&A

Authors: Sasan Mansouri, Edoardo Pilla, Mark Wahrenburg, Fabian Woebbeking

Organizations: University of Groningen · Goethe University Frankfurt · Halle Institute for Economic Research (IWH) and Martin Luther University Halle-Wittenberg

Abstract

Answering financial questions is often treated as an information retrieval problem. In practice, however, much of the relevant information is already available in curated vendor systems, especially for quantitative analysis. We study whether, and under which conditions, Model Context Protocol (MCP) offers a more reliable alternative to standard retrieval-augmented generation (RAG) by allowing large language models (LLMs) to interact directly with data rather than relying on document ingestion and chunk retrieval. We test this by building a custom MCP server that exposes LSEG APIs as tools and evaluating it on the FinDER benchmark. The approach performs particularly well on the Financials subset, achieving up to 80.4% accuracy on multi-step numerical questions when relevant context is retrieved. The paper thus provides both a baseline for MCP-based financial question answering (QA) and evidence on where this approach breaks down, such as for questions requiring qualitative or document-specific context. Overall, direct access to curated data is a lightweight and effective alternative to document-centric RAG for quantitative financial QA, but not a substitute for all financial QA tasks.

Figures & tables

Figure 1: LSEG APIs MCP server system architecture diagram.

Name	Arguments	Description
get_acquisitions	comp_name end_date start_date	Finds relevant M&A data
get_balancesheet_statement	comp_name period scale	Retrieves relevant balance sheet statement
get_business_segments	comp_name period scale	Retrieves relevant business segment revenue data
get_capital_structure	comp_name period scale	Retrieves relevant capital structure data
get_cashflow_statement	comp_name period scale	Retrieves relevant cash flow statement
get_earningscall_transcript	comp_name period	Outputs relevant earnings call transcript

Table 1: Implemented MCP tools.

Figure 2: FinDER average numeric density across answer and references by category. Carriage returns are removed from the Answer variable to avoid inflating the underlying number counts. Numeric density is defined at the granular level as the ratio between number and non-space substring occurrences.

Figure 3: FinDER average forward-looking word count across answer and references by category.

Parameter	Value
Parameter	Generator	Evaluator
max_completion_tokens	2048	4096
seed	17	17
temperature	0.01	0.01
top_p	0.15	0.15

Table 2: Baseline model parameters.

Figure 4: LSEG-powered LLM performance on FinDER by question category. Relative values computed over 990 and 4713 entries respectively for Financials and Other samples.

Figure 5: LSEG-powered LLM performance on FinDER by context quality and question category. An answer is defined as of high quality if the associated Context Relevance is equal or larger than 0.75. Relative values computed over 668 and 636 entries respectively for Financials and Other samples in the High-quality group, and over 322 and 4077 entries respectively for Financials and Other samples in the Low-quality group.

Figure 6: LSEG-powered LLM performance on FinDER Financials subset by context quality and task type. An answer is defined as of high quality if the associated Context Relevance is equal or larger than 0.75. Limited sample size reduces significance of results for Logical reasoning group.

Figure 7: LSEG-powered LLM performance on FinDER Financials subset by number occurrence group and question type. Limited sample size reduces significance of results for "No number occurrences" and "101 or more occurrences" groups.

Appendix figures & tables3 assets

Supplementary material from the paper’s appendix.

Appendix

Figure A.1: FinDER High-quality answer count (LHS) vs. average Answer Accuracy (RHS) by category.

Figure A.2: FinDER Context Relevance (LHS) vs. Response Groundedness (RHS) answer count by category.

Figure A.3: FinDER average Context Relevance (LHS) vs. average Response Groundedness (RHS) across number occurrences by category.

Financial QA

Financial question answering develops systems that answer questions about reports, markets, regulations, and other financial data with reliable evidence and reasoning.

See all 51 papers

Agentic Retrieval-Augmented Generation for Financial Document Question Answering
May 6, 2026Yang Shu, Yingmin Liu, Zequn XieFinancial QA Agentic RAG
College of Computer Science and Technology, Zhejiang University, Hangzhou, 310027, China
Enhancing Financial Question Answering: A Novel Benchmark Dataset of Banks' financial statements
Sep 3, 2026Arianna Miola, Bruno Spaccavento, Lorenzo Silotto +2Financial QA RAG Evaluation
Intesa Sanpaolo Innovation Center, Corso Inghilterra 3, 10138, Torino, Italy. · Universit`a degli Studi di Milano-Bicocca, Milan, Italy · Politecnico di Torino, Corso Duca degli Abruzzi 24, 10129, Turin, Italy. +2
FinMCP-Bench: Benchmarking LLM Agents for Real-World Financial Tool Use under the Model Context Protocol
Mar 26, 2026Jie Zhu, Yimin Tian, Boyang Li +8LLM Agent Evaluation LLM Tool Use
Qwen DianJin Team, Alibaba Cloud Computing · YINGMI Wealth Management · Soochow University

May 6, 2026J/K move · Enter open · S save

Agentic Retrieval-Augmented Generation for Financial Document Question Answering

Yang Shu, Yingmin Liu, Zequn Xie

College of Computer Science and Technology, Zhejiang University, Hangzhou, 310027, China

Financial QA Agentic RAG Finance LLM Reasoning

Open paper AskPDF arXiv

Financial document question answering (QA) demands complex multi-step numerical reasoning over heterogeneous evidence--structured tables, textual narratives, and footnotes--scattered across corporate filings. Existing retrieval-augmented generation (RAG) approaches adopt a single-pass retrieve-then-generate paradigm that struggles with the compositional reasoning chains prevalent in financial analysis. We propose FinAgent-RAG, an agentic RAG framework that orchestrates iterative retrieval-reasoning loops with self-verification, specifically engineered for the precision requirements of financial numerical reasoning. The framework integrates three domain-specific innovations: (1) a Contrastive Financial Retriever trained with hard negative mining to distinguish semantically similar but numerically distinct financial passages, (2) a Program-of-Thought reasoning module that generates executable Python code for precise arithmetic rather than relying on error-prone LLM-based mental computation, and (3) an Adaptive Strategy Router that dynamically allocates computational resources based on question complexity, reducing API costs by 41.3% on FinQA while preserving accuracy. Extensive experiments on three benchmark datasets--FinQA, ConvFinQA, and TAT-QA--demonstrate that FinAgent-RAG achieves 76.81%, 78.46%, and 74.96% execution accuracy respectively, outperforming the strongest baseline by 5.62--9.32 percentage points. Ablation studies, cross-backbone evaluation with four LLMs, and deployment cost analysis confirm the framework's robustness and practical viability for financial institutions.

Abstract

Financial QA

Explore similar work

Agentic Retrieval-Augmented Generation for Financial Document Question Answering

Enhancing Financial Question Answering: A Novel Benchmark Dataset of Banks' financial statements

FinMCP-Bench: Benchmarking LLM Agents for Real-World Financial Tool Use under the Model Context Protocol