From Rows to Reasoning: A Retrieval-Augmented Multimodal Framework for Spreadsheet Understanding

TL;DR

This paper introduces FRTR, a retrieval-augmented multimodal framework for understanding complex spreadsheets, significantly improving reasoning accuracy over large, multimodal enterprise spreadsheets compared to prior methods.

Contribution

The paper presents FRTR, a novel retrieval-augmented multimodal reasoning framework and a large-scale benchmark for complex spreadsheet understanding, addressing scalability and multimodal integration challenges.

Findings

FRTR achieves 74% accuracy on FRTR-Bench with Claude Sonnet 4.5.

FRTR attains 87% accuracy on SpreadsheetLLM with GPT-5.

FRTR reduces token usage by about 50% compared to serialization methods.

Abstract

Large Language Models (LLMs) struggle to reason over large-scale enterprise spreadsheets containing thousands of numeric rows, multiple linked sheets, and embedded visual content such as charts and receipts. Prior state-of-the-art spreadsheet reasoning approaches typically rely on single-sheet compression or full-context encoding, which limits scalability and fails to reflect how real users interact with complex, multimodal workbooks. We introduce FRTR-Bench, the first large-scale benchmark for multimodal spreadsheet reasoning, comprising 30 enterprise-grade Excel workbooks spanning nearly four million cells and more than 50 embedded images. To address these challenges, we present From Rows to Reasoning (FRTR), an advanced, multimodal retrieval-augmented generation framework that decomposes Excel workbooks into granular row, column, and block embeddings, employs hybrid lexical-dense retrieval with Reciprocal Rank Fusion (RRF), and integrates multimodal embeddings to reason over both numerical and visual information. We tested FRTR on six LLMs, achieving 74% answer accuracy on FRTR-Bench with Claude Sonnet 4.5, a substantial improvement over prior state-of-the-art approaches that reached only 24%. On the SpreadsheetLLM benchmark, FRTR achieved 87% accuracy with GPT-5 while reducing token usage by roughly 50% compared to direct serialization methods.