MADP用多智能体+人工介入，让企业文档处理自动化率超97%。
MADP: A Multi-Agent Pipeline for Sustainable Document Processing with Human-in-the-Loop
- 五类智能体协同分工，结合大模型提取与人工校验
- 年处理10万张发票可减少70%人力，97%流程无需人工干预
- 兼顾准确率与可持续性，碳排放降低69%
文档处理自动化在企业环境中仍面临挑战，传统人工方式耗时且易出错。本文提出MADP，一种多智能体架构，通过深度学习分类与解析结合大语言模型抽取，并引入人工在环（HITL）机制与新颖的反馈继承提示微调（PFTFI）方法，实现高精度自动化。系统包含分类器、拆分器、解析器、抽取器和验证器五个专用智能体。在年产10万张发票的生产场景中，全职人力需求预计减少约70%。2026年1月前实际部署于955份真实文档，全流程自动化率达97.0%，仅3%需非AI兜底。对分层抽样的100份文档（每类供应商/文档类型5份）进行消融实验显示，完整MADP配置在人工监督下达到98.5%文档级准确率。全面可持续性分析表明，相比传统人工处理，该混合AI+HITL方案可减少69%碳排放、69%能耗及63%用水量。多种大模型后端（Granite-Docling、Mistral-Small、DeepSeek-OCR）的基准对比为生产部署提供实用参考。
原文摘要 · Abstract (English)
Document processing automation remains a critical challenge in enterprise environments, where traditional manual approaches are labor-intensive and error-prone. We present MADP, a multi-agent architecture that addresses the challenge of automating document processing in enterprise settings by combining deep learning-based classification and parsing with large language model extraction, while maintaining accuracy through selective human validation. Our system integrates five specialized agents--Classificator, Splitter, Parser, Extraction, and Validator--with a Human-in-the-Loop (HITL) mechanism and a novel Prompt Fine Tuning with Feedback Inheritance (PFTFI) approach. The operational analysis on a production use-case scenario of 100,000 invoices per year indicates a potential reduction of Full-Time Equivalent (FTE) requirements by approximately 70%. Production deployment on 955 real-world documents processed through January 2026 achieves a 97.0% full-pipeline automation rate, with only 3% requiring non-AI fallback. Ablation evaluation on a stratified 100-document subset (5 documents per each of 20 supplier/document-type categories) demonstrates that the full MADP configuration with Human-in-the-Loop supervision attains 98.5% document-level accuracy. Additionally, we present a comprehensive sustainability analysis showing that our hybrid AI+HITL approach reduces CO2 emissions by 69%, energy consumption by 69%, and water usage by 63% compared to traditional manual processing. Benchmark comparisons of multiple LLM backends (Granite-Docling, Mistral-Small, DeepSeek-OCR) provide practical insights for deployment in production environments.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。