iATADocumentIntelligence: AI for Document Processing and Product Matching

In an increasingly digitized business environment focused on operational efficiency, automating document analysis is the key to reducing costs, time, and margin of error. This thesis describes the development of a Document Intelligence platform (iATADocumentIntelligence) that uses advanced artificial intelligence and computer vision technologies to automate the processing of various business documents. The system autonomously acquires and processes documents using OCR, computer vision, and natural language understanding techniques. The system can accurately recognize and interpret textual content and classify it in a way consistent with the semantic context. A supervised user interface ensures the reliability of the results by allowing operators to validate, correct, or refine the extracted data, creating a feedback loop that integrates artificial intelligence with human control. This "human in the loop" approach strikes a balance between automation and supervision, thereby improving the system's accuracy. The platform has demonstrated concrete benefits in real world scenarios, significantly improving operational efficiency compared to traditional methods. This thesis analyzes the main design choices, critical issues, and solutions, offering a comprehensive overview of the development process and intelligent automation's impact on business document flows.