MUMBAI, India, Sept. 28 -- Intellectual Property India has published a patent application (202641111454 A) filed by Sns College Of Technology on September 17, 2026, for Integrated Ai-Based Cross-Document Evidence Extraction, Consistency Verification And Workflow Decision Support System.
Inventors include Krishna Kiran P; Jainitha R S; Swetha T; Mathana Gopal G M; Lalu Prasadh. V; Madhan Venue. M; Yogasri V; Pranesh J; Vishnupriya A; and Aswini Selvakumari S.
The application for the patent was published on September 25, 2026, under issue no. 39/2026.
Abstract: The present invention relates to an integrated Artificial Intelligence (AI)-based document intelligence and workflow decision support system configured to extract, correlate, verify, and analyze information distributed across multiple heterogeneous business documents. The proposed system provides an intelligent platform that converts unstructured and semi-structured documents into interconnected evidence representations and automatically identifies inconsistencies, missing information, temporal conflicts, and unsupported claims across related documents. The invention employs Optical Character Recognition (OCR), Natural Language Processing (NLP), machine learning, semantic embeddings, entity extraction, event extraction, temporal analysis, and relationship modeling techniques to identify relevant facts from documents including invoices, purchase orders, delivery notes, receipts, contracts, reports, forms, emails, and other business records. Unlike conventional document processing systems that independently extract fields from individual documents, the proposed system establishes relationships between extracted facts, entities, events, quantities, dates, and claims originating from multiple documents. The system generates a structured evidence graph in which each extracted fact is associated with its source document, location, confidence score, timestamp, entity relationship, and supporting or contradicting evidence. A consistency verification engine analyzes the evidence graph to identify contradictory values, missing evidence, temporal inconsistencies, duplicate information, incomplete transactions, and deviations from predefined business rules. The system further generates evidence-grounded explanations describing the source documents and relationships responsible for each detected inconsistency. A workflow decision support engine uses the verified evidence state to determine appropriate next actions, including automatic approval, rejection, escalation, additional-document requests, human review, payment hold, or workflow routing. Confidence-aware processing enables low-confidence or conflicting cases to be transferred to human operators while allowing high-confidence cases to proceed automatically. The proposed invention thereby extends AI document processing from simple information extraction to cross-document evidence understanding, consistency verification, and evidence-driven workflow decision support, reducing manual verification effort, improving process accuracy, and enabling auditable business automation.
Disclaimer: Curated by HT Syndication.
