MUMBAI, India, Aug. 12 -- Intellectual Property India has published a patent application (202621075930 A) filed by Symbiosis International Deemed University on June 17, 2026, for Ai Driven Accounting And Bookkeeping System With Hybrid Nlp-Ml Transaction Categorization.
Inventors include Dr. Akhil Gupta; Anubhav Sethi; Tilak Barapatre; and Om Gadbail.
The application for the patent was published on August 07, 2026, under issue no. 32/2026.
Abstract: ABSTRACT AI DRIVEN ACCOUNTING AND BOOKKEEPING SYSTEM WITH HYBRID NLP-ML TRANSACTION CATEGORIZATION The present invention discloses an AI driven accounting and bookkeeping system (100) for micro, small and medium enterprises comprising a hybrid prediction engine (110) that fuses a rule-based natural language processing module (112) with a trained machine learning model (114) through a precedence-based decision logic unit (116) for automated transaction categorization with confidence scoring. The system (100) further comprises an OCR receipt scanning pipeline (120) with a multi-stage image preprocessing module (122) and a text extraction engine (124) for digitizing physical receipts, a GST computation module (130) for automated category-wise tax calculation aligned with Indian GST schedules, a human-in-the-loop approval workflow module (140) with a three-state pending-approved-rejected transition mechanism, a relational database layer (150) providing ACID-compliant persistent storage, and a confidence analytics dashboard (160). The hybrid engine (110) achieves approximately eighty-nine percent categorization accuracy by leveraging keyword-based text analysis and amount-based statistical modeling. [
Disclaimer: Curated by HT Syndication.
