MUMBAI, India, Sept. 22 -- Intellectual Property India has published a patent application (202641108519 A) filed by Geethanjali College Of Engineering And Technology, Hyderabad on September 09, 2026, for Enterprise Knowledge Assistant – Agentic Rag System For Hr, It And Finance Data.
Inventors include Dr. Sujatha Dandu; Dr. L. Kiran Kumar Reddy; S. Sai Shruthi; V. Lalithya Sai; V. Krisha Reddy; A. Srinidhi; and A. Indhumathi.
The application for the patent was published on September 18, 2026, under issue no. 38/2026.
Abstract: The present invention relates to an Enterprise Knowledge Assistant – Agentic RAG System for HR, IT and Finance Data that provides intelligent, context-aware, and grounded access to organizational knowledge through natural-language interaction. The proposed system integrates enterprise data ingestion, document preprocessing, semantic embeddings, vector retrieval, structured database querying, Large Language Models, specialized departmental agents, agentic orchestration, access control, response validation, source citation, and conversational interaction within a unified architecture. The system receives structured and unstructured enterprise information from Human Resources, Information Technology, Finance, document repositories, knowledge bases, and other authorized organizational sources. The proposed Agentic RAG architecture dynamically analyzes user queries and determines the appropriate sequence of retrieval, database, analytical, and reasoning operations. Specialized HR, IT, and Finance agents can operate individually or collaboratively to answer domain-specific and cross-domain questions. The system combines semantic retrieval with keyword, metadata, and structured-data retrieval to identify relevant information and employs controlled database and tool execution for analytical tasks. A validation mechanism evaluates generated responses against retrieved evidence to reduce hallucinations and improve reliability, while source citation mechanisms provide traceability to supporting enterprise information. The proposed invention enables organizations to provide employees with a centralized intelligent knowledge assistant capable of retrieving policies, answering departmental questions, analyzing structured information, summarizing enterprise documents, supporting technical troubleshooting, generating reports, and resolving complex multi-domain queries. By combining Agentic AI and Retrieval-Augmented Generation with enterprise security, authorization, evidence validation, and conversational interaction, the system provides a scalable and transparent approach for intelligent enterprise knowledge management.
Disclaimer: Curated by HT Syndication.
