MUMBAI, India, Aug. 12 -- Intellectual Property India has published a patent application (202621036768 A) filed by Persistent Systems on March 26, 2026, for System And Method For Dynamic Sensitivity Classification Of Data Entities In Ai Interaction Payloads.
Inventor includes Mr. Nitish Shrivastava.
The application for the patent was published on August 07, 2026, under issue no. 32/2026.
Abstract: The present disclosure relates to a system and method for constructing an organizational knowledge graph (400) to enable real-time dynamic sensitivity classification of data entities within artificial intelligence interaction payloads. The system integrates heterogeneous enterprise data sources through automated ingestion pipelines (420a–420f) including HRIS, CRM systems, project management platforms, financial systems, code repositories, and curator-managed intelligence sources. A knowledge graph repository (410) maintains multiple entity registries including a personnel registry (411), a customer registry (412), a project registry (413), a financial registry (414), a technical registry (415), and competitive registry (416) storing entity attributes, relationships, and sensitivity classifications. A relationship context traversal (440) utilizing HNSW vector indexing identifies previously unknown sensitive entities using similarity thresholds and contextual sensitivity amplification. A graph query engine (430) processes classification queries, while a compliance mapping module (450) enforces regulatory constraints. Temporal versioning and curator review queue (470) support continuous updates and validation.
Disclaimer: Curated by HT Syndication.
