MUMBAI, India, Nov. 2 -- Intellectual Property India has published a patent application (202621112107 A) filed by Chinmay Lalit Metha; Kaustubh Karne; Chitrakant Omprakash Banchor; Vinayak Tukaram More; and Kartik Mehta on September 18, 2026, for Agentic Conversational Honeypot System For Scam Detection, Adaptive Engagement, And Intelligence Extraction.
Inventors include Chinmay Lalit Metha; Kaustubh Karne; Chitrakant Omprakash Banchor; Vinayak Tukaram More; and Kartik Mehta.
The application for the patent was published on October 30, 2026, under issue no. 44/2026.
Abstract: The invention provides an Agentic AI-based conversational honeypot system for autonomous scam detection, adaptive scammer engagement, and structured intelligence extraction. The system detects suspected scam communications and engages malicious actors through adaptive multi-turn conversations to collect actionable threat intelligence. A conversation state management mechanism maintains contextual information, while an intelligence extraction mechanism identifies structured entities from ongoing interactions. A decision mechanism dynamically selects engagement strategies, and a response generation mechanism produces context-aware replies using a Large Language Model (LLM). A learning mechanism continuously evaluates conversation outcomes and optimizes future engagement strategies. The invention integrates autonomous reasoning, adaptive conversation management, continuous intelligence extraction, and feedback-driven learning into a unified framework for conversational cyber deception and threat intelligence collection.
Disclaimer: Curated by HT Syndication.
