MUMBAI, India, Jan. 9 -- Intellectual Property India has published a patent application (202541113103 A) filed by Gayathri. N; Dinesh J; Manickam. S; Sanjeevi Kumar. M; and Krishnan. T, Coimbatore, Tamil Nadu, on Nov. 18, 2025, for 'agentic ai browser workflow.'
Inventor(s) include Gayathri. N; Dinesh J; Manickam. S; Sanjeevi Kumar. M; and Krishnan. T.
The application for the patent was published on Jan. 9, under issue no. 02/2026.
According to the abstract released by the Intellectual Property India: "The Agentic Al Browser Workflow represents a transformative approach to intelligent web automation by integrating Large Language Models (LLMs) and Computer Vision within a coordinated multi-agent framework. The invention is designed to autonomously perform browser-based workflows such as data extraction, form submission, content analysis, and website interaction without the need for static scripts or predefined selectors. At its core, the system employs a Vision Module, a Reasoning Module, a Memory Module, and an Executor Module, all orchestrated by a Coordinator Agent. The Vision Module visually interprets web pages, recognizing elements and layouts through image-based understanding, while the Reasoning Module utilizes natural language reasoning to plan and sequence actions based on contextual understanding. The Memory Module stores taskrelated information for multi-step workflows, enabling continuity and adaptive behavior, whereas the Executor Module performs real-time browser interactions such as clicking, scrolling, and form filling. The operational process begins when a natural language instruction is provided by the user. The Reasoning Module converts this instruction into an actionable workflow, the Vision Module identifies the necessary interactive elements, and the Executor Module executes corresponding actions through automated browser control tools. The Memory Module ensures that context and prior states arc preserved across multiple steps, and the Coordinator Agent supervises overall synchronization between the modules. This invention eliminates the brittleness of traditional automation frameworks that depend on static DOM structures, offering instead a perceptive and adaptive automation mechanism that functions effectively across varied and dynamic web environments. By combining perception, reasoning, and execution within an agentic architecture, the system significantly enhances automation reliability, reduces maintenance overhead, and enables generalpurpose browser interaction. The Agentic Al Browser Workflow thus establishes a foundation for next-generation web automation systems that are context-aware, flexible, and capable of autonomous decision-making, bringing human-like adaptability to browserbased workflows."
Disclaimer: Curated by HT Syndication.
