MUMBAI, India, May 29 -- Intellectual Property India has published a patent application (202641064495 A) filed by Madhurima Kommuru, Hyderabad, Telangana, on May 21, for 'a system and method for adaptive retrieval-augmented generation in enterprise knowledge platforms.'
Inventor(s) include Madhurima Kommuru.
The application for the patent was published on May 29, under issue no. 22/2026.
According to the abstract released by the Intellectual Property India: "An adaptive retrieval-augmented generation system and method for enterprise knowledge platforms are disclosed. The system integrates heterogeneous enterprise data sources, including structured, semi-structured, and unstructured repositories, into semantically indexed knowledge representations. A query interpretation module determines contextual intent associated with user queries, while an adaptive retrieval engine selectively retrieves relevant knowledge segments. Retrieved contextual information is processed through a context compression module configured to summarize, filter, and prioritize information to reduce contextual payload size. A token optimization module dynamically allocates and manages token utilization according to model inference constraints and relevance parameters. A prompt construction module generates structured augmented prompts for processing by a generative artificial intelligence model to produce context-aware responses. An adaptive feedback module continuously evaluates retrieval relevance, response quality, and token consumption metrics to refine subsequent retrieval and optimization operations. The disclosed framework improves response accuracy, computational efficiency, scalability, and enterprise knowledge accessibility."
Disclaimer: Curated by HT Syndication.
