AI specialist 42Maru has received a patent registration decision in South Korea for its large language model (LLM)-based data visualization technology and has completed its U.S. patent application, aiming to strengthen global competitiveness in the enterprise AI market. The patent for a "Method and System for Generating Data Representation Based on Large Language Model" covers technology enabling an LLM to read information in unstructured enterprise documents and automatically organize requested answers into tables or charts, with a verification procedure checking whether figures in generated responses match actual enterprise data to suppress hallucination in enterprise-grade small language models (sLLMs) and improve accuracy. Beyond this patent, 42Maru is expanding its patent portfolio in agentic AI: a patent for a "Method and System for Generating Reports Using Enterprise Data," related to its conversational agentic AI service SiteBunny, has completed domestic registration in South Korea with examinations underway in the United States and Europe; a patent for generative AI-based conversational communication agent services has completed South Korean filing and Patent Cooperation Treaty (PCT) international application, followed by a U.S. application completed in January of this year; and the company recently filed a new patent for a security system leveraging language models. 42Maru's cumulative patents filed or secured domestically and internationally now total 122, comprising 63 in South Korea, 39 in the United States, 18 in Europe, and 2 PCT international applications, with 50 filed since 2023.