Generative AI startup FortyTwoMaru said on Sept. 21 it filed patent applications in South Korea and the United States for a "large language model-based data representation generation method and system".

The technology has a large language model (LLM) analyse information in unstructured documents within a company and generate answers in the form of tables or charts that users want.

It is designed to verify figures included in the answers based on actual corporate data and provide visualised results by combining data calculation and inference functions. FortyTwoMaru said it can reduce hallucinations in small language models (sLLMs) for enterprises and support business decision-making.

The patent has completed registration approval in South Korea, and the company has also finished filing in the United States.

FortyTwoMaru also registered in South Korea a patent for a "report generation method and system using enterprise data" related to its agentic AI service, SiteBunny. Patent examinations are under way in the United States and Europe.

A "method for providing a conversational communication agent service based on a generative AI model" was filed in the United States in January after filings in South Korea and an international application under the Patent Cooperation Treaty (PCT).

FortyTwoMaru has filed or secured a total of 122 patents at home and abroad. That includes 63 in South Korea, 39 in the United States, 18 in Europe and 2 PCT international applications, of which 50 were filed since 2023.

The company plans to apply the related technology sequentially to its enterprise LLM solution.

FortyTwoMaru CEO Donghwan Kim (김동환) said, "Now is the time to accumulate technology assets as the agentic AI market opens in earnest," adding, "We will continue to expand domestic and overseas filings of core technologies in the LLM and agentic AI fields."