AI Times, a Korean AI publication, has published Managing Partner Yongduck Kim's column on generative AI patents.
In this column, we introduce the enterprise AI agent architecture for generative AI to safely and efficiently utilize a company's diverse data, focusing on U.S. Patent No. 12111859 of C3.ai, an American enterprise AI company. In a corporate environment, there is distributed data in different systems such as databases (DB), ERP, CRM, sensor data, email, and PDF documents. Rather than having one AI process all information, the patent proposes a structure in which an orchestrator and multiple specialist AI agents cooperate to solve problems. It is designed to utilize corporate data more accurately and reliably.
In the column, we examine from a patent perspective the process by which the orchestrator analyzes user queries, distributes tasks, processes structured data and unstructured data in an optimal manner, and then integrates the results. We also analyze how the structure that increases the reliability of answers through repetitive information verification, the type system that connects data from different systems into a common model, and the multi-agent structure in which multiple AI agents collaborate in parallel improve the accuracy and usability of generative AI for enterprises.
In addition, we note that this patent moves away from relying on a single Large Language Model (LLM) and presents a technological direction for the era of Agentic AI, where various specialist AI agents cooperate. These technologies are expected to become a core foundation for effectively utilizing corporate data in industries that require high reliability, such as manufacturing, energy, finance, defense, and healthcare, and are expected to have important implications in future enterprise AI and generative AI patent strategies.
The full article can be found at the link below.
This article reflects the information available when it was published. Contact us to discuss your circumstances.