AI Schedules and Records Meetings Based on Preferences... KT to Commercialize Hyper-Personalized AI Agents

Commercialization of Hyper-Personalized and Vertical AI Agents in the Second Half of the Year

"Research Underway on Future-Oriented Agentic AI in the Form of OpenClaw"

KT's Proprietary 'K RAG' Validated... "Effectiveness Under Review"

#. AI agents (virtual assistants) gather in a chat room to schedule a meeting. They share their users' individual calendars to find available dates, and when deciding on the meeting location, they conduct a vote that reflects their owners' preferences. Once the human makes the final decision, the AI records the schedule in a memo.

Kim Junseok, Head of the Agentic AI Lab at KT AX Future Technology Institute, is presenting KT's AI agent status and development strategy on the 17th at KT West Building in Gwanghwamun, Seoul. Photo by Noh Kyungjo

KT will commercialize hyper-personalized artificial intelligence (AI) agents in the second half of this year. In addition, the company is developing a "future-oriented agentic AI" in the form of OpenClaw, where AI autonomously completes tasks, aiming to transform daily life and industrial sites.

On the 17th, KT announced that it will introduce its enhanced AI agents, including My K, Genie TV, and Market Easy, in the second half of the year by applying "long-term memory-based technology" and customer situation-tailored "actionable agent technology." Based on technology that utilizes structured data from users' preferences and past behaviors, the AI will not only make recommendations but also actually perform tasks such as changing subscription plans.

The company will also continue its research into "future-oriented agentic AI," which maximizes the autonomous execution capabilities of AI. This concept moves beyond simply providing methods to users; the AI will actually complete work processes, such as clicking buttons or moving files. The goal is to create a personalized AI agent that proactively identifies the user's potential needs—even those the user is not consciously aware of—suggests relevant tasks, autonomously plans how to address them, and executes the tasks directly.

Kim Junseok, Head of Agentic AI Lab at KT AX Future Technology Institute, stated, "In the future, agents will operate through various interfaces beyond just chat windows. We are currently researching autonomous agentic AI in the form of OpenClaw, where AI agents directly perform practical tasks on users' PCs and in cloud environments."

In the business-to-business (B2B) domain, KT is advancing vertical AI agents with expert capabilities that leverage industry-specific data to provide solutions. The company plans to secure domain-specialized models for areas such as network operations and legal services, and will validate their effectiveness by identifying successful cases in the second half of the year. In addition, KT's proprietary Retrieval-Augmented Generation (RAG) technology will be expanded to both business-to-consumer (B2C) and B2B agent services. RAG is a technology in which the AI searches for and verifies supporting information before generating a response. KT is currently applying RAG to various internal agent services, including KT Knowledge Hub and My K, to validate its effectiveness.

A KT representative said, "Going forward, we plan to pursue strategic partnerships on agentic AI with organizations such as the South Korea-China-Japan telecommunications alliance (SCFA). Through these efforts, we will generate synergy between companies and realize agentic AI that works immediately in customers' daily lives and in industrial settings."

This content was produced with the assistance of AI translation services.