KT unveils 'Agentic On', an AI agent platform linked to corporate systems

Step-by-step linkage of tasks such as document review, approval request, and system input

(Seoul = Yonhap News) Reporter Park Hyeong-bin = As generative artificial intelligence (AI) evolves into an 'AI agent' that performs actual corporate tasks, KT [030200] is introducing a platform that supports task automation by connecting AI agents to a company's existing system.

KT announced on the 27th that it will unveil 'Agentic On', an agentic AI platform that helps companies create AI agents and apply and operate them in actual work.

Agentic On is designed to allow companies to introduce AI agents while utilizing existing data, business systems, and AI models.

Rather than building a new, separate work system, the key is to connect AI to the existing environment to create and operate agents suited to each company's work.

This platform connects a company's internal documents and data so that AI can utilize them, and provides integrated functions to manage AI models and agents. The explanation is that you can build an AI agent that suits your business purpose by combining the necessary functions without having to do complex programming yourself.

Another characteristic is that multiple agents divide roles and handle one task. In tasks that require multiple steps, from document review to approval request to business system input, each AI agent divides roles and processes the final results in connection with the business system.

In the case of repetitive tasks, it also provides a function to remember the work process performed once and automate it later.

KT explained that even if a task requires up to five steps of instructions at first, once it is performed, the entire process can be executed with just one instruction.

A security function that manages access and execution rights to corporate data and AI was also applied. It manages the authority of data, AI models, and agents according to corporate policy and provides safety measures to protect personal information and prevent error responses. The AI ​​agent's judgments, actions, and work performance history can also be tracked and managed.

KT said, "What is important in a company's AX is not just creating a single AI agent, but connecting AI to existing data, systems, and work processes to produce practical business results." KT said, "KT will support the business innovation and productivity improvement of companies by supporting the application of Agentic On to a variety of corporate environments, not limited to specific industries."