KT Collects 122 AX Ideas as Staff Seek Automation for Repetitive Tasks, Not ‘AI That Can Do Anything’
IT DAILY ·
✦ AI Summary
122 AX ideas were submitted inside KT.
Many called for reducing repetitive work such as data collection and processing, and approvals and reviews.
KT is using the AX Design Clinic to promote standard workflows and reusable AI assets.
As companies race to adopt AI agents, KT received 122 AX ideas through an internal call for submissions. Many of the ideas focused on easing repetitive work such as data collection and processing, as well as approvals and reviews, rather than on complex AI services. Business demand also leaned toward tools that can handle repetitive tasks and simplify existing procedures, rather than an AI that can do anything.
KT employees are reviewing AI agent functions under development and how they could be applied to work through the "AX Design Clinic." Through this process, the company confirmed a trend in which frontline demand is concentrated on reducing repetitive work rather than on general-purpose, complex AI.
In response, KT plans to avoid one-off development by individual teams and instead build standard workflows. It also plans to turn reusable AI assets into companywide resources and expand them across the enterprise. The photo was provided by KT.
KT announced on the 9th that it is running an internal program called the "AX Design Clinic" to turn work innovation ideas proposed by employees into actual AX tasks. The program is designed to implement employees' proposed work innovation ideas as real AX projects.
The AX Design Clinic operates with joint participation from in-house AX experts and frontline employees. They analyze workflows and pain points and support AI adoption. The scope of support covers the entire process, from work diagnosis to workflow design, selection of implementation methods, development of AI agents and automation solutions, and verification.
KT has received a total of 122 work innovation ideas from 15 organizations inside the company. Of these, 71 were for automating data collection and processing, accounting for about 58.2%; 33 were for improving approval and review processes, or about 27.0%; and 12 were for generating reports and dashboards, or about 9.8%. Together, these three categories totaled 116 cases, or about 95% of all proposals.
The results show that AX demand on the ground is focused on improving existing operations rather than creating new services. The targets for such efficiency improvements included finding and organizing materials, approval requests, and report writing.
In the early days of generative AI adoption, use cases centered on document summarization, question-and-answer functions, and technical experimentation. But at the stage of actual business application, the key issues have shifted beyond AI model performance to how data is connected and how existing workflows are changed.
Based on this, KT is taking an approach that first analyzes the characteristics of frontline work and the data used, then selects the most suitable method from AI agents and automation technologies. KT avoids pushing ahead with agent development solely because AI can be applied, and is additionally reviewing whether the work can be handled with existing automation technologies.
In this program, KT placed added emphasis on reusability rather than the number of AI agents. This is because if similar agents are developed separately according to each organization’s needs, construction and operating costs can rise.
Developing similar agents separately for each organization also increases the likelihood of duplicate functions and data. KT is stressing reduced duplicate development over a race to increase the number of agents.
KT plans to establish standard workflows for verified frontline cases and develop reusable models. It is also considering reusing pre-developed functions and models in other organizations, with the goal of improving development efficiency. In addition, it plans to strengthen governance related to AI use.
This direction means managing in-house generative AI agents as assets. If separate agents are built for each organization, there is a chance of functional overlap, blind spots in management, and inconsistencies in work standards. KT is focusing on reducing functional overlap, management blind spots, and inconsistencies in work standards.
KT has not disclosed the planned reduction in work hours for the projects it is pursuing, nor has it disclosed the amount of cost savings. As a result, the program’s effectiveness can only be confirmed if it is backed by concrete performance indicators such as actual usage rates, shorter processing times, and fewer errors.
KT is advancing AI transformation through the AX Design Clinic. The areas targeted for AI transformation are purchasing and contracts, legal consultation reservations, and financial data analysis and reporting.
The SCM division is developing a "procurement and contract process integrated guidance agent." The target work involves procurement and software(SW) service contracts. The agent is being designed to provide guidance on regulations, procedures, and operating standards, as well as procurement and contract processes, work standards, forms, and how to use internal systems. It also aims to explain the relevant departments and follow-up procedures, reduce repeated inquiries to those departments, and build procurement and contract knowledge into an organizational asset.
The legal affairs division is pursuing a "daily legal consultation reservation and scheduling agent." The automation scope covers receiving consultation requests, assigning schedules, coordination, changes, cancellations, and guidance. The goal is to automate repetitive reservation-related tasks and reduce the workload of staff members.
In finance, KT is using an agent for querying, comparing, and analyzing management information based on IR materials. The information queried includes key management data such as performance, investment, and budget status, and it is queried in natural language. This is intended to reduce repetitive document searches and data organization work and to help users quickly check the management information they need.
KT plans to expand the cases identified and verified in the AX Design Clinic to other organizations and group affiliates. The goal is to move beyond AI tool usage at the individual or some-organization level and establish standardized work methods across the organization.
Chung Chan-ho, vice president and head of KT's IT Strategy Division, said the AX Design Clinic is not focused on creating a large number of AI agents, but on designing the agents and work methods needed on the front line and verifying their actual effects. He added that KT will work to accumulate and spread organization-level AX assets from verified cases and advance the quality of AI-based work methods.
Source: IT DAILY · Lee Jae-young
Original: https://www.itdaily.kr/news/articleView.html?idxno=241491
References
This article was produced with the help of an automated content generation algorithm.
Source: IT DAILY
View originalThis article was summarized and organized by BizCrush based on the original article from IT DAILY. For exact quotations and full details, please refer to the original article.