AI-based development practices, including KDS 2.0, take hold
Applied to services including Token Factory and KT.com

KT is using artificial intelligence (AI) to unify the customer experience across all its digital channels, including its applications, website and text messages. It is standardizing screen layouts and guidance messages, which previously varied by service, and introducing AI throughout design and customer communications to improve both customer convenience and work efficiency.


KT employees are using KDS 2.0 and the AI-powered Message Generation Agent to design digital service screens and review customer notification messages. Courtesy of KT.

KT employees are using KDS 2.0 and the AI-powered Message Generation Agent to design digital service screens and review customer notification messages. Courtesy of KT.

View original image

KT said on November 12 that it had built KDS 2.0, a user experience (UX) design system with enhanced AI capabilities, and introduced an AI agent to help write and review customer guidance messages.


The key feature of KDS 2.0 is that it has been upgraded so AI can directly read and use design guidelines that previously served only as a reference for people. KT built the “KDS Harness,” which enables AI to create screens according to design standards and check its own output, reducing the burden of repeated design revisions and creating a consistent user experience.


KT said it improved development efficiency by applying KDS in September to My KT, KT.com and Direct Shop, among other services. In particular, during development of the AI model platform “Token Factory,” planning, design and development proceeded in parallel, and repetitive coding was automated, allowing work that would have taken about eight months to be completed in 2.4 months. This shortened the development period by more than five months.


Visual elements such as graphics, typefaces and photographs follow a “premium minimalist” approach, reducing unnecessary elements to improve readability. KT has also transformed its approach to customer communications through AI transformation (AX). KT sends about 170 million guidance text messages each month. It has replaced difficult telecommunications jargon with simpler expressions, such as “prepayment” with “pay in advance” and “usage-based billing” with “charges based on how much you use.” To put these changes into practice, KT officially launched the “Message Generation Agent” internally on October 30.


By entering only the key information, the agent generates and reviews messages in accordance with KT’s writing guidelines and legal requirements. It also handles image recommendations for next-generation rich communication services (RCS) messages and multilingual translation in one step. KT said that, after its introduction, the time needed to write and review a message fell from as long as 60 minutes to about five minutes, while image production time dropped from four hours to 15 minutes. Translation work that previously took more than two days can now be completed in 30 seconds.


In preliminary tests, customer response metrics rose by as much as 3.4 times. KT plans to introduce the “Customer Language Agent” on November 20 to refine customer-facing messages using simpler expressions. It will also extend KDS to external partners, a wider range of digital channels and AI devices, continuing its AI-based innovation.



Won Manho, head of KT’s AX Platform Division and an executive director, said, “Customers experience all of KT’s digital touchpoints—from apps and websites to text messages and consultations—as one. With KDS 2.0, we are improving design consistency and transforming customer communications through AX to enhance both work productivity and the customer experience.” He added, “We will continue to expand the scope of AX and create ‘one KT experience’ that is easy, consistent and trustworthy, no matter which touchpoint customers use.”


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

© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.

Today’s Briefing