You have probably sat through a vendor demo that promised to collapse your workflow into minutes, then watched the clock tick past the old baseline. KT ran the opposite experiment. The South Korean telecommunications company held a two-day event in October at its Pangyo building to show what happens when employees build AI agents for their own work, then share the results across a group of affiliates. The format combined exhibits, presentation sessions and a live hackathon. The agenda covered business process improvements, innovation in manufacturing and logistics, workplace tooling and core technology R&D. KT framed the gathering as a step toward becoming an AX platform company, its term for an organisation built around AI transformation.
KT merged two existing events, AID-X Demo Day and AX Tech Connect, into a single showcase. The programme ran across four tracks: business, innovation, work and technology. In the business track, KT presented MAGMA BI/DW, a data analysis platform that accepts natural-language questions and returns dashboards without manual query writing. The company also demonstrated technology that analyses customer app reviews and consultation transcripts. KT said the AI cuts work that previously required about eight hours to within ten minutes.
The innovation track featured AI applications for manufacturing, logistics and mobility. Exhibits included a Robot Skill Flywheel that accumulates work experience from humanoid robots and uses that data to learn new tasks, a Maintenance and Warranty AX Agent Pack that predicts vehicle failures, and an urban air mobility traffic control support platform.
In the work track, KT introduced AX Works, a system that lets employees develop and share AI agents directly. The company also showed technology that links screen planning to design and code generation. KT reported that projects using the toolchain reduced development time by about 30 percent, from nine months to six months.
The technology track covered AI-based security vulnerability diagnostics, abnormal transaction detection, and next-generation network research including 6G and quantum security. KT shared R&D results for KT-Klo, a personalised expert agent applied to Everyone's AI, and Mideum K 3.0, the company's in-house AI model.
Presentation sessions focused on trial, error and iteration. KT walked through the build process for kode, an AI-based software development system created jointly by KT and KT DS, and a CFO agent that analyses financial data in natural language.
The event also included an AID-X hackathon in which 32 three-person teams built AI agents and work automation services to solve problems they encounter on the job.
KT reported three quantified outcomes. Customer analysis work dropped from about eight hours to within ten minutes when the company applied AI to app reviews and consultation content. Development projects using the screen-to-code pipeline finished in six months instead of nine, a reduction of roughly 30 percent. Thirty-two teams, each with three members, reached the final round of the internal hackathon.
KT's approach separates capability building from vendor dependency. AX Works gives employees the tools to prototype agents for their own workflows, then share working examples across the organisation. The model assumes that the people closest to a process understand its friction points better than a central team. When a customer service analyst can see the eight-hour task collapse to ten minutes, and a developer can watch a nine-month cycle shrink to six, the case for AI literacy stops being theoretical.
The hackathon structure matters because it treats AI fluency as a team sport. Small three-person squads push people to combine domain knowledge with technical build skills. That configuration mirrors the way most organisations will deploy AI: not as a solo coder's trick, but as a capability that crosses functions. The presentation sessions on trial and error signal that KT expects mistakes and wants people to learn from them in public. Organisations that hide failure stories end up repeating them in every department.
Ok Kyung-hwa, vice president and chief information officer at KT's IT division, said, "An important competitive edge of AX lies in accumulating field-verified experience as organisational capability and spreading it quickly." She added, "We will spread verified AX cases and know-how across the organisation to lead to AX innovation in which customers and employees can feel real change."
The risk for L&D leaders watching from outside is mistaking the event for a one-off showcase. KT merged two recurring programmes into a single calendar slot, which suggests the company intends to repeat the format. The value lies not in the October gathering but in the infrastructure that lets employees build, test and share agents year-round. If your organisation runs lunch-and-learns on prompt engineering but lacks a platform where people can deploy what they learn, you are training for a capability you have not enabled. KT's model works because AX Works exists before the demo day, not after it.
The broader stake is organisational learning speed. KT's goal is to move verified AI cases across the group faster than competitors can. That requires more than a training budget. It requires a system where an agent built in one division becomes a template in another, where a failed experiment in logistics informs a successful one in finance, and where the distance between "we tried this" and "now you can use it" shrinks to weeks instead of quarters. L&D's role is to build the scaffolding (the shared language, the peer learning rituals, the permission to prototype) that makes that speed possible.
---
Sources:
.png)
The WorkFlow podcast is hosted by Steve Glaveski with a mission to help you unlock your potential to do more great work in far less time, whether you're working as part of a team or flying solo, and to set you up for a richer life.
To help you avoid stepping into these all too common pitfalls, we’ve reflected on our five years as an organization working on corporate innovation programs across the globe, and have prepared 100 DOs and DON’Ts.
