Park Hye-young, a Senior Solutions Architect at Amazon Web Services (AWS), emphasized that the success of artificial intelligence (AI) adoption in companies is most effective when led by employees who are directly involved in the work.
On July 9, at the '2026 ABC Launch Commemoration AI Ecosystem Innovation Forum' held at The Plaza Hotel in Seoul, Park presented a session titled 'How Successful Organizations Implement Enterprise AI.'
Park explained why only a small number of companies achieve results after adopting AI and how they can utilize AI to reduce costs and improve profitability, citing examples from the domestic oil refining and media industries.
According to McKinsey's Global AI Survey 2025, while 88% of companies have adopted AI, only 6% have seen improvements in operating profit. Park pointed out that the issue lies not with the AI technology itself but with the implementation stages and methods.
He identified the lack of success in AI adoption as stemming from a 'top-down approach led by IT departments' and stagnation during the proof of concept (PoC) phase. Even if internal IT departments build systems, they often fail to accurately diagnose real-world problems, and if the AI systems are not used in practice, it is difficult to achieve a return on investment.
About 30% of AI pilot programs do not lead to company-wide implementation. He stressed, 'It is more valuable to properly apply one project to actual work than to conduct ten pilots.'
Park introduced real-world AI adoption success stories from GS Caltex, LG Electronics' Korean Sales Division, KBS, and CJ ENM.
GS Caltex, where two-thirds of employees work in production, achieved results by cultivating on-site staff such as process engineers, equipment managers, and quality controllers as AI-utilizing personnel through a bottom-up approach. He noted, 'GS Caltex allowed on-site employees to diagnose needs and create AI agents, resulting in the production of over 1,000 AI agents, with 85% of employees currently using them in their work.'
LG Electronics' Korean Sales Division established a structure where collaborative employees propose automation tasks, and the IT organization combines technical components like prompts, images, videos, market research, and predictive models to implement agents. Park stated, 'This has enabled marketing automation that analyzes problems and consumer responses through AI, focusing on customer relationship management (CRM).'
He also shared examples of KBS and CJ ENM using AI to reduce costs and generate new revenue while broadcasting K-pop content globally. KBS developed the 'Vertigo' solution, which automatically generates individual fan cams of idol members using just one 8K camera. The post-production process, which used to take hours or days for editing, color correction, and encoding, has been reduced to just a few minutes.
Park explained, 'In the past, producing fan cam videos for idol groups required as many cameras and personnel as there were members, but now we can create fan cams by separating and extracting each member from footage shot with a single 8K camera, enhancing production efficiency.' He added that this technology has been licensed to Japan's KDDI, resulting in cost savings and the creation of a new revenue model through AI adoption.
CJ ENM applied real-time multilingual subtitles using AI while streaming K-pop live videos to 190 countries. He noted, 'Previously, one stenographer was needed for each language, but by switching to AI subtitles, we significantly reduced the workforce burden. Non-Korean viewers have a better understanding of the content with real-time subtitles, improving the global fan experience.'
Park concluded, 'GS Caltex, LG Electronics, KBS, and CJ ENM, despite being in different industries, achieved results by centering AI use around employees who need to use it in practice. The key to successful AI adoption in companies lies not with the IT department but with the operational teams.' He also introduced 'Amazon Quick' as a tool to support AI adoption centered on operational staff.
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.