On the 31st in Seoul's Gangnam District, 27th floor of the GS Tower. Like a small Consumer Electronics Show (CES), about 20 booths were set up. It was the site of GS Caltex's third Deep Transformation Day (DT Day), showcasing best practices in combining artificial intelligence (AI) and digital transformation (DX).
The most eye-catching spot was the equipment management booth. At the GS Caltex Yeosu plant, 400,000 tons of hydrogen are produced annually. The hydrogen produced is sent to other plants and equipment through compressors, machines that compress air. In the past, when a compressor broke down, quick response and repairs were required, but the response depended solely on the person in charge's skill level. That is because experience and knowledge accumulated during past inspection and repair processes were not properly turned into data.
But the situation changed after GS Caltex introduced the Asset Plus process optimization platform. By connecting 800,000 devices, instruments, and piping facilities as data, equipment management can now be handled on a single platform.
For example, when a compressor issue occurred, searching for the compressor in Asset IU pulled up information on related symptoms, causes, and actions. By migrating information that had previously remained only in workers' PC Excel files into Asset IU, about 1,700 items of information became searchable.
Choi Chang-seong, a manager on the GS Caltex equipment planning team, said, "In the past, when a new person came in due to personnel changes, retraining was needed, or it took time for skill levels in inspection and repair to catch up," and added, "With Asset Plus introduced, past knowledge and experience can be searched, enabling anyone to respond quickly."
In particular, AI calculates and manages equipment life spans, inspections, and replacement cycles on its own, and even checks for equipment anomalies by analyzing vibrations. This can block problems in advance through preventive inspections and maximize efficiency. Workers' unique experience and knowledge are turned into data resources, while simple, repetitive tasks are handed off to AI.
The Yeosu plant also introduced AI closed-circuit televisions (CCTVs). In the past, people in the situation room watched CCTV footage directly to detect situations. The problem is that the Yeosu plant spans 6 million square meters (about 1.82 million pyeong), twice the size of Yeouido. There are limits to people monitoring 24 hours a day.
Since the introduction of AI CCTVs, AI automatically detects flames, smoke, and intruders and sends warning messages to managers. So far, about 170 AI cameras have been installed, and with two years of accumulated data, accuracy has reached 99%, according to GS Caltex.
GS Caltex's AI shift is bringing change not only to plants but also to offices. Lee Eun-ju, head of the GS Caltex DX Center, said that while preparing for interviews with experienced hires recently, the in-house AI platform AIU was being used.
Lee said, "In the past, I had to read and organize applicants' documents one by one," and added, "Now, when I ask AIU, it organizes key experience and differentiators and even suggests questions that might be missed, improving the completeness of the interviews."
The booth for "Studio Ballal" also drew attention. Studio Ballal is an AI creative group voluntarily formed by members of the PR2 team in the public relations division. They built an integrated communications management system and, in collaboration with other departments, showcased 20 in-house produced Generative AI video contents over the past six months.
These changes at GS Caltex stem from the determination of President Hur Sae-hong. Hur presented AI as a core pillar of corporate competitiveness and even provided AI education to executives in Jul. to spread AI within the company.
Hur said, "AI is not just a technology but a new colleague that changes how we think and work," and added, "We will evolve into an organization capable of faster, more precise decision-making and creative collaboration through AI."