As AI transformation (AX) spreads, more companies are moving quickly to help business units develop needed AI services or agents directly without going through IT staff. That is likely to lead to a change in the role of IT departments.

Eun Ju Lee (이은주), executive director and head of the DX Center at refiner GS Caltex, highlights training for business staff and measurement and monitoring of AI use as IT’s increasingly important role in the AX era.

Lee met with a reporter at the AWS Summit Seoul 2026 conference hosted recently by Amazon Web Services and shared the company’s AX results. "In the past, IT departments focused on making services. Now the role has grown after services are created, and it will grow further," she said. "IT staff now need to focus on understanding how well services created by business units are used and helping improve them if there are problems," she added.

Before discussing GS Caltex’s AX and the changing role of its IT department, it is necessary to explain the company’s business environment. GS Caltex is structured in a way that requires it to maximize efficiency through technology. GS Caltex has revenue of about 44 trillion won, but its workforce is about 3,000 people and operating profit is less than 1 trillion won.

Because the company has relatively few people for its scale, technology was already a key keyword inside GS Caltex even before the terms digital transformation (DX) and AX emerged. The company has especially focused on using data. "GS Caltex has a lot of data, and if we use it well we can optimize production," Lee said. "Beyond developing new services using AI, we have also steadily worked to digitize even design drawings from about 60 years ago. Data quality is not a perfect 100 points, but we are continuing to accumulate it while finding effective ways to use it," she added.

GS Caltex established a DX Center in 2022 and began its DX strategy in earnest.

GS Caltex’s DX strategy can be summed up as bottom-up first.

Unlike top-down efforts led by management, bottom-up DX focuses on spreading a culture in which staff on the ground think through what they need for their work, build it themselves and share it with colleagues. The same applies to AX. The priority is to build an environment where business staff can create the AI-based apps they need through various methods, including low-code, robotic process automation (RPA), machine learning and generative AI, and use them properly. "The core of a bottom-up strategy is building a structure where it is not something people do because they are told from above, but something created on its own from below. Leadership also has a role here," Lee said.

With the spread of generative AI, field-led app development inside GS Caltex appears to be gaining further momentum. Apps for various uses, from drafting contracts to screening hires and checking purchasing information, are being created directly in the field and spreading internally, the company said. Lee said it built a generative AI platform based on Amazon Bedrock and provides an environment where business staff can develop machine-learning services even without coding through Amazon SageMaker Canvas. "Various cases are already coming out. About 3,000 employees have made around 1,900 apps, and about 60 apps have been used hundreds of times or more. There are 10 apps used jointly across the company, and 9 of them were made directly by employees," she said.

Not only has the volume of services created in the field increased, but meaningful numbers are also emerging in business terms.

In marketing, the company built an environment to analyse customer complaint (VOC) data in natural language using an ontology and generative AI. That led to a sharp rise in users of GS Caltex’s Energy Plus mobile app, which provides mobility services such as refuelling, car washing and electric vehicle charging. "Energy Plus users increased from 150,000 to 2.4 million, and the app rating rose from 1.6 points to 4.6 points," Lee said.

As the company encourages field staff to make and use apps directly, the role of GS Caltex’s IT department is gradually shifting from developing what business units need to building an environment where business units can create and use what they need on their own. The IT function still has many things to focus on in terms of support.

Even if AI has made things easier, there is still a barrier to entry for business units to build and use apps with AI. That makes education more important. Education does not produce results immediately, and Lee emphasises training backed by details tailored to business situations.

"Beyond simply informing business units about technology, we provide technology that can be learned as easily as possible. If the barrier to entry is lowered, it can greatly reduce disruptions in work when 담당자 changes," Lee said. "If people write in natural language through text SQL rather than using data with Python queries, a person newly assigned to the work can adapt more easily," she added.

Training periods are also set with the realities faced by business staff in mind. "As generative AI evolves, we can significantly shorten training periods. For machine learning, there used to be 12-week courses, but now it has become possible for staff to take a one-day course and use it in their work," Lee said. "We have been running training since 2023, and these days the number of people who want to take the training is more than twice what we expected," she added.

Monitoring is also becoming a keyword with growing weight inside GS Caltex’s DX Center. It would be good if business units make apps and continue to use them well on their own, but there are cases where that does not happen for various reasons. Lee said monitoring how and how much apps created in the field are used, and supporting fixes when there are problems, will become the most important role of IT departments going forward.

"At GS Caltex’s DX Center, we look at how many people use apps each month, how many apps are being used, and which apps are used well and then not used, on a weekly and daily basis," Lee said. "For apps that stop being used after being used well, we are expanding the weight of interviewing business staff to ask why and supporting them in resolving the issues," she added.