Noly Universe employees are using AI to cut repetitive work such as converting concert seating charts and consolidating event information. The company has applied 97 AX transition projects to actual work, and its own tally puts monthly time savings at about 3,600 hours.

Joo-bong Yoo (유주봉), leader of Noly Universe's AX Excellence team, presented these results and cases at DigitalToday’s 'Digital Insight 2026' conference held at COEX in Samsung-dong, Seoul, on Sept. 29. He spoke on the theme, "Change that began out of sight - Noly Universe’s internal AX transition."

Noly Universe decided to start by changing internal work because a lot of manual work remains at operational sites in its travel, leisure and culture businesses. Document formats sent by partners such as affiliate stores and concert promoters vary, so employees must cross-check and re-enter data. Some tasks were small in scope and were pushed down the development priority list.

Yoo said, "If one person in charge is out, it is extremely difficult to replace the work." He said the company pursued internal AX from the idea of "let’s change how we work first before going B2C."

In November last year, Yanolja Group held an internal AI contest. A total of 342 employees from group affiliates participated in 74 teams, and teams that included both developers and non-developers received extra points. It was meant to encourage employees who know on-the-ground work to solve problems together with employees who can implement solutions.

One tool that emerged from frontline teams was a concert seating-chart converter. Previously, employees transferred seating charts sent by concert promoters into an internal registration format. Now the tool applies common rules by venue to create a registration file, and the person in charge checks exception sections.

For large venues, the previous manual work took about 6 to 8 hours. In a demonstration, converting a file took about 5 minutes, but the person in charge must review and correct items such as section labels that differ by concert because accuracy is about 90 percent. The tool is currently undergoing validation in operations teams.

A nationwide event calendar was created directly by a planner using AI. It is a tool that compiles local festivals and concert schedules to identify events that could affect lodging demand. Yoo said work that would have taken about 130 hours to plan and develop the conventional way was completed in about 3 hours. It was a case in which a planner waiting for developers implemented the needed screens directly.

An operations manager also built a tool to assign ticket buyers to pickup counters at concert venues. For a concert of about 5,100 people, it previously took 4 to 5 hours to split lists by name, but the tool can instantly calculate assignment results while changing conditions such as the number of counters. Another tool that organizes concert guidance into customer service questions and answers cut the work from about 180 minutes per concert to about 5 minutes, based on the responsible organization’s tally.

Noly Universe said 30 teams and departments have so far registered 151 AX transition projects, of which 97 have been applied to work. The estimated monthly savings of about 3,600 hours is an internal tally based on project owners comparing time spent before and after adoption.

In a question-and-answer session, Yoo introduced a case in which the security team detected a non-developer-built tool after an API key or personal information was mistakenly entered. He said, "The hardest part was budget." He said the company had to prove productivity gains to explain the need for additional budget after increased AI use exceeded the initial budget.

As the number of tools grew, problems also emerged, including other teams redeveloping functions already built elsewhere and having no one to handle maintenance after tools were created. Noly Universe is supporting reuse by collecting each team’s tools and use cases through the AX Excellence team and an internal portal, 'AX Hub.' For projects that frontline teams can build themselves, it provides training and templates, and it collaborates with development teams on projects that require system changes.

Yoo said the company also reviewed automated generation of marketing copy, but postponed adoption because the time required for input and review was similar to writing directly. He said whether AI application actually reduces work time was used as the criterion for selecting projects.