Through SK's Icheon Forum 'Full-Speed AX' and the special lecture by Song In-hyuk, CEO of Unique Good Company, invited by SK's SUPEX Pursuit Council
Examining changes in how we work and leaders' organizational restructuring
In the age of one person, one agent, how much faster has the organization become?
If more employees use AI well, will the organization also work better?
If all members leverage AI and each person has an agent, will the company move that much faster? As AI adoption spreads rapidly, companies are now asking the next level of questions. Because they know that in reality, this is not the case. They are coming to understand that an individual quickly producing outputs like documents, code, and design is fundamentally different from multiple departments pursuing a single goal, and further, creating synergy that translates to company productivity. Beyond simply using AI tools to improve productivity, they need to examine organizational review and approval processes, roles and responsibilities, and the way information is communicated. From this perspective, following the flow of SK Group's full-speed AX discussions this year and the nanoventure insight special lecture by Song In-hyuk, CEO of Unique Good Company, held in September through the SUPEX Pursuit Council invitation, we can find clues to the answer to this question.
In June, Chairman Choi Tae-won's Icheon Forum: The core of SK's declared 'Full-Speed AX' is O/I
SK held the '2026 New Icheon Forum' at the SKMS Research Institute in Icheon, Gyeonggi Province from June 11-13, 2026. It was an integrated event combining management strategy meetings and the Icheon Forum, intensively discussing AI-centric management transformation. At this event, Chairman Choi Tae-won emphasized comprehensive and rapid AI transformation, namely 'Full-Speed AX.' He also presented the direction that AI must evolve from 'My AI' to 'Our AI.' This means developing individual AI usage into AI that connects the entire organization's work and performance. Additionally, he explained the essence of AX as operational improvement, O/I (Operation Improvement). The explanation is that AX is about precisely defining what our organization does, changing how we work, and building the capacity to execute it. Management executives also agreed that business processes must be fundamentally redesigned.
AI Made Outputs Faster, But What About Our Organization's Speed?
CEO Song In-hyuk introduced the organizational management difficulties he had experienced in a special lecture to SK's president-level executives. Realworld, which was Unique Good Company's flagship business model, as a game content and platform required comprehensive work areas such as planning, design, and marketing, and in particular, the advancement of the development department was also very important, with characteristics where efficiency between each work process was critical. However, by 2025, Unique Good's members were already using AI extensively, but the work of reviewing and revising results and the existing work procedures remained unchanged, so organizational bottlenecks were not resolved at all and naturally did not lead to productivity improvements—an experience he frankly shared.
For example, imagine a team using AI to create planning drafts much faster than before.
As drafts increase, the documents to be reviewed also increase. If the criteria for what to pass through are unclear, revision requests are repeated. If coordination with other departments and final approval proceed in the existing manner, even if the time to create drafts decreases, the time to complete the project may not change significantly. Rather, as output production speed increases, bottlenecks between processes are intensifying, creating more problems. Meetings increased and speed slowed down.
In the end, the issue isn't the adoption of productivity tools themselves, but rather examining together where work becomes bottlenecked, who decides by what criteria, and why the same revisions are repeated. The key insight was that distinguishing between individual work speed and organizational task completion speed could be a starting point.
The Key to Organizational Productivity is AO (Agentic Organization)
Therefore, the concept presented by CEO Song In-hyuk is AO (Agentic Organization), an agentic organization. It's an operational structure where organizational conversations, directives, and feedback are connected to agents' task performance, and this process builds up as organizational standards and knowledge. It can be said to be one way to actually operationalize 'Our AI' as mentioned in the Icheon Forum. AO goes beyond how members each use AI individually, creating an operational structure that connects organizational conversations, directives, and feedback to agents' task performance. It focuses on how the entire organization's work is connected and improved rather than individual tool usage. For companies to achieve O/I, they had to redesign the organization itself so that synergy with agents would occur.
All members' directives and feedback must be made readable to AI.
It's AO, not AX. Why AO? The reason organizations don't actually change even after adopting AI is because AI usage remains in a black box area. No one else can know what work each person did or how they build skills through what feedback. Since individual AI accounts were distributed, it's arguably natural. They simply share outputs on collaboration tools or repositories and each person looks at them and converses—that's all. For agents, process and feedback are important, but these are completely left out. Of course, as people work, there are also incorrect results, and these failure records are very important too. Because they can be organized as reference points for the next task. Most importantly! Even if the person in charge changes, previous decisions and experiences must be able to be reused.
Rieder must ensure this.
Change Never Happens Through Delegation Alone
What's important is not expecting AI to understand and improve the organization on its own. Leaders must directly design what information to keep, what to validate, and how to reflect it in the next execution. CEO Song In-hyuk also initially, because he had a development organization of over 10 people, ordered focus for realizing AX, but concluded that delegation completely failed. Because organizational design wasn't something a few members could do. Each member was optimized for their given tasks but lacked the perspective and authority to 'break' the organization's work processes. Furthermore, teams of various sizes, like cell membranes each with their own worldviews as mini-companies, experience a sense of crisis as if facing survival threats when dismantling their own domains, so without leadership or a CEO with clear decision-making authority, change doesn't even get a foothold. He also shared this experience. In the end, after paying the expensive tuition, CEO Song emphasized that he directly designed the organization's AO and, until it was fully advanced, never ceased his role as chief developer.
If you would like to see the detailed lecture content by CEO Song In-hyuk, please click the link below!
AI Education Must Change to Realize Collective Impact
If we translate this transformation into education, we can divide questions according to the decision-making scope for each level. Executives and management address which tasks to change first and authority structures, team leaders and key talent address how to share and build the team's outputs, and new employees address the responsibility of connecting and completing assigned work through to the end.
In an agentic organization, regardless of rank, each individual member becomes an executor. The boundaries between planners and executors, between those who direct and those who process, fade, and anyone can directly implement ideas with an agent. This is also why a small team of a few people can accomplish as much as an organization of over 100 people.
At this point, an important question is—if it's used to process existing work slightly faster, the scope of change is smaller than expected. The power of an agentic organization truly emerges when challenging work that has been delayed due to lack of personnel and time, persistent problems between departments that have long remained unsolved, and company challenges that no one dared to tackle.
Therefore, the goals of AI education must also change. Tool usage is actually not very meaningful. Beyond that, all members must transform into executors, together define the challenges our company faces, and have experiences solving them directly with agents. Education itself becomes a venue for tackling the company's real problems.
The leader's role also becomes clear here. It is about bringing out the problems the organization really must solve, establishing authority and decision criteria so members can challenge them, and ensuring that experiences gained in the process build up in the organization. The three decisions by leaders we examined earlier are ultimately conditions for opening such a venue for challenge.
From Special Lectures to On-Site Experiences: Realworld's Corporate Education
There's also a way to connect this discussion to experiences shared with colleagues. Realworld operates programs that combine special lectures, content experiences, and classroom usage. Realworld's corporate education, along with AI·AX special lectures, covers participatory workshops where attendees divide roles and solve missions. At this point, it's important not to end the meaning of the experience simply at "experienced new technology." You can reflect on who had what information, how decisions were made when opinions differed, and how roles were divided for a shared goal.
In the lecture, you gain questions for viewing the organization; in the experience, you create concrete experiences to discuss those questions with colleagues.
For those planning AI special lectures and corporate workshops, it's worth first determining what work and behaviors members will revisit after education, rather than the program's name.
Thank you!
Through nanoventure, design and prepare in advance for an era when one person can command 10,000 AI organization members.
https://www.nanoventure.org/