Hello, this is Unique Good Company.
These days, the interest you're showing in "AX for Leaders" and "AI for Nano Ventures" is truly hot. Every day, many people contact us personally and through their companies, and I feel in my bones that AX is now an unstoppable trend. I sincerely thank you.
Today, I'd like to talk about SK Group Chairman Choi Tae-won's presentation at the Icheon Forum.
It was a declaration: "We must embark on AX (AI Transformation) in all directions, at full speed."
I felt that the Chairman's declaration was setting quite an important milestone for the future of the company. Looking at recent trends, that's even more true. From Samsung Chairman Lee Jae-yong's AI Transformation Declaration, Hyundai Motor Chairman Chung Euisun's AI Internalization Declaration, to LG President Ryu Jae-chul's AI Identity Declaration. Now, the highest decision-makers of companies are directly at the front lines talking about AX. This is because the gap between a company that uses AI as a tool and a company driven by AI is that significant. As we're just entering the stage of transitioning to an agent-based organization, I understand it well. These leaders' declarations are not meant to be consumed as mere slogans. It's a matter of survival. So today, I'll unpack the story from the Icheon Forum from our own multi-directional (?) perspective.
Individual task productivity? ≠ Company productivity
Individual speed must translate into organizational execution speed
AX is often measured by "AI usage rate." We count what percentage of employees use AI, how many times they input prompts a day. When the numbers go up, we come to believe that some transformation is happening. But this number makes us mistake much more than we think.
Individual work speed has clearly accelerated. But meetings still happen on the same cycle, and approvals still go through the same stages. Who decides what, how information flows—it's all the same. Even with 90% of employees using AI, the company's overall speed can be exactly the same as last year.
Writing reports faster, organizing emails faster, writing code faster—these are individual-level efficiencies. This efficiency doesn't automatically transfer into organizational decision-making speed or execution quality. If a draft produced in 10 minutes goes through five approval stages and gets approved only 2 weeks later, what good is it?
So instead of asking "how much AI is our company using," we should ask "is individual speed translating into organizational execution?" The moment individual productivity connects with organizational execution capacity, that's when real AX begins.
Simply distributing accounts doesn't cause transformation.
AX is moving AI from an individual's chat window to the company's operating system.
When you decide to do AX, Stage 1 is usually similar across organizations. You distribute AI accounts to all employees, hold in-house seminars explaining what AI is, and train people on prompt writing. Once you've finished here, some experience a moment of magic, and there's a sense of accomplishment as if you've crossed the threshold of great change called AI.
But this process isn't much different from deploying new software and handing out user manuals. Understanding how to use a tool and that tool changing the company are completely different dimensions of issues.
Real AX is connecting a company's goals, data, work processes, decision-making, execution, and verification into a single flow. If AI remains as a conversational partner confined to one browser tab in an individual's hands, that flow won't be created. From the moment you set goals to the moment you verify results, AI must be involved across all segments of work. I'd appreciate it if you understood this as moving AI from an individual's chat window to the company's operating system. Account distribution is just the first page, and the real story unfolds from there.
From My AI to Our AI
You need to create connection points between agents.
Early in adoption, "each person's AI" naturally emerges. Going beyond simple AI utilization, you now move into the stage of using each person's own agents. Marketers use marketing agents, developers use coding agents. Each person's day certainly becomes easier.
However, these agents remain scattered like islands, unconnected to each other. You're still at the level of enhancing individual productivity.
To move to the next stage, you need to broaden your perspective one step. From one person using one agent well to creating a structure that connects and directs multiple agents. This is multi-agent orchestration. Imagine a flow where a data analysis agent validates campaign ideas created by a marketing agent, and the result is automatically passed to an execution agent. The ability to design and direct this flow becomes a new organizational capability. If "My AI" was a personal assistant, "Our AI" is closer to a neural network connecting the entire organization.
In traditional organizations, work is divided and fixed by department. Planning teams, marketing teams, and development teams each guard their own domains, and collaboration happens through meetings and emails. There's nothing to fault in this structure itself. Even though it's a bit slower, it's the most stable and optimized for maintaining processes regardless of people coming and going. Past O/I (Operational Improvement) was ultimately about how well you managed organizational processes.
But now that AI can work across departmental boundaries, organizations need to adapt accordingly. Beyond the level of one person having one AI assistant, you now need a structure that directs multiple specialized agents and connects them. You're moving from fixed, department-centered organizations to "execution networks" that freely combine based on goals. It varies by organization, but execution networks generally follow this sequence.
Personal assistance → Task execution → Specialized role differentiation → Inter-agent collaboration
When a project starts, the necessary people and agents gather around the goal, and when the project ends, they disperse again. Departments shift from being name tags marking affiliation to nodes in a network where execution flows. We know this isn't as easy as it sounds. But it's also clear that many organizations are already finding their way through trial and error, and actually delivering results.
We understand it's not as easy as it sounds. But it's also true that many organizations and companies are already finding their way through trial and error and delivering results.
The Essence of AX Lies in Task Decomposition and Redefinition
It Starts with Viewing Your Work Metacognitively and Defining It Precisely
When AX is discussed, most people first think of new tools, new models, new platforms. But you can't go far just by constantly changing tools. AX is the work of redesigning your entire way of working, and its first shovel strikes at a place one step ahead of AI adoption. That's "task decomposition."
Ask yourself four things.
Definition: What exactly are we doing?
Decomposition: Can we divide this work into input, judgment, execution, and verification?
Data: What do we need to remember and continue to accumulate?
Responsibility: How much does AI handle, and who makes the final judgment?
Good prompts? They're not that important here. What we need is a good task structure that can accommodate AI.
How work flows in what sequence, who decides what, how results are verified. Reconstructing the backbone of this operation is what AX really is. AI is there as a means to make that redesign possible.
Ultimately, the moment you define your work precisely and metacognitively, that task becomes something that can finally be automated and upgraded. "What is this task made of?", "What judgments are needed?", "Where ends the repetitive and where begins the unique judgment?". Only by decomposing it this way do the parts to hand over to AI and the parts people must hold onto become clear.
A 100-Year-Old Growth Formula Is Breaking
For the past 100 years, the formula for corporate growth was simple. When work increased, you hired people. Growth meant hiring.
This formula is shaking. The new formula reads like this:
Growth = System × Agent
Instead of increasing the number of people when work increases, you refine the work system carefully and grow the number and capabilities of agents operating within it. Hiring is still necessary. But hiring, which always held a fixed constant position in the growth equation, now shifts to become one of multiple variables.
People's roles actually grow bigger.
At the end of such talk comes the worry: "Isn't AI eliminating people's jobs?" Based on what we've experienced, it's nearly the opposite. In fact, the problem is that people aren't adapting to change.
Hand over repetitive performing, writing, and delivering work to agents now. Instead, people move into positions where they set goals, make judgments, build relationships, and take responsibility. These four areas are where AI cannot substitute for humans. Goals require understanding and commitment to organizational direction, judgment is decision-making amid uncertainty, relationships are built on trust and communication, and responsibility requires the existence of a person to take on the results.
The Company Must Become One That Accomplishes Bigger Tasks With a Smaller Structure
Until now, the bottleneck in company growth has always been "execution." How much work has been postponed because of insufficient people and time despite having good strategies? Now, a significant portion of execution can be handled by agents. So there's no reason people should continue standing at that bottleneck. It's time to shift the time spent as hands and feet in execution toward goal-setting, judgment, and responsibility.
Organizations Must Transform to Be Driven by AI
This isn't just SK's story. Samsung is carrying out "AI Transformation," changing how it works and its organizational DNA, Hyundai Motor Group has declared "AI Internalization," and LG is restructuring its business with AI identity at the center. Regardless of scale or industry, the future where the structure of enterprises themselves will be different is already at the door.
Now organizations must be driven by AI.
Real transformation begins when you can answer this question yourself.
An organization that accomplishes much bigger work with a small number of people. It's the face of the company we'll face in the future!
Thus, I've analyzed SK Chairman Choi Tae-won's Full Speed AX today.
I hope Chairman Choi Tae-won's insights reach those pondering organizational AX properly!
The Era Where One Person Moves Ten Thousand
Introducing the New Book Nano Venture
There's a book that tells this entire story from beginning to end. 『Nano Venture』(HumanCube) captures the process that Unique Good Company experienced while restructuring its organization around agents. This year, our organization shrank from 58 people to 7 people. But we were completely reborn through AX and are delivering performance beyond last year. This book vividly records how Unique Good Company, despite hardship, transformed crisis into opportunity through AX. When you see how, with a small workforce, they redefined existing tasks and processes, established a structure for execution and verification, and how they transitioned to agents, you'll see that a new environment has unfolded where not only does one person create amazing productivity, but even small organizations can accomplish work at the scale of over 100 people!
You can check the blueprint in this book!
Easy ordering via KakaoTalk
Order via Yes24
Order via Kyobo Bookstore