Moving beyond digital transformation to artificial intelligence transformation... All about 'AX,' which redesigns how we work around AI
"We adopted AI, but why is our way of working still the same?" This is a question many companies have recently encountered. Generative AI tools such as ChatGPT and Claude have been brought into the workplace, but they often fail to lead to noticeable results. Emerging to fill this gap is the concept of "AX (AI Transformation)." AX goes beyond simply adding AI to a few tasks; it refers to a transformation where an organization completely redesigns its actual way of working around AI. It is not about changing tools, but about changing the structure of work.
Changing the 'Structure of Work,' Not Just Tools
To understand AX, one must first examine its difference from the preceding "DX (Digital Transformation)." DX was the task of moving analog to digital. The core was changing paper documents into electronic documents, offline stores into online shopping malls, and manual ledger books into systems. In contrast, AX is the stage where AI independently learns, predicts, and suggests optimal judgments on top of such digitized data and processes. Microsoft CEO Satya Nadella defined it in November 2024 by stating, "If DX is automation, AX is autonomy." This means the center of gravity shifts from a stage where humans operated systems to increase efficiency to a stage where AI analyzes and suggests independently, collaborating with humans.
Taking a restaurant as an example makes the difference clear. If changing the method where customers ordered directly from a waiter to a kiosk or app is DX, then AX is analyzing customers' past order history along with weather and time zones to recommend menu items beforehand. Global IT firm IBM defined AX in a 2024 report as a strategic initiative integrating AI across operations, products, and services to foster innovation, efficiency, and growth. Boston Consulting Group (BCG) similarly explained it the same year as a strategy that converts various tasks end-to-end on the foundation of AI infrastructure and member capabilities. The important point is that DX and AX are not opposing concepts, but exist on a continuum. AI can finally play its proper role only when digital data and infrastructure are sufficiently established.
The Speed of Transformation Shown in Numbers
Behind the emergence of AX as an urgent task are concrete figures. According to a 2024 survey by global consulting firm McKinsey, approximately 65% of companies worldwide already utilize AI in daily operations. This is nearly double from 33% in 2023 within the span of a year. The 2025 AI Index Report published by Stanford University's Human-Centered AI Institute (HAI) also responded that as of 2024, 78% of organizations globally applied AI to at least one task. Adoption itself has become closer to a default value rather than a competitive advantage.
Performance is also being confirmed by figures. In a domestic government survey (2024), 77.8% of companies that adopted AI answered that work efficiency improved, and the Organisation for Economic Co-operation and Development (OECD) reported that productivity of companies introducing AI automation increased by an average of 15% to 40%. Data compiled by Deloitte Korea in 2026 also showed that 66% of responding companies selected efficiency and productivity improvement as the greatest achievement of AI adoption. The future outlook is even stronger. The World Economic Forum (WEF) estimated that about 86% of global business transformations will be driven by AI by 2030, and the Korea International Trade Association projected that around 36% of domestic manufacturing will adopt AI technology by the same year.
Both Government and Corporations Declare 'The First Year of AX'
Changes are appearing simultaneously in policy and corporate fields. In October 2025, the ministers of three ministries—the Ministry of Science and ICT, the Ministry of Trade, Industry and Energy, and the Ministry of SMEs and Startups—gathered and agreed to cooperate for the spread of AX in the domestic industrial sector. The government has begun separately classifying companies leading work innovation through AI-based software as "AX Companies." The speed in corporate fields is even faster. Samsung conducted an "AX Bootcamp" targeting presidents of all affiliates to have management directly handle AI from the top, setting a plan to complete training for all employees within 2026. Chairman Lee Jae-yong emphasized in his New Year's address that "the way we work and our corporate DNA must be completely changed," instructing the incorporation of AI into all tasks ranging from R&D to production, marketing, and support. Daesang Group declared 2026 as "the first year of AI-based work transformation," entering the establishment of "AI agents" where AI performs specific tasks in their entirety.
Beyond Efficiency to 'Redesigning Work'
So, how will actual work change? The core lies in "redesigning work" rather than just "efficiency." If AI is used simply as a tool to process the same work faster, it ends up merely spitting out two to three times more existing outputs. True AX starts from redefining what humans should focus on. Repetitive and labor-intensive tasks such as report writing, invoice processing, or data organization are handled by AI agents, while humans shift to areas requiring judgment and creativity.
Indeed, companies pushing for AI transformation start by dividing work finely into small units that AI can perform, standardizing each unit, and automating them. An industry executive described AX not as a means to process work faster, but as a process of finding the essence of work. In the end, diagnoses show that success or failure is determined not by what to have AI assist with, but by what to entrust to it and what new roles humans will take on.
As the flow of transformation steepens, human roles change alongside it. Analysis suggests that the advent of generative AI has gone beyond mere productivity improvement into entering a stage of redesigning the organization's job design and workforce structure itself. As repetitive tasks shift to AI, humans focus more on verifying the results produced by AI, imparting context, and making final judgments. For this reason, experts emphasize that progress should not be evaluated solely by time and cost savings. Qualitative criteria, such as whether results were achieved at a level to change company policy or constitution, and whether entirely new values impossible before were created, must be examined together. This is why AX must be viewed as a matter of working "how differently" rather than "how fast."
Four Conditions That Complete AX
AX is not completed simply by introducing AI tools. Experts believe that four elements—data, infrastructure, people, and culture—must interlock in a balanced way. The starting point is data. Without accurate and consistent data, AI cannot properly learn nor create correct grounds for judgment. Second is infrastructure, requiring a system environment where AI can operate stably. Third is people. No matter how much AI develops, the ultimate subject of final judgment is still humans, and members must possess "AI literacy" to interpret AI results and connect them to business value. Last is culture. An open atmosphere of judging based on data, experimenting with new tools, and repeating improvements must take root so AI can naturally permeate into the way of working.
The Gap Between Expectations and Reality, and Security
However, rosy forecasts do not tell the whole story. There are still many companies that introduce AI tools yet fail to connect them to actual changes in working methods or organizational-level performance. In a survey targeting practitioners, a difficulty commonly cited by respondents was the blankness of "not knowing how to apply it to my work." This means change does not happen simply by being introduced to tools. Pointers also indicate that the biggest wall blocking AX lies in the work structure rather than the technology itself. Unless knowledge and experience staying only inside human minds are drawn out into organizational systems and work flows are structured into a form AI can handle, automation cannot expand. Security is also a new challenge. As data and systems are closely connected through AI as a medium, the complexity and impact of security threats also grow together, making the design of safety devices alongside the speed of transformation critical.
A Greater Opportunity for Small and Medium-Sized Enterprises
An interesting point is that AX is not the exclusive domain of large corporations. Views even exist that it is a more realistic opportunity for small and medium-sized enterprises or small practitioner organizations. While cloud computing and the Internet of Things (IoT)—core technologies of the DX era—demanded massive costs and specialized personnel for adoption itself, generative AI allows powerful functions to be rented at relatively low costs through Large Language Models (LLMs) such as ChatGPT, Gemini, and Claude. Without massive organizational restructuring or large-scale investments, it can be started immediately in small work units like contract analysis or report automation. Introducing small tools to reduce repetitive work and improve customer experiences is also strictly a starting point for AX. Ultimately, the essence of AX is closer to a perspective shift looking anew at how we work, rather than massive technology.
Where to Begin
There is no fixed correct answer on how to start AX. However, a realistic approach commonly recommended by experts is to start small from areas with measurable value rather than a large project changing everything at once, verify results, and expand gradually. This method is also advantageous for reducing risk and forming consensus within the organization. Even if slightly behind in digital transformation, catching new opportunities quickly can surpass leading companies, while conversely missing the current flow will widen the gap further, according to the diagnosis of many experts. AX is not a temporary buzzword, but an ongoing transformation rewriting the way of working itself. The important question is shifting from "Will we adopt AI?" to "How will we redesign our organization's work?"