South Korea’s industrial competitiveness has always been rooted in its manufacturing sector. In industries such as shipbuilding, batteries, and home appliances, the competitive edge is largely determined by how well factories operate rather than what they produce. Even with similar equipment and workforce, differences in productivity and quality often come down to operational efficiency.
Many manufacturing companies are asking: “How much will production efficiency improve if we introduce artificial intelligence (AI)?” While this question is valid, it doesn’t fully capture the changes happening in modern factories. More pressing concerns include whether to build infrastructure first or start with small AI tasks when data is limited. Should implementation wait until human-level accuracy is achieved? Although this cautious approach seems logical, it often leads to delays. As infrastructure is being built, on-site issues remain unaddressed, and individual tasks fail to translate into organizational performance. Waiting for perfection only slows progress, leaving companies in a state where they prepare for AI without seeing any real change in competitiveness.
A good example of how to address these challenges is Foxconn, a global electronics manufacturing company known for its collaboration with Apple. Working with BCG, Foxconn took a different approach by fundamentally redesigning factory operations. Instead of viewing AI as a tool for improving specific processes, it started by asking: “How should factories be operated?”
Foxconn’s Chosen Approach: ① Redesigning the Factory Operating Model
Through the “Genesis” project, Foxconn integrated physical processes, digital twins, and AI-based decision-making into a single system. Factories were transformed from human-centered operations to structures where AI continuously optimizes performance. Initially, a “Human-in-the-loop” approach was used, where humans actively intervened in AI systems to supplement judgments, enabling rapid on-site application and parallel learning.
The biggest change appeared in the design process. Previously, factories were built first, and issues were corrected through trial runs. However, Foxconn designed processes first in a digital twin environment, derived optimal conditions through AI simulations, and then constructed the actual factory. Changes in the operational phase are also clear. AI agents handle repetitive tasks such as defect correction, process coordination, and performance optimization, while humans focus on exceptional situations and complex judgments. This structure expanded into a cross-factory learning system. The central model is deployed to global hubs, each factory relearns based on data, and results are gathered back to the center. Consequently, experiences and know-how between factories are shared in real time, and operational methods are continuously improved in response to changes in raw materials, equipment, products, and other environments.
Factories are no longer fixed systems but are evolving structures that learn and adapt on their own. This is not a simple shift to automation; it is a fundamental change in how factories are operated.
Foxconn’s Chosen Approach: ② Direction and Speed, Not Perfection
What is more important in Foxconn’s case is its decision-making approach. The key to success lies not in which AI is applied but in defining how factories should be operated. Before applying AI, Foxconn first established the direction of its AI operating model. This direction was defined not by the field but by CEO Young Liu. Specific goals such as yield, quality, productivity, and cost structure were linked to the direction of AI utilization, aligning AI as a means to achieve these objectives.
Another difference is speed. Foxconn did not wait for perfect preparation. Even in situations with insufficient data, it accumulated data and learned in parallel using AI. Initially, human intervention supplemented results while rapidly applying AI. This is a prerequisite for fast AI adoption. Similar patterns appear in overseas battery material companies. Although there is resistance due to initial data shortages and infrastructure issues, when top management clarifies the direction and pushes for execution, AI-based decision-making is rapidly adopted. As a result, quality and yield improve within a short period, and positive changes occur in productivity and cost structures. Ultimately, the speed of AI transformation is determined not by technology but by management’s resolve.
Ultimately, Competitiveness Emerges from ‘Redesigned Factories’
The environment facing South Korean manufacturing is becoming increasingly complex. Amid competition with China, companies must secure both cost competitiveness and quality superiority, but skilled labor is decreasing, global production bases are expanding, and demand volatility is growing. Particularly, due to global expansion into regions like the U.S., Europe, and Southeast Asia, organizational cultures and operational methods differ across regions, making past Korean production methods difficult to apply as-is. Added to this, demand volatility and supply risks are significantly expanding the uncertainty of production operations. In this environment, operational methods relying on experience and intuition are insufficient to maintain competitiveness. The issue is not productivity but the very way factories are operated.
Now, manufacturing companies are at a point where delaying choices is difficult. The choice is whether to remain at the level of adding AI to existing processes or to fundamentally redesign factory operations around AI. AX (AI transformation) in manufacturing is not just digital conversion. It is a change in operational methods where equipment, processes, workforce, and data are connected into a single system. While this transition is not easy, once started, the competitive gap widens rapidly. The problem companies are currently grappling with is not a lack of technology—sufficient technology already exists. What matters is how quickly they set a direction and move to execution. Observing leading cases or starting with infrastructure construction makes it difficult to keep up with the pace of change. Now, companies must shift to a structure where they execute on-site, learn through execution, and reflect the results back into operations.
South Korea’s manufacturing competitiveness still comes from factories. The only difference is that these factories are no longer operated in the same way as before. Today’s factories are being redesigned around AI, and the speed of this change will determine future competitive gaps.
월 5900원 멤버십, 신문 독자에게는 2900원, 조선멤버십
55000원 상당의 신문-잡지 8종 마음껏 보기, 조선멤버십
현금처럼 쓸 7000포인트 받아 알뜰한 쇼핑, 조선멤버십
