POSCO DX Transitions to AI Native: Equipping Both Physical and Agentic AI
Physical AI Deployed in Industrial Sites: Robots Introduced for High-Intensity and Hazardous Processes
Development of 'Agentee' AI Agent Platform: Advancing Autonomy in Office and Management Operations
POSCO DX has declared its transition into an artificial intelligence (AI) native company, embarking on both technology development and commercialization. The company has set a goal to support AI transformation (AX) in industrial sites, with a dual strategy focused on physical AI and agentic AI.
Chae Hwan Choi, Head of Management Planning at POSCO DX, is making a presentation at a press conference held at Lotte Hotel Seoul on the morning of the 29th. Photo by Myunghwan Lee
View original imageChoi Taehwan, Head of Management Planning at POSCO DX, stated at a press conference held at Lotte Hotel in Jung-gu, Seoul on the morning of the 29th, "POSCO DX aims to evolve into an AI native company," adding, "We are the only company in the country that understands both the physical domain of factories and the management domain of the office."
POSCO DX has outlined a strategy to realize intelligent factories through physical AI and to automate office and management areas using its proprietary AI agents. To establish factories where three entities—humans, AI, and on-site equipment or robots—collaborate, the company is integrating physical AI technologies into industrial sites to achieve 'robotization of equipment,' enabling control over large-scale machinery.
Physical AI has already been applied to steelworks equipment, including large cranes used to transport steel products, reclaimers that move raw materials stacked in yards, and ship loaders handling raw materials at the docks—all weighing dozens of tons. With this approach, POSCO DX is enhancing worker safety and productivity. In some cases, the AI learns from human expertise to collaborate directly with on-site workers.
The company is also promoting the application of industrial robots in high-intensity and high-risk environments. It offers a comprehensive suite of robotics automation services, including consulting, design, construction, and operation, and has jointly secured a cross-functional robot operation platform with POSCO. POSCO DX is advancing robot walking and manipulation technologies by designing robot systems based on simulation and developing vision-language-action (VLA) models for mobile robots, such as humanoids.
At the press conference, POSCO DX showcased several examples of robots performing demanding and hazardous tasks, including a robot for impurity removal in zinc galvanizing furnaces at steelworks, a band cutter robot for the cold rolling process, and an autonomous guided vehicle (AGV) capable of automatically transporting steel products weighing up to 65 tons.
POSCO DX is also advancing the autonomy of office and management work through "AI staff" specialized for each area. To achieve this, the company has independently developed 'Agentee,' an AI agent platform that integrates the life cycle management of employees from recruitment to evaluation, training, promotion, and retirement. All processes follow a standardized methodology, enabling rapid deployment of AI agents.
Hot Picks Today
"No Wonder It Soared Today... Morgan Stanley Eyes 3 Million Won Target—Which Stock Is It?"
- "Who Wants to Hire 40 Skilled Kimbap-Making Women?"... A 'Suspicious Ad' Appears Online
- "Why Waste Scorching Subway Heat? New York Launches 1.1 Billion Won 'Innovative Experiment'"
- "If You're So Sure, Try It Yourself": 44-Year-Old Father Admitted to Prestigious University After Daughter's Challenge
- "This Is Not Right, Chairman"...Flood of Criticism over SK hynix's Pursuit of 'Five-Layer Duplicate Listing'
There were also cases shared where AI agents were utilized in finance, human resources, labor, and business management operations. With the adoption of AI agents, the recurring "pre-closing risk checks"—previously taking three hours—can now be completed in about one minute. POSCO DX expects that, if AI agents are implemented across the entire financial closing process, overall productivity could increase by at least 30% compared to previous methods.
© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.