POSCO E&C has held a company-wide artificial intelligence (AI) contest to identify AI-based models for automating construction site operations and improving productivity.
The company said Tuesday that it had concluded the "Company-Wide AI Challenge," held over roughly two months from March. A total of 1,887 employees participated across four categories — video, report, AI agent and Golden Bell — representing about half of the company's entire workforce.
The "Work Report Automation AI Agent" in the AI agent category was selected as the top entry. At construction sites, partner company employees report their daily work through social media (SNS), and a person in charge manually compiles the reports to produce a daily work log. The winning system automates this process, cutting repetitive tasks that previously took an average of more than 90 minutes per day and laying the groundwork for a data-driven site management system. On an annual basis, this translates into roughly 375 hours (about two months) of work saved per site manager. POSCO E&C plans to push ahead with upgrades to apply the system to actual construction sites.
The finals also introduced a method for evaluating real-time AI capabilities. Participants in the report category drafted and presented reports within a set time limit on topics provided on the spot, such as "AI-Based Innovation in Construction Site Productivity" and "Strategies to Enhance Brand Value."
POSCO E&C plans to offer the top winner in each category an opportunity to attend an overseas AI conference.
"Through forward-looking and creative work innovation using AI, we have laid the foundation for a smart-work culture that reduces repetitive tasks and allows employees to focus on more essential work," a POSCO E&C official said. "We will build an optimized AI platform and continue to advance our AI agents."
