"The questions corporations ask about artificial intelligence (AI) have changed. In the past they asked, 'Which model is better?' Now they ask, 'In which business areas, what can we actually execute?'"
Gong Seong-bae, chief AI officer (CAIO) and vice president of the AIR institutional sector at Megazone Cloud, said this on the afternoon of the 26th at SMARTCLOUD SHOW 2026, the country's largest tech conference, held at the Westin Josun Hotel in Jung District, Seoul. He said the era of chatbots competing on answer quality is fading, and as agents move into actually performing tasks, the focus of competition has shifted to execution. He gave a lecture that day on "The AI direction of corporations seen through the latest trends."
The backdrop to Vice President Gong's emphasis on execution is concern over the "AI-era productivity paradox." The idea is that while AI speeds up individuals' task handling, its effects do not appear in financial performance or labor productivity. Vice President Gong said, "Everyone uses AI, but some use it well and others fall behind, and the results are not immediately visible," adding, "It's exactly like when Robert Solow in the 1980s said, 'You can see the computer age everywhere but in the productivity statistics.'"
Vice President Gong predicted that productivity will rise only when AI technology, people, and processes are combined. He said, "When PCs first appeared, we thought innovation would happen immediately, but the real effects showed up 10 to 15 years later when work systems such as customer relationship management (CRM) and supply chain management (SCM) were computerized. Likewise, adopting AI does not automatically raise productivity."
Vice President Gong said, "Efficiency is saving time, and productivity is creating new business value," adding, "Finishing a report that used to take 10 hours in 10 minutes is merely efficiency; you have to convert the saved time into revenue, customer satisfaction, and quality for it to truly be productivity improvement," stressing the need to distinguish between efficiency and productivity.
Vice President Gong advised that corporations should make the harness the core when introducing AI. A harness refers to the execution environment for efficiently and reliably controlling and managing AI models. He emphasized, "A harness is a device that fixes results so that similar outcomes emerge even if you change the model," adding, "The constraints managed while operating the harness remain as the company's asset."
He also addressed the trend of treating AI as a work colleague and the need for corresponding controls. Vice President Gong said, "In the past we treated AI like a slave, but now we should accept it as a colleague," while noting, "If you give a colleague excessive authority, accidents can happen, such as deleting company documents, so appropriate permissions and controls are essential."
Vice President Gong predicted that the indicators used to measure AI performance will change. He said, "Right now we look at ROI (return on investment) such as processing time, error rate, and rework rate, but later the focus will shift to ROE (return on employee) from the perspective of employee productivity, and ROF (return on future) from the perspective of future value." On ontology, a recent hot topic, he urged caution. He said, "If you recklessly apply it companywide, it will fail 100%," adding, "You should start with small areas."