'Agent Economy' to Reshape Corporate Workforce as AI Fills Vacancies Left by Resignations

When three employees resign, a company may choose to install AI agents on the computers they previously used instead of hiring new staff. This scenario is already becoming a reality at some game development companies in China. In an era where AI fills the vacancies left by departing employees and remaining staff use AI to perform the work of multiple people, companies expect to reduce the costs of recruiting and training new personnel, though this shift hides a new crisis in organizational management.

According to an analysis by tech writer Kim Ji-hyun reported by T Times TV, corporate management in the AI era will likely succeed or fail based on how "tokens" are managed. The AI industry functions like a manufacturing sector that uses GPUs as raw materials to produce tokens. If data centers are the factories producing tokens, companies become the subjects of "Tokenomics," consuming these tokens to create added value.

"If you cannot measure it, you cannot manage it"... The Importance of Token ROI

Kim identified "Token ROI (Return on Investment)" as the most critical metric for companies establishing business plans for 2027. Rather than simply measuring how much AI is used, companies must measure how much value the invested tokens produce. For example, spending 10,000 won worth of tokens to contribute to securing a 100 million won contract is a successful investment, whereas pouring tokens into creating hundreds of reports that no one reads is a clear waste.

It has been noted that many companies are indiscriminately consuming token usage during the early stages of AI adoption. "As Peter Drucker said, if you cannot measure it, you cannot manage it," Kim emphasized, adding, "Beyond simply knowing the total usage fee, companies must manage usage by attaching at least a minimal label to which department used tokens for what purpose and what results were achieved." This serves as a guideline to prevent indiscriminate consumption while aiding efficient task execution.

On-device AI and 'Computer Use' to Change Device Purchasing Standards

The era of "Hybrid AI," where AI technology moves from infrastructure (data centers) down to devices, is approaching. While previous AI centered on massive cloud-based models, "on-device AI"—where AI runs directly on smartphones or laptops—is expected to become mainstream starting in 2026. Consequently, purchasing strategies for individuals and companies must also change.

Kim suggested the necessity of a "Personal AI Agent (PAI)" as a criterion for device selection. The core feature is "Computer Use," which learns a user's data and work style to operate apps on their behalf without the need to upload files. "Depending on whether you are satisfied with simply using cloud AI or whether you want to utilize your device's resources to perform personalized assistant functions, you will decide whether to buy a standard laptop in the 2 million won range or a high-performance AI PC in the 7 to 8 million won range," Kim explained.

The 'Agent Economy' where screens disappear, causing a shift in the advertising market

The rise of the "Agent Economy," where AI searches, watches YouTube, and executes apps on behalf of users, is shaking the foundations of existing business models. For the past 30 years, humanity has been accustomed to a "GUI (Graphical User Interface)" environment involving looking at screens and clicking mice or touching with fingers. However, when the era arrives where agents act directly upon user commands, users will no longer need to stare at screens.

In this process, existing advertising models may face a crisis. If users receive only the results without looking at the screen, portal banner ads or sponsored search links could disappear. Kim predicted, "Ultimately, the company that preempts the Personal AI Agent will play the role of the new portal and dominate the market." A massive competition is expected to begin between existing leaders like Microsoft's Copilot, Samsung Electronics' Bixby, and Apple's Siri, and new forms of agents like Meta's Muse.

However, resistance from existing platform companies is expected to be significant. Companies such as Google, Naver, and Coupang will likely attempt to embed powerful agents within their own apps to prevent the penetration of external agents. Kim suggested the uncertainty of this change, noting, "Just as the winners of the web era were not necessarily the winners of the mobile era, the new winners of the agent era cannot be predicted by anyone."

Meanwhile, Kim added a warning regarding organizational management concerning the phenomenon where work accumulates on employees who are proficient in using AI. If core personnel who produce high performance through AI are left to work alone, the sustainability of the organization decreases. "To ensure long-term survival, companies must support and manage human resources so that employees who use AI well can continue to learn more," Kim advised.

Source: original video (YouTube)