AI is compressing the middle layer, and the function of the middle layer has already been rewritten

Writing: Fang Dao

The discussion about “middle management disappearing” is, in essence, not a question of whether a job category lives or dies. Rather, it’s about a set of functions within organizations being systematically replaced.

When AI Agents start to have the ability to decompose tasks, execute processes, and provide result feedback, the information-processing and SOP-execution work that enterprises previously relied on humans to do is transforming into a capability that can be automatically completed by systems. What is being compressed is never a particular hierarchical level; it is “the portion of work that can be written into a process.”

Traditional companies need middle management not because management itself is the problem, but because the cost of information processing is high. The strategic plans from the top need to be translated into concrete tasks, and execution at the frontline needs to be consolidated into structured feedback—this back-and-forth forms the foundation for the existence of middle management.

But when information can be generated and synchronized in real time by systems, and when tasks can be automatically broken down and tracked by Agents, the necessity of this intermediate link begins to decline. Organizations no longer rely on “people to transmit information,” but start to rely on “systems to run processes.”

This shift shares the same underlying logic as the replacement of frontline roles. Frontline roles get replaced because their work is highly dependent on standardized processes. And the parts of middle management being compressed are also centered on processes—breaking down processes, transmitting processes, monitoring processes. When the system takes over the process itself, the difference between middle and frontline roles along this dimension starts to converge. What is being replaced is not a position, but the work structure.

Therefore, the organizational-structure change that is happening is not simply “flattening,” but a form of “eliminating the transfer layer.” Information no longer needs to be passed step by step through layers, and execution no longer depends on manual oversight. The internal logic of hierarchy within enterprises shifts from “information asymmetry” to “capability distribution.” In this new structure, what is retained is no longer a position, but capability itself.

Middle management is beginning to be re-divided into two types of roles. One type depends on process operation; its value declines as system capabilities improve. The other type can define goals, organize resources, and be responsible for results; its value is amplified instead. The core of the former is to ensure the process doesn’t go wrong. The core of the latter is to get the system running through uncertainty.

This is also why “knowing how to use AI” is discussed so frequently. But more accurately, the statement should be: the ability to embed AI into business processes and form stable outputs is becoming the new threshold. AI itself will not automatically create value. It needs to be scheduled, constrained, and incorporated into the organization’s operating system. Whoever can complete this orchestration will occupy the new structural position.

If we push this change one step further, we can see that the enterprise form is undergoing a deeper transformation. Companies are no longer just a human labor organization; they are gradually evolving into a “capability scheduling system.” In this system, some capabilities come from people, some from models. What is truly scarce is the ability to integrate the two into a sustainable operating structure.

Middle management will not disappear, but “the pricing logic of middle management” has already changed. In the past, value came from information asymmetry; in the future, value will come from system control. After AI compresses the value of information, what truly remains within organizations will only be those nodes capable of defining rules, orchestrating systems, and being responsible for outcomes. In this sense, enterprises no longer need more “intermediaries,” but need more “system nodes.”