Summary of Key Points
Over the past decade, middle managers in large companies have been able to secure their positions by acting as a link between higher management and lower-level staff, responsible for conveying information, coordinating resources, and monitoring progress. They often relied on seniority or aligning with certain individuals within the company. However, with the rise of AI, these traditional roles have begun to be automated (for example, AI can break down tasks, monitor progress, and generate preliminary drafts). As a result, companies are adjusting their management models: ByteDance requires managers to work on the front lines, Tencent is piloting a project-based system, JD.com has eliminated middle levels of management, and Amazon has cut staff, particularly among middle-level employees. Middle managers must now prove their unique value—either by using AI to improve efficiency or by demonstrating skills that AI cannot replace, such as judgment and interpersonal abilities. The experiences of the five middle managers discussed in this article reflect their anxiety, struggle, and attempts at transformation in the face of these changes.
1. What Traditional Roles Have AI Taken Over from Middle Managers?
The core value of middle managers traditionally revolved around acting as a hub for information and processes:
- Conveying Information: Breaking down big goals into smaller tasks for subordinates and reporting on their progress to higher management (as mentioned by Huo Shan).
- Monitoring Progress: Urging subordinates to work and detecting any delays (AI can monitor progress 24/7 and make automatic adjustments).
- Initial Review: Checking documents and code drafts for errors (AI can quickly generate structured drafts, although they may contain mistakes, eliminating the need for middle managers to create the basic framework).
- Scheduling and Estimating Time: Setting deadlines for projects (Yin Jie’s subordinate completed what would have taken two weeks in three days using AI, rendering the traditional scheduling method ineffective).
AI performs these tasks faster and more efficiently, making middle managers dispensable if they only focus on these functions.
2. What Measures Are Large Companies Using to Force Middle Managers to Transform?
The requirements and management models for middle managers are undergoing fundamental changes:
- Elimination of Middle Levels and Reduction of Staffing: JD.com has removed two levels of management, leaving fewer people in charge of broader responsibilities; Amazon has laid off 78% of its middle-level staff (L5-L7).
- Changes in Evaluation: Performance is no longer measured by the number of employees managed or meetings held; instead, it focuses on whether AI can improve efficiency and generate real business value. For example, a manager who used to earn 100,000 yuan for a performance of 1 million yuan now needs to achieve 1.5 million yuan to receive the same salary, otherwise, their bonus will be significantly reduced.
- Require Managers to Get Involved Personally: ByteDance emphasizes that managers must work on the front lines and not just give orders; Zeng Xiaojian is now responsible for generating initial drafts using AI before allowing his team to iterate.
- Project-Based Systems Replacing Fixed Roles: Tencent is piloting a project-based system where team leaders and directors are assigned to specific projects, and their roles may end after the project is completed, making these positions less stable.
3. The Current Challenges Faced by Middle Managers
The experiences of the five middle managers highlight various difficulties:
- Increased Workload: Zeng Xiaojian’s team’s output has doubled with AI, but he must verify the accuracy of the AI-generated content and handle any issues (such as data errors or compatibility problems in the code environment), doubling his workload.
- Pressure to Focus on Metrics: Dong Ning, who was initially skeptical of AI, is now forced to focus on metrics required by management, such as AI usage and the proportion of AI-generated code.
- Ineffective Management Tools: Yin Jie’s subordinate completed the work in three days using AI, rendering the traditional scheduling and estimating systems obsolete; he even found that managing AI agents was similar to managing direct subordinates.
- Being Overseen by Subordinates: Lin Lin’s subordinate’s work is reviewed directly by higher management, who skip her for updates from the subordinates, making her feel marginalized.
4. What New Skills Are Required for Middle Managers to Survive?
AI does not aim to completely replace middle managers; instead, it requires them to develop skills that AI cannot provide:
- Using AI Effectively and Critically: Being able to use AI while identifying its limitations (for example, identifying data issues in AI-generated documents or making architectural decisions).
- Getting Involved Personally: Managers must not only direct tasks but also understand the business context (for example, actively learning how to implement AI or participating in technical work).
- Dealing with Human Issues: Tasks such as coordinating across teams, securing resources from higher management, and communicating team performance cannot be automated by AI (Yin Jie mentioned that 60% of work remains beyond AI’s capabilities, such as accident assessment and interpersonal coordination).
- Continuous Learning: AI evolves rapidly, and skills learned three years ago may become obsolete. Only middle managers who embrace change can keep up.
In short, middle managers cannot simply act as messengers; they need to become skilled operators who combine AI with human expertise to improve efficiency and address issues that AI cannot solve.
Conclusion
AI is not meant to eliminate middle managers altogether but those who are unable to adapt. In the past, middle managers relied on their position for success; now, they must rely on their abilities. Those who can adapt to changes, use AI effectively, understand the business, and handle human interactions will be more valuable. Those who remain stuck in their ways or focus only on superficial tasks may be eliminated. This wave of AI represents both a crisis and an opportunity for middle managers.