As artificial intelligence continues to reshape the corporate landscape, the traditional hierarchical structures that defined industrial-era organizations are being fundamentally challenged. The role of middle management, once indispensable for information relay and process control, is now under unprecedented scrutiny as AI agents take on execution tasks and organizations pivot toward scenario-driven workflows. The shift from department-based structures to project-oriented frameworks signals a profound transformation in how value is created and how leadership is defined.

Tan Liang, an Assistant Dean at the University of Hong Kong Business School and a professor of management and business strategy, offers a provocative take: in the AI-driven enterprise, there is no room for purely administrative managers. Those whose primary function is information transmission, procedural oversight, or reliance on positional authority will see their value erode dramatically. Tan, who pioneered the "1391 Strategy to Execution Framework" and has guided numerous Fortune 500 companies and leading Chinese tech firms through AI transformation, argues that this shift doesn't simply mean AI eliminates middle management—it fundamentally recalibrates what makes professionals valuable, amplifying judgment, expertise, and project leadership while making traditional competencies obsolete.

Who Gets Left Behind in the AI Era

The question of whether AI will replace or empower managers depends largely on the managers themselves, Tan emphasizes. His strong conviction is that organizations in the AI age should have no purely administrative cadres. These positions, which focus solely on management without touching core business operations, will face intense pressure. Middle managers must first avoid three dangerous states. The first is being a pure information conduit—merely passing down directives from above without any translation or contextualization. This strips away the middle manager's essential value: converting high-uncertainty strategic decisions into clear directives that frontline teams can execute rapidly. The essence of the role lies in decomposition and operational translation.

The second pitfall is complacency and ego-driven leadership. A manager whose self-importance exceeds their vision, demanding the team orbit around them rather than fostering collaboration, will find rapid elimination in the AI era. The third failure mode is using information asymmetry as a mechanism for process control—hoarding knowledge rather than sharing it. AI fundamentally democratizes information, and those who still rely on knowledge gaps for authority will have no ground to stand on. Rather than waiting for this shift to happen passively, Tan advises proactive self-reinvention: abandon the mentality of "I am a middle-level cadre," return to your essence as a professional, and ask: What industry am I in? What expertise do I possess? Can I become a trusted master craftsman? In the new organizational paradigm, there are no fixed positions but rather scenarios, projects, and task flows—what projects can I lead, and what value can I create in various contexts?

Beyond the Paradox of Middle Management

Observation that the internet once promised to eliminate middlemen, only to discover that distributors often provided essential robustness to business operations by absorbing shocks and smoothing fluctuations, raises a compelling question about middle managers. Do they serve a similar stabilizing function? Tan acknowledges the public narrative that AI will eradicate middle management but urges a more objective perspective. The middle manager role emerged from the industrial-era pyramid structure of bureaucracy. When paradigms rapidly reconstruct, all positions face redefinition. Echoing Tolstoy's observation in Anna Karenina about everything being in chaos yet simultaneously being rebuilt, Tan suggests we are in exactly such a moment of flux.

The meaningful question, he posits, is how those currently in middle management should approach the paradigm shift. His counsel: nothing is stable or permanent, so rather than being passively broken by change, one should actively embrace self-dismantling. The misfortune is experiencing paradigm reconstruction; the fortune is that new paradigms can liberate us. The path forward involves first breaking down old cognitive frameworks, then constructing new ones through action. This includes developing AI literacy, engaging in self-diagnosis to identify areas of professional depth where one might become a master expert, and understanding that new organizations are driven by scenarios, projects, and task flows—then asking whether one can collaborate with young "fresh builders" to form human-machine special operations teams.

The judgment becomes sharper here. The key phrase for what Tan calls the "A to A era" is whether capabilities are needed—not a question of effort, but whether certain skill sets retain relevance. A K-shaped differentiation will emerge in employment: those with deep expertise become "master craftsmen" with no need to clock out, their judgment amplified by AI agents into 24/7 productivity; while those who merely relay information will find themselves with no role to fill, entirely replaceable by AI agents. The growth path for a master craftsman follows a pattern: starting with intuition, undergoing rigorous rational training, then returning to intuition—the final insight appearing similar to the first, yet fundamentally transformed by the deliberate practice between. The distinction between an expert and a novice lies in that intervening period of structured learning and refinement. AI does not eliminate middle management per se—it eliminates outdated middle management capabilities.

What Truly Distinguishes Great Leaders

Drawing on years of cultivating successful executives, Tan identifies core characteristics that separate exceptional leaders from managers. These are fundamentally different roles: leadership is about doing the right things, while management is about doing things right. Leadership emanates from internal motivation and inspiration, while management relies on authority, titles, and resource allocation. From his experience at General Electric and observing industry best practices, Tan highlights three central requirements for true leaders. First, articulating expectations clearly—and importantly, this means not just outcomes but the full landscape of expectations involving both operational goals and human development. A true leader holds tasks and people in equal view, developing people through achievement.

Second, actively helping team members achieve their goals. Many leaders set KPIs and then disengage, but genuine leadership involves supporting the team's execution. Tan candidly shares his early career misstep: assuming his team could accomplish what he could personally achieve. The reality was different—position, cognition, and resource access all differ between leader and team. If others could do exactly what you do, why would you be needed? In the AI era, this helping dynamic also changes: leaders must become coaches who guide through questioning rather than directives. Third, leaders must allow employees to take responsibility for outcomes. Tan references Jack Welch's classic insight: before you become a leader, success is all about yourself; from the day you become a leader, success is all about others. In Alibaba's corporate lineage, this translates to "management as altruism"—the manager's essence is helping others succeed. Alibaba requires its senior leaders to demonstrate what it calls "three views and two missions": a global perspective, a holistic perspective, and a future-oriented perspective, combined with mission-driven action and the ability to mobilize people toward a shared vision.

The most accomplished leaders share two core qualities: seeing people—understanding human nature to inspire effectively—and seeing far, maintaining a compelling vision. Most business leaders operate from the premise that survival today precedes planning for tomorrow. But a more powerful mindset asks what the day after tomorrow should look like, then works backward to determine today's actions. In the paradigm-reconstructing AI age, this future-backward thinking offers a more expansive horizon for organizations and individuals alike.

Why Prompt Engineering Isn't the Priority

When asked what managers should most critically learn about AI and what pitfalls to avoid, Tan offers clear guidance. First, develop fundamental AI literacy. The English adage holds that common sense does not equal common behavior—without basic AI awareness, communication becomes ineffective across a chasm of misunderstanding. Second, approach learning through business orientation by clarifying three questions: which scenario in your team's operations best demonstrates AI value? What tasks can be reshaped by AI? What projects could achieve highest AI integration? The implementation path runs through entering scenarios, initiating projects, and restructuring workflows. Third, identify two key roles within your organization: the veterans with deep domain expertise and the fresh AI builders—typically young people who can rapidly prototype and iterate at speed. Notably, some large tech companies have already reformed their campus recruitment to eliminate traditional interviews entirely, replacing them with practical testing: give candidates an account, allow 48 hours to build something, then evaluate the results with engineers on-site.

Regarding concerns that AI-savvy managers might replace employees with AI agents and reduce headcount, Tan''s stance is nuanced. Businesses must use AI—hands-on experience is essential, since those who don't engage will retain outdated perspectives. Once you''ve built that familiarity, make a key decision: identify the three problems you''ll solve with AI in the next six months. As a leader who sees people, motivate and encourage collective growth. Business development and human growth can be organically linked—but organizations should not tolerate idle workers, those with outdated mindsets who refuse to adapt, or those simply coasting. Leaders must be decisive, but through inspiration and opportunity provision. If people fall behind despite the chances provided, there should be no regret—commit to the path and accept the consequences.

Small Self in a Grand Era

For Chinese enterprises advancing toward technological self-reliance and AI transformation, Tan offers one core piece of advice: be courageous and be a long-termist. We live in an era of grand narratives, but individuals should maintain a "small self" perspective. He references the Alibaba founder''s transformative moment at the Greenwich Observatory, where beholding the vastness of the cosmos brought profound humility—every person needs such moments to understand their place in the larger scheme. Human subjectivity operates on three levels: doing (active engagement), being (presence and existence), and becoming (evolution and growth). For years, under the great division of labor in society, people have been defined by doing, rarely returning to a state of simple being. Many organizations advocate "becoming a better self"—but that''s still doing, doing more. Tan''s philosophy: instead of focusing exclusively on being better, focus on better being oneself. The former emphasizes action and achievement; the latter emphasizes authentic presence. Balancing these two states enables genuine becoming.

In the AI era, drive comes from the fire within—that burning desire that manifests as energy and determination. Maintain openness, as exemplified by the recurring phrase "I see you" in Avatar—only through openness can we truly see best practices and spark curiosity. Maintain continuous learning—everything observed or heard should ultimately be processed through questions: how does this apply to my scenario? What concrete changes does it bring to me? Learning must translate into personal decisions and changes; otherwise, it''s merely information gathering, not genuine development. This conversation occurs in 2026, marking a profound intersection: AI completes its 70-year journey from laboratory to everyday presence, while humanity undergoes 70 years of deep division of labor that placed individuals as mere links in organizational processes. Now, an AI that can create from nothing and do almost anything meets a human being who has been fixed for 70 years into the formula "person equals position, position equals person." For most people, the initial response to AI is anxiety and fear. But the other side of the coin is liberation—AI releases humans from standardized work and offers the opportunity to return to things that make us distinctively human: aesthetics, curiosity, and judgment. Be friends with time; be courageous. Time ultimately rewards those who dare to embrace change and act boldly.