AI Automates Routine Management and Raises the Value of Leadership

Executive summary

Artificial intelligence is unlikely to eliminate executive roles, but it is changing the tasks that make those roles valuable. Drawing on research into task automation, 45 interviews with executives and AI specialists, and an exploratory assessment by seven generative AI systems, this policy paper identifies a consistent asymmetry. AI is increasingly capable of supporting or absorbing structured, data-rich management tasks such as performance monitoring, workflow coordination, compliance, and first-pass analysis. It is less able to replace leadership work that depends on judgment, trust, social intelligence, and accountability. This creates a subtraction effect: as routine management occupies less executive time, the human work of setting direction, shaping culture, motivating people, and leading change becomes more important. Organizations should respond by auditing executive roles at task level, automating routine management without delegating accountability, redesigning roles around leadership, and developing capabilities such as strategic judgment, narrative, negotiation, and cross-boundary influence. The aim is to allow AI to handle structured work while leaders remain responsible for people, ethics, and change.

Context and challenge

When Google co-founder Sergey Brin said that “management is like the easiest thing to do with AI,” he was describing how he had used AI to summarize internal communication, assign follow-up work and highlight the contribution of a quieter engineer who had been overlooked in promotion discussions. The remark may sound like Silicon Valley bravado, but it captures a question many executives are beginning to ask: if AI can do more of the work called management, what remains for managers to do?1

The anxiety extends beyond technology companies. In a 2024 survey of 600 CEOs, 43% said they believed AI could replace the CEO’s job. That result appears to signal a broader unease. Executives are no longer asking only how AI will change their teams. They are asking how it will change their own roles.2

The most useful unit of analysis is the task. Executive roles combine administrative, analytical, relational, political, and symbolic activities. AI will not affect each component equally. It is already making structured tasks such as monitoring performance, coordinating workflows, enforcing rules, summarizing information, and preparing analysis cheaper and faster. Other activities depend heavily on judgment, trust, timing, and relationships.

Management and leadership respond differently

A long-standing distinction associated with Harvard professor John Kotter helps reveal this asymmetry. Management maintains stability: it reduces uncertainty, allocates work, monitors progress, and keeps people and processes performing as expected. Leadership creates change: it sets direction, sustains a shared identity and culture, builds bridges across boundaries, and mobilizes people around a common purpose.6

These are functions, not separate categories of people. Most executives perform both. The distinction is important because management tasks are often structured, rule-based, and data-rich, while leadership relies more heavily on creative and social intelligence: generating responses in ambiguous situations, reading people and contexts, building trust, persuading, negotiating, and coordinating collective action. These capabilities are difficult to automate because their value depends on context, culture, relationships, and accountable judgment.7

The challenge for organizations is therefore to decide which managerial tasks AI should support or absorb, which decisions require stronger human oversight, and how executive roles should be redesigned when routine control work no longer occupies the same share of managerial time.

Key insights: the evidence

The exploratory task assessment and the interview evidence point in the same direction. The more a task depends on standardized information, formal rules, and measurable outputs, the more exposed it is to AI support or automation. The more it depends on imagination, trust, legitimacy, and relationships, the more central human intervention remains.

This does not produce a simple divide between management and leadership. Team development and complex client support are managerial tasks but remain strongly human. Competitive intelligence and drafting communications are associated with leadership but can be substantially augmented by AI. The consistent pattern is that AI is strongest at structured control work and weaker where a task requires meaning, accountability or social legitimacy.

Note: These ratings are exploratory relative indicators. They compare tasks within this analysis; they are not probabilities or forecasts of job replacement.

The subtraction effect

AI does not merely add tools to the executive role; it can remove or reduce tasks that once defined it. As performance monitoring, scheduling, compliance, and first-pass analysis are increasingly delegated to AI systems, the role does not disappear, but its content changes. Managers spend less time supervising processes and more time interpreting situations; less time enforcing routines and more time exercising judgment; less time controlling operations and more time engaging with people. Executive value shifts from maintaining control toward providing direction, trust, meaning, and change.

The solution framework

Organizations should not simply insert AI into existing routines. They should redesign executive work around four coordinated moves. The organization must change roles, incentives and governance, while individual executives change their calendars, skills and professional identity.

An accountability diagnostic

Before delegating a managerial task to AI, executives should ask:

5. Is the task rule-based and repetitive?
6. Are the relevant data reliable, representative and available?
7. Are outcomes measurable, and can errors be detected?
8. Could the task affect employment, dignity, trust or fairness?
9. Is there a clear process for human review, explanation and appeal?

Tasks that are structured, data-rich, and measurable, have limited human consequences, and sit within a robust review process are stronger candidates for automation. Where decisions materially affect people, AI should augment human judgment rather than replace it. The goal is not to let AI manage people; it is to let AI handle structured work while humans retain responsibility for ethics, exceptions, and trust.

Applications: the shift in practice

The interviews suggest that the subtraction effect is already emerging across sectors. In each example, AI reduces structured coordination or analysis and moves human attention toward interpretation, alignment, and judgment.

Sales and marketing

At a global wine and spirits company, AI helps teams monitor performance and allocate resources in real time. Faster, clearer information reduces the need for weekly meetings devoted to reconciling figures. The harder questions remain human: why local teams are not making the most of a campaign, whether incentives are aligned, and whether strategic priorities are understood on the ground. The technology improves information; leadership determines adoption, alignment, and responsible use.

Shipping operations

In a shipping group, AI systems optimize vessel schedules and reallocate resources as conditions change. Work that once required layers of managerial coordination can increasingly be supported algorithmically. Human intervention moves upstream: setting priorities, deciding which trade-offs are acceptable, communicating direction, and enabling teams to act confidently under pressure.

Financial services

In merchant banking, AI tools give portfolio advisers faster access to financial insights, reducing the time needed for client reviews and risk evaluation. The adviser’s value shifts from gathering and producing analysis toward interpreting it: explaining what the information means, how it should shape commercial strategy, and how clients should act.

Conclusion and key takeaways

AI is separating activities that have long been bundled inside executive roles. Routine management does not vanish, and leadership is not untouched by technology. But the balance changes. Structured monitoring, coordination, compliance, and analysis become easier to support or delegate, while judgment, legitimacy, trust, meaning, and change become more important sources of executive value.

This raises the standard for leaders. Operational busyness can no longer serve as proof of value when AI can handle more of the underlying process work. Executives must reclaim time for conversations and decisions that cannot be automated: listening to weak signals, aligning peers, coaching successors, negotiating across boundaries, and helping others make sense of uncertainty. AI may make executives less valuable as supervisors of process and more valuable as leaders of people.

Methodology and sources

The analysis combined four elements: a review of research on task automation and managerial work; the management-leadership distinction used to identify recurring executive tasks; a structured elicitation exercise with Claude Sonnet 4.5, ChatGPT-5, DeepSeek 3.2, Gemini 2.5 Flash, Grok 4, Mistral Large 2, and Perplexity AI; and 45 semi-structured interviews with executives and AI specialists from global firms in consumer goods, financial services, shipping, healthcare, aviation, insurance, and technology.

Each model received the same task list and prompt protocol, rating the creative intelligence, social intelligence, and relative susceptibility to AI support or automation associated with each task. Averages and ranges were used as directional indicators. The interviews served as an interpretive check, showing where model assessments aligned with organizational practice and where implementation, governance, or social concerns complicated technical feasibility.

Limitations

The results are exploratory, not forecasts. AI systems are imperfect judges of their own capabilities, and their outputs are not fully replicable. The interviews are interpretive rather than statistically representative. The task ratings should therefore guide comparison and discussion, not be read as automation probabilities or predictions of job replacement.

About the authors

Julien Jourdan is Professor of Management and Human Resources at HEC Paris, where he teaches management and leadership across the PhD, MSc, and MBA programs. His research examines reputation, legitimacy, and other social evaluations of organizations, and how these evaluations shape organizational behavior, strategy, and performance. 

Emmanuel Coblence is Professor of Management and Human Resources at HEC Paris. He teaches in the Executive MBA and Executive Education programs, where he coordinates leadership courses. His work focuses on leadership, organizational change, and managerial practices, particularly in complex and creative environments.

Catherine Tanneau is an adjunct professor of leadership at HEC Paris, an executive coach, and CEO of a coaching and consulting boutique advising CEOs and C-suite executives on strategy, organizational transformation, and leadership development. She partners with global companies and co-hosts the Transform Tomorrow Today podcast. She is also the author of several books on leadership and coaching.