Throughout my career I have worked alongside both deep specialists and broad generalists. Both matter. A surgeon should not be a generalist. Neither should a structural engineer.
But leadership is different. And AI is making that difference more important than ever.
What David Epstein Got Right
In his 2019 book Range, researcher David Epstein made a provocative argument: in complex, unpredictable environments, generalists consistently outperform specialists. He called these environments “wicked,” meaning the rules are not fixed, feedback is delayed or misleading, and experience in one narrow domain does not reliably transfer to the next problem.
Epstein contrasted two paths to excellence. Tiger Woods, who held a golf club at eighteen months and was optimized for one domain from the very beginning. And Roger Federer, who played a dozen sports as a child, came to tennis relatively late, and built a foundation of adaptable movement and pattern recognition that eventually made him one of the greatest players in history.
Epstein’s point was not that specialization is bad. It is that early, narrow specialization is optimized for stable environments with clear rules. And most leadership environments are anything but stable.
The argument was compelling in 2019. In 2025, it has become urgent.
What AI Is Actually Very Good At
To understand why, it helps to be clear about what AI does well.
AI excels at defined tasks with large amounts of training data, pattern recognition in structured domains, generating first drafts, summarizing information, analyzing data, writing code, and identifying anomalies in predictable systems. In short, AI is an extraordinarily capable specialist.
The World Economic Forum’s Future of Jobs Report 2025 assessed more than 2 800 distinct workplace skills and found that none of them had high potential to be fully replaced by current AI tools. The skills showing the lowest substitution potential were those rooted in human interaction, empathy, active listening, nuanced judgment, and contextual reasoning.
This is not a coincidence. It reflects something fundamental about how AI works. It optimizes within defined parameters. It does not know when the parameters themselves need to change.
The Generalist’s Unfair Advantage
This is precisely where the generalist leader comes in.
Consider a leadership decision about whether to automate a significant part of a company’s operations. A deep specialist in logistics technology might see the efficiency gains clearly. A deep specialist in organizational psychology might see the human cost clearly. But the leader who needs to make the decision has to hold both simultaneously, and also factor in customer experience, competitive dynamics, regulatory context, and the organization’s capacity for change.
That kind of thinking, connecting insights across domains, knowing enough about several fields to ask the right questions, is what Epstein describes as the core generalist skill. And it is precisely what AI cannot do.
McKinsey describes the emerging model as “human-led, AI-enabled,” where productivity gains come not from replacing human judgment but from freeing leaders to focus on it. The implication is clear: the leader’s job is shifting from execution and information processing toward orchestration and synthesis. From knowing the answer to knowing which questions matter.
The Data Behind the Argument
The WEF Future of Jobs Report 2025 identifies the skills most valued by employers through 2030. Analytical thinking tops the list, followed by resilience, flexibility and agility, and leadership and social influence. Creative thinking, curiosity and lifelong learning all rank in the top ten.
These are not specialist skills. They are generalist skills. The capacity to move between contexts, to see patterns across domains, to adapt when the environment shifts. The report also finds that eight of the ten core skills projected to grow most by 2030 are what researchers call durable human skills, competencies that develop through breadth of experience rather than depth in a single field.
Deloitte’s Human Capital Trends report adds a concrete business case: organizations that invest in developing these broader human capabilities are 1.8 times more likely to report stronger financial results than those that do not.
The Paradox at the Heart of the AI Age
Here is what makes this genuinely interesting rather than just reassuring.
As AI gets better at specialist tasks, the relative value of generalist thinking increases. Not because generalists become smarter, but because the bottleneck shifts. When information processing and routine analysis are handled by machines, the scarce resource becomes the judgment to direct them wisely.
Epstein captures this in Range with a deceptively simple observation: the most important skill in a changing world is the ability to recognize when the rules have changed. Specialists, by definition, have invested deeply in one set of rules. Generalists have learned to question them.
In a world where AI handles an increasing share of the execution, the leader’s primary contribution becomes something closer to wisdom than expertise. Knowing not just how to do something, but whether it should be done, for whom, and at what cost to everything else.
What This Means in Practice
None of this is an argument against deep expertise. The best generalist leaders are not shallow. They have developed what Epstein calls “outside-in” thinking, the ability to bring frameworks from one domain into another and see connections that pure specialists miss.
What it does suggest is that organizations may need to rethink how they develop and evaluate leaders. A career path that rewards narrow, deep expertise at every stage may be optimizing for a world that AI is rapidly changing. The leader who has worked across functions, industries, or disciplines, who has been uncomfortable and adapted, may be more valuable than their CV currently signals.
It also suggests something about how individual leaders should invest their own development time. Reading widely, not just in your field. Taking on projects outside your comfort zone. Building relationships across disciplines. These habits, often dismissed as unfocused, may turn out to be exactly the right preparation for leading in an AI-enabled world.
The Leader AI Cannot Replace
The question is not whether AI will change leadership. It already is. The question is what kind of leader becomes more valuable as a result.
The evidence points consistently in one direction. Not the narrowest expert, but the most adaptable thinker. Not the one who knows the most about one thing, but the one who can hold complexity, connect dots across domains, and make wise decisions in situations where there is no established playbook.
Federer did not win because he specialized earliest. He won because he built a foundation broad enough to adapt to anything.
The same may be true of leadership in the years ahead.