AI has moved from experimentation to operating reality faster than most leadership models can absorb. As agentic systems begin to plan and act across workflows, the biggest constraint is not technical feasibility but whether leaders understand AI well enough to reimagine work, build trust, and capture value at scale.
In our examination of more than two dozen AI transformations, a distinct pattern emerged: AI adoption is outpacing leadership readiness.
The leader of a life sciences company told us that the biggest bottleneck to AI adoption and scaling is change management, which traditionally has meant adapting processes or structures. To manage AI-related change requires rethinking how organizations operate and create value, something that many leadership teams struggle with right now.
Those organizations moving the fastest are spurring leaders to use AI themselves so they can help teams operate with confidence as expertise, accountability, and roles are rapidly being rewritten.
Fluency comes from doing
For years, executives could lead technology transformations by asking sharp questions, sponsoring investment, and relying on experts. That model is no longer sufficient. Agentic AI changes the unit of work from a tool or a use case to an evolving workflow. Leaders need enough practical fluency to understand what AI can do and how enterprise-level value is created.
Not every leader must become an engineer, but they can stop treating AI as something they are briefed on. “You need to actually build your own agents to understand how it’s done, what the limitations are, and what the risk factors are,” one McKinsey practitioner who works on transformations told us. Another expert put it more directly: “The worst thing you can do as a role model is to explain AI without having used it yourself.”
The most effective examples make this shift tangible. One practitioner described a leadership immersion session when executives picked a major pain point from their personal lives and built an agent to solve it. The emphasis was not the use case itself but creating an “aha” moment when a leader feels that AI is a collaborator rather than a concept.
In another example, a global bank put senior leaders through a hands-on AI training session. Leaders left energized and with a new sense of the company’s value creation potential.
AI fluency can become a management expectation, not an optional learning module. Leaders who follow this mantra work with AI on real problems, not generic demos. They can do this through reverse mentorship, by teaming up with digitally fluent colleagues who may be junior in hierarchy but senior in practice. This reverses a long-standing leadership instinct: Expertise no longer always flows downward.
Human–agent teams require new management
The manager’s role is also changing dramatically. In the AI era, managers will increasingly orchestrate work across humans, agents, and systems. In a human-in-the-loop workflow, agents do parts of the work while humans approve critical decisions.
That shift creates a new burden of discernment. Managers will see more output faster, but of more variable quality. They will need to ask: Is this answer right? Is it relevant? Can the team defend the logic? Where does the risk sit? One practitioner we interviewed framed the managerial challenge as, teams will bring leaders “AI slop,” and managers must know both how to spot it and how to coach the person who produced it.
A corporate services example illustrates what this looks like in practice. One six-person team, working on an ambitious agentic build, mapped accounting and bookkeeping processes quickly. The team captured business and functional requirements and moved away from PowerPoint-based sprint reviews toward live, codified outputs. The client was “super impressed,” not because the team used AI as a presentation enhancer, but because it changed the speed and structure of problem-solving itself.
Organizations must keep humans at the center, especially where adoption is high but tangible impact on business outcomes is murky. Leaders can set the expectation that managers evaluate output, challenge assumptions, set escalation paths, and decide where human judgment must remain paramount. This is where ethical judgment becomes operational, as accountability gets designed into workflows, governance, and day-to-day management.
At a three-day summit for a global life sciences organization, 150 HR leaders worked to reimagine the future with AI. One focus of the summit was on translating the accountability principle into the mechanics of work. Leaders explored how a weekly operating review could be redesigned for the AI era by adding a rotating role where someone wore a “challenger” hat. This person’s brief was to provide a different perspective and intentionally seek out evidence that contradicted the leading conclusion before decisions were made. The person kept a simple log showing what AI presented compared with human decision-making. The goal was to increase speed while preserving accountability, judgment, and trust.
Adaptive capacity is key
AI raises the performance bar, but it also increases fear. Teams worry about job relevance, emotional and cognitive overload, deskilling, and whether today’s expertise will matter tomorrow. Leaders who ignore that fear will slow the transformation, as will those who overindex on reassurance without changing the work.
The answer to these challenges is stronger, more adaptive leadership. AI will continue to evolve faster than most organizations can absorb. The companies that capture value can’t wait for perfect tools or perfect certainty. Instead, they can support leaders who get their hands on the technology, managers who can orchestrate human–agent work, and cultures that learn fast without losing accountability or humanity.
In the organizations moving fastest, AI fluency among senior leaders is already a differentiator. The gap between leaders who have built something with AI and those who have only been briefed on it is showing up in decisions, expectations, and the confidence they can give their teams. Leaders who have experienced AI’s limitations firsthand can hold its complexity honestly and set expectations without overselling or catastrophizing. They can also address the harder human questions: what work stays human, what the productivity dividend is for, and what the organization stands for in a transition that employees are watching closely. That credibility is hard to fake. Staying on the sidelines may be comfortable, but it is increasingly untenable.