When the chiefs of Coca-Cola and Walmart cited AI as a reason for stepping down, they raised a question every board should be asking: Does your leadership have what this moment actually demands?
A few weeks ago, two tech chief executives let AI avatars stand in for them during earnings calls. Employees and investors pushed back, unsettled by the blurred lines of accountability. The AI or the CEO — whose message was it, exactly?
Since then, two more chief executives have answered that question by leaving altogether.
James Quincey of Coca-Cola and Doug McMillon of Walmart both recently stepped down, and both cited AI as a factor in their decisions. Quincey told CNBC it was “time to put someone else on the field for the next wave of growth.” McMillon said he could start the AI transformation, but couldn’t finish it — that his company needed someone faster. Both framed their exits as acts of self-awareness: leaders reading the moment clearly enough to know when to hand off the baton.
That framing deserves to be taken seriously.
Skeptics will argue that AI is providing a clear, forward-looking explanation for exits that had more immediate causes. They may be right. But the more interesting question isn’t whether these men were right to leave. It’s what their reasoning reveals about what the AI transition actually demands of leaders.
Stepping Aside Can Be the Right Call
It’s genuinely difficult to know why any chief executive leaves. The factors are rarely singular, and outsiders are never privy to most of them. Stock performance, revenue trends, investor pressure — these offer clues, but they’re always a fraction of a much larger picture. Reducing these departures to any single explanation would be a mistake.
What is worth taking at face value is the self-diagnosis. A chief executive who looks honestly at what’s coming and concludes they’re not the right person to see it through is doing something difficult and valuable. That kind of introspection is rare at any level of leadership. At the CEO level, it’s rarer still. Done genuinely, it’s a service to the organization.
McMillon’s language was notably specific. He said he could see what agentic commerce was going to look like, understood the scope of the transformation ahead, and concluded that the timeline didn’t match his own. That’s not a failure of nerve. It’s a leader being honest about fit.
The question worth sitting with is what that fit actually requires. What does it take to lead a company through an AI transition — and why might the leader who built an organization to where it stands today not be the right person to take it where it needs to go?
What the Transition Actually Demands
We may be entering the single greatest change in how businesses operate since the Industrial Revolution. That’s not hyperbole. Every organization, regardless of industry, has large chunks of knowledge work and technology infrastructure woven into how it functions. AI is going to touch all of it.
Three distinct capabilities define what that transition demands — and each represents a genuine gap for some leaders, regardless of how accomplished they’ve been.
The first is vision. AI can make existing things faster, better, or cheaper. More significantly, it makes previously impossible things possible: new business models, new ways of serving customers, and organizational structures that didn’t exist before. Seeing that future clearly and building toward it requires a different kind of thinking than optimizing what already works. Not every leader who excelled at the latter possesses the former — and that’s not a criticism. It’s a recognition that the skills that built a company to where it is aren’t always the skills that take it where it needs to go.
The second is the ability to manage competitive threat in a fundamentally different environment. Every organization now faces disruption not just from established rivals but from companies reconfiguring entire business models from scratch — not improving on what exists, but replacing it. Leaders who built their careers navigating known competitive landscapes may find this particular challenge sits outside their range.
The third is change management at a scale the existing playbooks weren’t written for. Rethinking what’s done by people and what’s done by AI agents, what the organization looks like, how people are developed and moved through it, which processes get rebuilt from the ground up — this is a structural transformation of an unprecedented order. McMillon’s admission — “I could start this but couldn’t finish it” — points to a time-horizon problem more than a skills gap. Transformation at this scale requires sustained presence through years of disruption, not just clarity at the starting line.
The Dimension That Gets the Least Attention
There’s a fourth area that deserves more focus than it typically receives, and it may be the most demanding of all: stakeholder balance.
A chief executive’s job is to lead an organization on behalf of multiple constituencies — customers, employees, shareholders, and what might be called the fourth stakeholder: society at large. In a period of radical change, that balance becomes extraordinarily difficult. You may need to eliminate roles you only recently hired for. You may need to absorb lower margins for years while investing in capabilities whose returns are distant. You need to reckon honestly with what AI-driven decisions mean for the people they affect and for the broader systems your organization operates within.
Pure profit optimization is, in a sense, the simpler version of leadership. The math drives the decisions. It’s when a leader is genuinely trying to balance across all four stakeholders that the job becomes most demanding — and that balance is going to get harder before it gets easier.
A chief executive who looks at that complexity and concludes they’re not equipped to navigate it isn’t failing. They’re being precise about what the moment requires.
What Comes Next
The question these departures leave behind isn’t really about the leaders who chose to go. It’s about what organizations do with what’s coming.
AI is going to keep demanding more from leaders, and faster than most are prepared for. The incoming generation — at every level, not just in the CEO chair — will be asked to lead through something genuinely without precedent. The ones who do it well won’t necessarily be the fastest or the most visionary. They’ll be the ones honest enough to know which of these capabilities they have, clear-eyed about where they need to build or bring others in, and present enough to stay accountable through years of disruption rather than stepping aside when the difficulty peaks.
That combination of self-awareness and sustained commitment is rare. Right now, it may be the most important leadership capability of all.
