CEOs Must Lead Change, Not Chase It

Treating the CEO as the singular "change agent" sounds efficient. It’s also a neat way to outsource responsibility — and to misunderstand how big organizations actually change.

McKinsey’s "agentic age" framing is useful in one respect: it forces CEOs to think explicitly about how goals become decisions and how decisions trigger action. Sure, but the phrase itself conjures a lone conductor waving a baton while the orchestra magically changes tempo. Isaac Asimov’s Foundation played with that fantasy — one genius plan steering history — and showed how brittle it becomes the moment reality stops cooperating. The article does real work outlining what CEOs should be considering as agents — software, new business units, organizational initiatives — start acting on their behalf. It just leans a bit too hard into a hero narrative that underplays three risks most boards and executives should lose some sleep over.

Let’s start where McKinsey is directionally right: CEOs can’t duck this. They are accountable for how AI agents, automation, and reorganized teams get deployed. They do have to cut through inertia. A passive CEO in an "agentic age" is basically a referee standing in the middle of a rugby scrum, hoping the rules enforce themselves.

Yeah, no.

The trouble begins when the CEO is cast as the change agent rather than a node in a bigger system.

First: governance friction. When you present the CEO as the central driver of change, you risk positioning them against every other formal power center. Boards are not stage props; they hold fiduciary duties that complicate any "move fast" agenda. Shareholders press for returns on specific timelines, employees anchor around stability and predictability, and regulators care less about transformation than compliance. The McKinsey piece notes implications for CEOs, but it doesn’t really dig into the chain reaction: decisions framed as bold agency moves can activate board risk committees, audit reviews, and regulatory scrutiny that slow or even reverse those same moves.

That’s not a small footnote — that’s the plumbing. Treating decisions as if they flow unimpeded from CEO intent to organizational action makes them look cleaner than they ever are. A CEO can set a goal on Monday and run into board risk flags, union provisions, or creaky IT systems by Wednesday. That gap between intent and implementation is not just operational noise; it’s exactly where many "agentic" strategies die or turn into performance theater designed to satisfy headlines and investor decks.

Then there’s the way goals become metrics, and metrics become weapons.

The McKinsey framing orbits around goals and decisions but doesn’t linger on how quickly measurement can distort behavior. Attach "change" to a CEO’s dashboard and you create incentives for everyone beneath them. People optimize what’s counted, not what’s meaningful. If the CEO’s mandate is "show measurable transformation," middle managers will carve up projects to generate visible wins, even if that undermines longer-term capability-building.

This isn’t an argument against data. It’s an argument against unexamined metrics. Targets set at the top become the language of the organization; choose the wrong language and you teach people to speak in short-term, surface-level progress. I’ve watched a large bank proudly hit its "AI use case" target, only to discover that most use cases were trivial workflow tweaks that added reporting burden without freeing real capacity. On paper: agentic triumph. On the ground: resentment and more meetings.

That’s how you get a brittle form of agency — movement you can point to on a slide, but not the kind of transformation that survives leadership changes, economic shocks, or regulatory updates.

There’s a darker twist, too: agency as alibi.

When you center CEOs as sole change agents, you give the rest of the system a convenient story: if something fails, it must have been a bad call by the hero at the top. That narrative erases the harder work of rethinking processes, reskilling teams, or realigning incentives. The CEO becomes the lightning rod for both praise and blame, and everyone else keeps their structures largely intact.

That framing feeds a churn cycle: hire a "transformational" CEO, anoint them as the agent of change, watch them collide with governance constraints and incentive mismatches, then fire them for not transforming fast enough. Reset. Repeat. The organization learns almost nothing, except how to survive another leadership swap.

There is an alternative path hiding inside McKinsey’s own thesis about agents: treat agency as a distributed property. Change is more credible when boards, executive teams, unions where they exist, and the technology stack all share explicit responsibility for outcomes.

Here’s the thing: that doesn’t make the CEO weaker; it makes their job more realistic. A CEO who walks into the boardroom with not just a transformation goal and timeline but a governance map — which committees own which risks, which policies must be rewritten, how unions or worker councils are engaged, where regulators will be briefed — is less romantic and far more likely to get real authority. The same applies inside the firm: executive teams should be evaluated on how they enable others to act — through budgets, tools, clear decision rights — not just on what the CEO orders and announces.

History backs this up. When Toyota pushed lean manufacturing, the mythology often centered on visionary leadership. But what actually changed factories was a system of shared agency: line workers empowered to stop production, managers rewarded for surfacing problems, and processes that treated information flow as everyone’s job. No single "change agent" could have willed that into existence from the corner office.

The McKinsey piece is right about the moment: agents — human and algorithmic — are no longer side characters. But when you crown CEOs as "the change agents" without also rewriting governance, measurement, and shared accountability, you don’t get an agentic age. You get faster decisions at the top and a more dramatic crash when those decisions hit the same old walls.

Treat CEOs like catalysts in a lab full of other reactive elements, and the experiment might actually produce something new instead of the same formula rebranded for the next earnings call.