Jetty Luppens · 29 March 2026

Over-governed in meetings, under-governed in decisions.

Over-governed in meetings, under-governed in decisions.

Why organizations stall today — and how you break through it in the AI era

There's a pattern you see in a lot of organizations today: calendars are packed, consultation structures are elaborate, everyone gets involved … and still, decisions don't get made.

Or worse: they seem to get made, but nobody actually feels ownership of them.

In other words, we're often over-governed in meetings, but under-governed in decision-making.

In a relatively stable environment, that's already a problem. In a context of fast technological change, rising market pressure and the advance of AI, it becomes outright dangerous. Organizations that can't decide clearly, quickly and consistently today will lose not just speed tomorrow, but direction.

Research and publications from, among others, McKinsey and Harvard Business School keep pointing to the same underlying weakness: many organizations confuse consultation with governance, being consulted with ownership, and alignment with decision mandate.

The real problem: the appearance of governance

Many organizations don't have too little governance. What they lack is clear governance.

There are steering committees, working groups, leadership meetings, transformation boards and project forums. Roles seem filled, check-ins are scheduled, decision-making bodies exist on paper. But the moment a critical choice has to be made, confusion sets in:

McKinsey has long described how organizations get tangled up in unclear decision rights. The limits of RACI—and a better way to make decisions makes it equally clear that models and roles alone aren't enough when nobody has explicitly defined who ultimately decides.

The result is predictable:

Why this is more acute today than ever

The rise of AI doesn't just make this weakness more visible — it makes it more damaging.

AI forces organizations to make choices:

Those aren't just technical questions. They're governance questions.

In organizations where decision-making is diffuse, you see the same pattern over and over:

But in reality, that's often not inclusiveness. It's postponement.

A further warning sign: when real expertise is kept out of the core

There's a second, often overlooked symptom of weak governance: the people with the most substantive expertise aren't genuinely included in decision-making.

Not because they contribute nothing, but often precisely because they bring something that's hard to hear.

The technical expert points out limitations. The risk owner names vulnerabilities. The data expert exposes that the foundation isn't in order. The legal or compliance officer slows down the desired enthusiasm with necessary nuance. The operational people make clear that an idea sounds good on paper but doesn't work in reality.

And that's exactly where it usually goes wrong.

Harvard Business School shows how leaders themselves often suppress dissent: by not really listening, cutting people off, or entering a discussion with the outcome already fixed in their mind. McKinsey makes a similar point: leaders often say they value pushback, but react defensively the moment that pushback threatens the group's direction or comfort.

That's a crucial observation. In many organizations, experts aren't excluded because they know too little, but because they see too much.

As long as expertise confirms what people already wanted to do, it's welcome. The moment expertise causes friction, it gets labeled as:

What's left then isn't strong leadership — it's a form of institutional self-confirmation.

The danger of this in an AI context

In classic change programs, marginalizing critical expertise is already risky. In AI, it's outright dangerous.

AI specifically requires organizations to make room for:

When exactly those voices are kept out of the core of decision-making, you get false certainty. It looks like alignment, but in reality it's selective information processing. Then leadership mostly hears what sounds workable, desirable or reassuring — not necessarily what's true.

Harvard Business School connects this to the suppression of dissent in organizations. McKinsey stresses the importance of "contributory dissent" as a precondition for better decisions. MIT Sloan Management Review adds an important nuance: expertise shouldn't be followed blindly, but it should be weighed critically and seriously. That's fundamentally different from systematically shutting expertise out because it's inconvenient.

How do you recognize this pattern in your own organization?

Many organizations are deeper into this than they realize. A few signals:

A simple litmus test: can you say, in one sentence, for an important issue, who ultimately decides, who advises, and which experts must be heard?

If you can't answer that sharply, you don't have a capacity problem — you have a governance problem.

The core of the solution: clarity, transparency and leading by example at the top

Many organizations try to solve this by adding even more structure: extra meetings, additional governance boards, new roles, heavier checklists.

That rarely helps.

The solution isn't more governance, but clearer governance:

But for me, this unmistakably starts at the very top.

If the top of an organization stays ambiguous about mandate, responsibility and direction, that ambiguity trickles down to every level below it. Teams then start feeling their way, protecting themselves, escalating, and reconfirming things too. The organization looks perpetually "busy", but everything feels like it's stuck in escalation mode without tangible results.

Harvard Business School is clear on this: if open debate and clear decision-making don't start at the top, they won't survive anywhere else in the organization either.

Leading by example: what the top actually needs to do

If leaders genuinely want to reverse this, it takes leading by example. Not slogans — discipline.

1. Make decisions explicit Every strategic meeting should end with three clear outcomes:

No implicit consensus. No "we're roughly on the same page". No decision that only turns out to have been made later.

2. Separate advice from decision Not everyone at the table decides. And that's healthy. But then make explicit:

That's also the essence of the critique McKinsey makes of superficial use of RACI: frameworks don't help if the real mandate stays unclear.

3. Organize critical expertise structurally, not optionally If the expert only joins the table when it's convenient, expertise isn't part of governance — it's decoration. Decide in advance which expertise must be involved in strategic choices, especially in AI matters:

4. Visibly reward pushback McKinsey stresses that leaders shouldn't just tolerate dissent — they should actively ask for it. That means:

5. Stop using escalation as the default mechanism Not every tension needs to go upward. When everything has to go to the top, that's usually a signal that mandates lower in the organization aren't clear enough. Escalation should be the exception, not the operating model.

6. As the top, be willing to actually change your mind Nothing undermines psychological safety faster than leaders who formally ask for input but have already decided in substance. If expertise or good counter-arguments never change anything, the organization learns quickly enough that speaking up isn't worth much.

What this delivers

When the top is clear on mandate, transparent about decision-making, and visibly makes room for expertise, the organization changes fundamentally:

That's when something emerges that many organizations are missing today: direction with confidence.

In closing

In many organizations, it looks like there's permanent motion but no real progress. That's rarely a lack of intelligence, effort or ambition. It's usually a lack of clarity about who decides, who gets heard, and which behavior actually gets rewarded at the top.

And it's precisely in an era of AI that this becomes mercilessly visible.

Because AI doesn't just call for new technology. It calls for mature governance. It calls for leaders who dare to decide quickly, who are transparent about mandate, and who are strong enough to hear what they'd rather not hear.

So the real test of governance isn't how many meetings you have, how many structures exist, or how many people get involved.

The real test is: are the right decisions being made clearly, by the right people, based on the right expertise — even when that expertise is uncomfortable?

Because if ambiguity persists even at the top, it inevitably trickles through the rest of the organization. And then you get no agility, no execution power, and no real AI transformation.

What you get instead is an organization that constantly escalates, but rarely actually decides.

This article is part of The Organizational Brain — our framework for organizations that get stuck on thinking.

Hidden Connections · Back to Our Thinking