In a workshop, a senior manager explained to me why a particular decision needed five approvals. “Because the people below don’t have the full picture.”
Then he heard himself. His own teams now work alongside AI that retrieves every policy, every precedent, every market signal in seconds. He corrected himself, quietly, almost to himself: “They didn’t used to have the full picture.”
That two-second pause explains this entire decade of organizational life.
Hierarchy was an answer, not an ideology
Hierarchy was never mainly about power. It was a rational answer to information scarcity.
This matters because most critiques of hierarchy treat it as a moral failing — the residue of ego, empire-building, or a controlling temperament. Sometimes it is. But the structure itself was a sensible engineering response to a real constraint. When context lived in filing cabinets, in regional offices, and in the heads of people who had been there twenty years, concentrating decisions where the context had accumulated was simply correct. You moved the decision to the information, because the information could not be moved to the decision.
Everything we recognise as corporate architecture follows from that one constraint. Layers exist to aggregate context upward. Approval chains exist because the person closest to the work genuinely could not see enough. Reporting lines, escalation paths, the annual planning cycle — all of it is machinery for routing scarce information to scarce judgement.
The org pyramid is fossilized information scarcity. It is the shape of a constraint that no longer exists, preserved in policy.
Why Teal kept stalling — and why the sceptics were right
For years, Frédéric Laloux and others showed us organisations that run like living systems: self-managing teams, distributed authority, no permission required to do obvious work. The examples were real and the results were real. And at enterprise scale, it kept stalling.
I want to be fair to the sceptics here, because I think they were largely right, and pretending otherwise is why so many self-management initiatives collapsed in a heap of goodwill. Their objection was not philosophical. It was economic: coordination was too expensive.
Give three hundred people genuine autonomy and you do not get emergence. You get three hundred people each spending a large fraction of their week finding out what everyone else decided. Someone has to reconcile it. Coordination cost rises faster than headcount, and past a certain size the meetings required to stay aligned consume the capacity that autonomy was meant to release. Hierarchy, for all its friction, was the cheaper coordination technology. That was a real finding, not a failure of nerve.
What AI actually changed
AI just took the sceptics’ point away.
Not by making people better at self-management, and not by making autonomy more appealing. By collapsing the price of the thing that made autonomy unaffordable:
- Context is abundant. The full picture — policy, precedent, current state, what the other team decided on Tuesday — is now retrievable in seconds by whoever needs it.
- Coordination is nearly free. Synthesis, summarisation, translation between functions, keeping track of what changed: this was the expensive human labour that made distributed work impractical, and it is precisely what these systems are best at.
- Expertise flows to the work instead of being hoarded above it. The senior person’s advantage was never only judgement; it was accumulated context. That part has largely commoditised.
So the calculation inverts. Self-management stops being idealism and becomes the economically rational design. And the five-layer machine — which used to be the efficient option — is now the one carrying the extra cost. Every layer is a delay you are paying for, buying a benefit that has evaporated.
This is why I think the AI conversation inside organisations is aimed at the wrong target. Teams are asking how to use AI to make the existing structure faster. The more consequential question is which parts of the existing structure the technology has just made unnecessary.
Teal isn’t a destination — it’s permission
I want to be careful with the conclusion, because “abolish hierarchy” is not it, and organisations that hear that message tend to remove the structure without replacing the function it served.
Teal isn’t a destination. It’s permission.
Not a target state to be migrated to over eighteen months with a change programme and a new operating manual. Permission — to stop defending a layer whose original justification has dissolved, to let a decision sit where the work is, to stop asking a question upward that the person asking can now answer better than the person being asked.
Which means the practical move is small and specific rather than sweeping. Not a redesign. A subtraction, tested.
The human truth underneath it
And here is the part that no operating model will solve for you, which I see constantly in the coaching room.
Giving authority away feels like loss before it feels like liberation.
Leaders spend thirty years being rewarded for holding the picture. That is what got them promoted. It is what makes them useful in a crisis, what they are known for, and quite often what they privately believe justifies their salary. Letting the system hold the picture instead is not a process change. It is a change in what a person understands themselves to be for — and it lands in the body long before it makes sense in the mind.
That is real inner work, and it deserves compassion rather than a change-management slide. Every distributed-authority initiative I have seen fail, failed here — not because the design was wrong, but because it asked senior people to hand over the one thing that had made them feel valuable, and offered them nothing in exchange but a diagram.
One experiment for this week
Take a single decision that currently needs approval from someone one or two layers above the work. Ask one question about it: what does that approver know that the person doing the work cannot now find out in under a minute?
Sometimes the answer is real — a legal exposure, a commitment made elsewhere, a piece of judgement earned over years. Keep that layer; it is doing work.
More often, the honest answer is: nothing. That approval is a fossil. Remove it for ninety days and watch what actually happens.
Hierarchy answered a question that AI has since answered better. The layers that remain should be the ones that can say what they are for.
This is the Distributed Authority pillar from The Living Organization — The New Operating Model for Flourishing in the AI-Native Era.
📚 Start here: AI-Native Organization Design — the full research hub, with every article and video in one place.
More in this series
- 67% of AI Adoption Blockers Are Organizational, Not Individual
- $2M Revenue Per Employee: The Case for the Living Organization
- The Living Organization: Five Pillars of an AI-Native Operating Model
- Is HR Ready to Design Human and AI Agent Productivity Together?
- How Much Does a CEO Really Need to Understand About AI?
- The Five Steps of AI Adoption: From Gated to AI-Native
- Your AI Isn’t Failing — Your Specification Is
- Stop Measuring Tomorrow’s Talent With Yesterday’s Rulers
Watch the full series
This article accompanies a video from AI-Native Operating Model and Org Design — a series on how organisations actually absorb AI, and where they break.
- ▶️ Watch the full playlist — AI-Native Operating Model and Org Design
- 📺 Subscribe on YouTube — @BreezeDONG
Chunfeng “Breeze” Dong is an executive coach (ICF PCC, CPCC) and founder of Springbreeze Ventures, with twenty years in organisational development inside Fortune 100 companies — Roland Berger, Siemens, ABB and Roche. She writes on AI-native organisation design, human–agent governance and change leadership.
📘 The Living Organization · 📘 A Soulful Transition · 🔗 LinkedIn
