AI-First Organization: The Company With No Departments

The next great organizational shift may not be a new technology. It may be the quiet end of the department itself.

Every company I have worked in was built the same way. Finance over here. Operations over there. IT, HR, Product, each with its own leader, its own budget, its own reporting line. We have organized work like this for so long that we rarely stop to ask why. The department simply feels like the natural shape of a company.

It is not. It is a solution to a problem we no longer have.

Please note: The views expressed on this website are my own and reflect my personal professional experience, research and publicly available information. They do not represent the views, strategies or confidential information of my employer or its affiliated companies.

For more than a century, coordination was slow and expensive. Information moved at the speed of meetings and memos. So we grouped people by specialism and stacked managers above them whose real job was to move information between the groups. The department was the most efficient answer to a genuine constraint.

AI has quietly removed that constraint. And once the reason for a structure disappears, the structure is worth questioning.

We organize around functions. Work happens around outcomes.

Most companies are organized around what people do. Finance does finance. Operations does operations. The customer, meanwhile, has never cared which department was involved. A customer wants an outcome - a problem solved, a product delivered, a claim paid.

The gap between how we are organized and what we are trying to produce has always existed. We bridged it with cross-functional teams, matrix reporting and an enormous amount of coordination - most of it spent simply carrying information across silos.

That bridging work is exactly what AI now does on its own. When a system can pull data across functions, see what matters and route the work to where it belongs, the cost of crossing a silo collapses. And when coordination becomes cheap, the silo becomes optional.

This is why I have come to believe the most important organizational shift of the next decade is not the matrix, the agile model or the product organization. It is the move from organizing around functions to organizing around outcomes.

This is not a new idea. It has already been tested.

The most striking proof that an outcome-based, networked organization can beat a functional hierarchy does not come from a strategy deck. It comes from a war.

In Team of Teams, General Stanley McChrystal describes commanding a force in Iraq that was better trained, better equipped and better resourced than its enemy - and losing. The reason was structural. His organization was a disciplined hierarchy, superb at predictable problems. The enemy was a decentralised network, able to adapt and reconfigure overnight.

His response was not to become more efficient. It was to become more adaptive. He did two things. He made information radically transparent, so everyone could see the whole picture at once - he called it shared consciousness. And he pushed decisions down to the people closest to the problem rather than escalating everything upward. His own role shifted from chess master, moving every piece, to gardener, creating the conditions and then stepping back.

The result was a team of teams: the trust and speed of a small unit, scaled across a large enterprise.

It is the same structure I am describing here. The argument is not that companies should imitate the military. It is that the hardest real-world test of dismantling functional silos in favour of an outcome-focused network was run under combat conditions, against a live adversary, and it worked.

What is new: the team now includes machines.

Here is the part McChrystal could not have written.

Shared consciousness was the engine of his transformation - and it was punishingly expensive to produce. He achieved it through sheer human effort: thousands of people in daily briefings, liaison officers embedded everywhere, a vast manual machine whose only purpose was keeping everyone aware of everything. It worked, but it cost an extraordinary amount of human time and attention.

AI does cheaply what he did expensively. Shared awareness across an entire organization, in real time, is now an automated capability rather than a heroic feat of logistics. The single biggest barrier to this kind of organization has fallen.

And there is a deeper change still. In McChrystal's day, every member of every team was a person. That is no longer true. AI agents are becoming working members of teams - not tools we pick up, but participants that monitor, coordinate, synthesise, draft and increasingly act. We keep discussing AI as a productivity story. It is really a structural one.

A word of caution follows directly. A company can run only one operating model, the way a car can have only one steering wheel. If we simply bolt AI onto the existing functional hierarchy, we end up running two models at once - one for people, one for machines - and the result is fragility, not advantage. The task is not to add AI to the old structure. It is to design one structure in which people and machines work as a single team of teams.

The questions worth asking

If I were sitting in a strategy session today, I would not spend the time choosing between models and vendors. I would put a harder set of questions on the table:

1. What happens to our structure when AI coordinates work across our silos?
2. If coordination is no longer expensive, do we still need departments in their current form?
3. How much of our management exists to move information rather than create value?
4. Could we organize around a handful of outcomes instead of a dozen functions?
5. What is the role of leadership when information is instantly available to everyone?

None of these are technology questions. They are questions about what the company is.

The honest part

I am not predicting that every company abandons its org chart next year. Structures are sticky, and for good reasons - control, compliance, deep expertise. The department will not vanish overnight, and it should not.

But I am increasingly convinced that the companies pulling ahead over the next decade will be the ones willing to ask the uncomfortable question most leadership teams still avoid. Not how do we use AI? - but what does our company become when the work is done by teams of people and machines, and the departments we built for an industrial age are no longer the point?

The technology behind this shift will be available to everyone. New models will come and go. The structure is the one thing no competitor can buy.

The real transformation was never going to be technological.

It is structural.

Sources

• Stanley McChrystal, Team of Teams: New Rules of Engagement for a Complex World (2015).
• McKinsey & Company, The agentic organization: Contours of the next paradigm for the AI era (2025).
• Bloor Research, via UNLEASH, on the single operating model for human and digital work (2026).
• Microsoft, 2025: The Year the Frontier Firm Is Born (2025).