Most conversations about AI in business are still stuck at the productivity layer. Faster writing. Better note-taking. Less admin. Quicker turnaround times. That’s all real, and some of it is already valuable.
But it’s only the first-order effect.
The more consequential shift isn’t that AI helps your current organisation run a bit better. It’s that AI may start to challenge whether parts of that organisation need to exist in their current form at all.
That’s a much bigger idea. And for agencies, consultancies and B2B service businesses, it’s becoming one of the most important strategic questions of the next decade.
The org chart was built for a different problem
Most management structures weren’t designed from first principles. They evolved.
A founder starts with a small team. Everyone does a bit of everything. Growth happens. Specialists are added. Client services becomes more formal. Project management appears. Department heads emerge. Reporting lines harden. Soon the business is running a surprisingly expensive internal machine just to stay coordinated.
As businesses scaled, founders lost direct visibility. Teams specialised. Different departments developed different priorities. Communication gaps widened. Decisions slowed. So companies added management to compensate — and on paper, that sounds entirely sensible.
In practice, it often creates a different set of problems.
The business gets larger, but not necessarily sharper. More people are added to drive growth, but decision-making slows. Layers emerge to improve coordination, but they create distance from the customer and from the truth of what’s actually happening on the ground.
At a certain point, the business is no longer constrained by ambition or market opportunity. It’s constrained by internal drag.
That drag rarely arrives all at once. It builds quietly.
A founder who used to make a decision in ten minutes now needs three meetings and a deck. A client issue that would once have been resolved in a phone call now travels through account management, delivery, ops and leadership before anyone owns it properly. Teams begin spending more time reporting on work than doing work.
This isn’t a people problem. It’s a structural one. And it’s a pattern we see consistently across the businesses we work with at Scaled.
A lot of middle management is really an information-routing function
That might sound harsher than intended, but it’s worth being honest about what many roles in growing businesses actually do.
In theory, management is about leadership, coaching, performance and strategic thinking. In reality, a significant proportion of management time is consumed by something more basic: gathering information, passing it on, translating it, chasing updates, clarifying priorities and helping the organisation function despite the fact that no one person can see the whole picture.
Put simply, a lot of management is coordination overhead.
Until recently, that overhead was necessary. If a company had no shared system of understanding, humans had to become the system. Managers sat between leadership and delivery — or between teams, or between client need and internal execution. They knew where things were. They knew what was blocked. They carried context.
That context-carrying role has historically been essential.
But this is precisely where AI becomes interesting. Because if a system can increasingly capture, update, structure and surface context in real time, some of the functions that justified those layers start to weaken.
The question becomes: if better systems can carry more of the informational burden, what should the people be doing instead?
That’s where the real redesign begins.
The first wave of AI improves productivity. The second wave changes structure.
Many businesses are currently treating AI as a local optimisation tool. The current model survives. People just work faster inside it.
That’s not unimportant — but it isn’t transformational either.
The bigger opportunity is structural.
Once AI starts helping a business maintain a live picture of what’s happening across clients, projects, pipeline, delivery, margin, resource allocation, risks and operational performance, it starts to reduce the need for context to be manually carried through layers of people.
That doesn’t remove the need for leadership. It doesn’t remove the need for judgement, trust or difficult calls. But it may reduce the need for large numbers of people whose role exists mainly because the company is otherwise too fragmented to coordinate itself.
This matters particularly in people-based businesses because they are often full of handoffs. Someone sells the work. Someone else scopes it. Someone else delivers it. Someone else manages the relationship. Someone else reports on performance. Someone else escalates concerns. Someone else translates what leadership wants into what the team actually does.
You can see quickly how value gets diluted across that chain.
For agencies specifically, this is one of the clearest examples we see. A large amount of effort already sits between the work and the customer. Leadership teams are often pulled back into the middle because the structure hasn’t created genuine autonomy — only more supervision. Delivery decisions are slow. Ownership is blurred. Profitability varies wildly by client, partly because nobody has clean enough visibility to intervene early.
The businesses that redesign around this reality — rather than waiting for it to become a crisis — will have a significant structural advantage.
People don’t become less important. Their value shifts.
The lazy version of this argument is that AI reduces headcount and businesses simply need fewer people. That may happen in some cases, but it isn’t the most useful framing.
The deeper shift is that people become more valuable when they’re not spending most of their time doing work a system could support more effectively.
In our view, the future organisation in many sectors will place a premium on four things.
The first is deep specialist capability — people who can genuinely build, diagnose, solve and create. Not people who sit adjacent to work and coordinate it, but people who can materially move something forward.
The second is direct ownership. Fewer passengers, more people who can hold outcomes end to end. The ability to own a problem, make decisions and drive resolution will matter more than maintaining a comfortable position inside a hierarchy.
The third is player-coach leadership. One of the biggest casualties of the old org model may be the detached manager whose primary role is supervision and coordination. The stronger businesses will favour leaders who still contribute meaningfully to the work, but who also develop others and raise standards around them.
The fourth is human judgement at the edge. AI may get very good at surfacing patterns, carrying context and recommending action. But businesses still need people who can interpret nuance, build trust, read a room, handle ambiguity and act confidently where the stakes are high and the situation is genuinely complex.
That’s where human value deepens — not disappears.
The strategic risk isn’t just inefficiency. It’s speed.
The businesses that win over the next decade are unlikely to be the ones that merely automate tasks inside bloated structures. They’ll be the ones that redesign around faster information flow, clearer ownership and better judgement.
Speed compounds. A business that sees issues earlier, allocates resources faster, learns quicker and keeps people closer to customers will consistently outperform one that remains structurally slow — even if both are using the same tools.
This is why AI and org design are now inseparable topics. The conversation isn’t just about what AI can do. It’s about what kind of company AI makes possible.
Some businesses will use AI to make the current machine slightly more efficient.
Others will use it to rethink the machine itself.
The second group will be the more interesting one to watch — and in our experience, the more valuable one to be.
The question more leadership teams need to sit with
If you were designing your business today, from scratch, in an AI-enabled world — would you really build it this way?
Which roles genuinely create value, and which exist because the organisation is too fragmented to function without them? Where does judgement matter, and where are you simply paying for human routing? How much of your structure is serving your customers, and how much is serving the internal machine?
These aren’t comfortable questions. But they’re the right ones to be asking now — before the pressure of the market forces the conversation on less favourable terms.
It’s a conversation we’re having regularly with founders and leadership teams at Scaled, and increasingly it sits at the heart of how we think about growth strategy for the businesses we work with.
Because sustainable growth isn’t just about adding resource or improving marketing. It’s about building the kind of organisation that can actually execute — cleanly, quickly and at scale.
If you’re thinking about how your structure is affecting your growth trajectory, talk to the Scaled team →
