Hierarchy was the old answer to complexity.

Every org chart you have ever seen is a solution to a single problem: no one person can hold the whole company in their head. So we broke the work into boxes, stacked the boxes into layers, and pointed every box at the one above it. Hierarchy was not an accident of ego or tradition. It was the best tool we had for managing complexity that outgrew any single mind.

For a hundred years it worked. And now, quietly, the problem it was built to solve is changing shape.

We are starting to put machines inside these companies. Not just tools that people use, but systems that reconcile the books, watch the pipeline, flag the anomaly, draft the analysis, and increasingly make or shape decisions that used to sit in one of those boxes. And here is the thing almost everyone gets wrong about that shift.

The mismatch nobody budgeted for

History has run this play before. Early factories bolted powerful new machines onto a management system built for manual workshops, and it took decades to realize the structure, not the machinery, was the bottleneck. Railways spread across countries before anyone had designed a way to run them safely at scale. In every case the technology arrived first and the organization lagged behind it, and the lag was where the damage happened.

We are living through the same gap right now. The technology is ready. Most organizations are not. We are running a new kind of intelligence on a management structure designed for hierarchy, silos, and one-thing-after-another execution. That structural mismatch is exactly where the value leaks out.

Which is why most of these efforts stall. They do not fail because the technology is immature. They fail because a company tries to run a genuinely new operating model on top of an authority structure that was built for a slower, more human world, and never asks whether that structure still fits.

The real work is invisible

When people picture AI transformation, they picture the visible things. Productivity. Automation. Tools. Faster output.

But the part that actually determines whether any of it holds together is invisible. It is the fabric underneath the org chart: trust, coordination, authority, how decisions get made, who is accountable when a machine is in the loop, what still counts as legitimate. Those were always the hard parts of running a company. Introduce systems that think, and they do not get easier. They get harder, and they get more important.

Consider one narrow example from my own world. A model flags that margins are slipping in a product line and recommends a price change. Who owns that call now? The analyst who used to build the number by hand and no longer does? The operator who trusts the model but cannot see inside it? The CFO who has to defend it to a board? The math got easier. The question of authority got harder. No tool answers that for you. Only leadership does.

Two structures, one transition

The old structure was built to move information up a chain and decisions back down it. The emerging one has to do something different: hold humans and machines in the same workflow without losing the things that make a company trustworthy.

Why this needs leaders more, not less

It is tempting to read all of this as a story about humans mattering less. I read it the opposite way.

When machines absorb the mechanical work, the only things left on the human side of the line are the hardest things: judgment, trust, direction, and the authority to decide what a number means and what to do about it. Those do not automate. They get more valuable as everything around them gets automated. The more capable the machines become, the more the whole thing depends on the quality of the person steering it.

An organization that buys the tools and skips the leadership does not get a smarter company. It gets a faster version of its old confusion, now with a machine in the middle of it that no one is quite accountable for. That is the failure mode I watch companies walk into. It is not a technology failure. It is a leadership vacuum with good software in it.

What this means for finance specifically

I spend my time in one corner of this transition, and it happens to be the corner where the stakes are clearest.

Finance is where a company's judgment gets encoded into numbers. When you wire machines into it, you are not just automating a close or a report. You are deciding who gets to define the numbers, who is trusted to act on them, and who answers for them when they are wrong. Get the tools right and the leadership wrong, and you have built a very fast way to make confident mistakes.

So when I take on a company's finance function, I am not really selling automation. I am doing the part the software cannot do: setting the definitions, drawing the lines of accountability, deciding what the machine is allowed to decide, and standing behind the result. The build matters. But the judgment about how humans and machines share the work is the actual job.

The bottom line

Every major technology has forced organizations to reinvent how they hold together, and the reinvention always lagged the machine. We are in that lag right now. The companies that come through it well will not be the ones with the best tools, because everyone will have those. They will be the ones whose leaders understood that wiring intelligence into a company is a question of trust and authority long before it is a question of software.

Hierarchy was our old answer to complexity. It served us well, and it is not enough anymore. The next answer will not come out of a box. It has to be led.

This essay was sparked by Florian Bankoley's TEDxBerlin talk, Why AI forces us to rethink leadership and organizations. The framing that hierarchy was our old answer to complexity is his. The finance lens, and where it leads, is mine.