Introduction: Your Hierarchy Is the Slowest Thing in the Room
Every executive wants to know how to make AI work faster.
Wrong question.
The real question is why your organization is so slow to begin with.
I’ve made the case, in AI on Your Org Chart and The Future of Middle Management, that decision rights are migrating from people to systems and that management is shifting from coordination to judgment.
Here’s what comes next.
If decisions no longer have to move up a hierarchy to get made, the hierarchy has no job left.
Most leaders won’t say that out loud.
But they’re already acting like it’s true. They route fewer approvals through layers that used to own them. They let AI surface answers managers used to gatekeep. They just haven’t redesigned the structure to match.
That gap is where the damage happens.
In my book Rogue Waves, I argued that organizations don’t get killed by gradual trends. They get killed by sudden shifts that overwhelm systems built for a different world.
AI is one of those shifts.
Not because it automates work.
Because it moves faster than the structure meant to contain it.
The bottleneck used to be information. Now it’s the org chart itself.
The 150-Year-Old Assumption Behind Every Org Chart
For more than 150 years, businesses have been organized around a simple premise.
Information moves slowly.
That premise shaped everything.
- Decisions flowed upward
- Approval flowed downward
- Managers became translators between the people doing the work and the executives deciding what happened next
It was an elegant solution for the Industrial Age.
Today, the premise no longer holds.
AI can synthesize information, identify patterns, and recommend actions in seconds. Markets shift overnight. Supply chains reroute in real time. Customers expect immediate responses.
Picture a regional sales dip. A store manager notices it first and flags that foot traffic is fine but conversion is down, maybe pricing or staffing is off. Under the old model, that observation climbs from the store to a regional manager, who builds a case for a promotional discount or a staffing change. The regional manager pitches it to a VP, who needs to sign off on the margin hit or the labor cost before anything happens. Each handoff waits on someone’s calendar. By the time approval comes back down, the quarter is already over.
AI can flag the same dip, model whether the cause is price, staffing, or a competitor promotion, and recommend the fix the same day. The data moved at machine speed. The decision still moved at hierarchy speed, because someone with budget authority still had to personally approve a discount before a store could act on it.
The organizations that win won’t necessarily have the smartest executives, or even the best AI.
They’ll have the fastest decision systems.
From Hierarchies to Networks: The Octopus Model
In AI and the Octopus Organization, I describe a different model for the future enterprise.An octopus doesn’t rely on a single brain directing every movement. Nearly two-thirds of its neurons sit in its arms, letting each limb sense, react, and solve problems independently while staying aligned with the organism’s overall intent.
It’s an extraordinary balance between autonomy and coordination.
Businesses increasingly need the same capability.
For decades, organizations concentrated intelligence at the top because information was scarce. Today, intelligence is abundant. AI gives people across the enterprise access to analysis, recommendations, and expertise that once required layers of management to reach them.
We’re already seeing early versions of this emerge. At Uber, AI engineers were embedded directly into finance, HR, legal, and operations instead of operating as one centralized AI team. Each group built AI agents around its own work while leadership focused on governance and strategic priorities. The results were dramatic. Financial reporting that once took two days now takes about ten minutes, while capital allocation across more than 150 cities dropped from fifteen hours to roughly thirty minutes. The organization became faster not because every decision flowed upward, but because intelligence moved closer to where the work happened.
Leadership still matters.
Perhaps more than ever.
But its job changes.
Instead of making every decision, leaders create the conditions that let thousands of good decisions happen simultaneously.
That’s distributed intelligence. Leadership sets direction. AI distributes expertise. Teams make better decisions closer to the work. Organizations become faster because they stop treating every decision like it has to climb a ladder before it can move.
AI Doesn’t Flatten Organizations. It Redistributes Intelligence.
Many executives assume AI simply removes management layers.
That’s the wrong question.
The real transformation isn’t fewer managers.
It’s a different decision architecture.
Microsoft describes this evolution as moving beyond the traditional org chart toward a “Work Chart,” where AI agents, specialists, and teams dynamically assemble around problems instead of relying solely on fixed reporting lines.
It’s a subtle shift with enormous implications.
The hierarchy doesn’t disappear. It becomes a flexible network that forms around opportunities rather than organizational boundaries.
Consider a customer complaint.
In most organizations today, that issue travels from customer service to a supervisor, then to product management, operations, engineering, and finally leadership before action happens.
In a distributed-intelligence organization, AI agents identify similar issues across thousands of customer interactions, estimate business impact, recommend corrective actions, and coordinate work across multiple teams. All of that can happen before the next executive meeting is even scheduled.
Humans still exercise judgment.
They just spend less time moving information and more time improving decisions.
Decision Velocity Is the New Competitive Advantage
For years, organizations measured productivity.
Increasingly, they need to measure decision velocity:
- How quickly can your organization recognize a problem?
- How quickly can it decide?
- How quickly can it learn whether that decision worked?
Those three measurements are becoming a primary source of competitive advantage.
McKinsey’s 2025 State of AI report found that 88% of organizations now use AI in at least one business function. Yet only about a third have begun scaling AI across the enterprise.
The gap isn’t about technology.
It’s about organizational design.
Companies are deploying AI into structures built for slower decision cycles.
Adding AI to an unchanged hierarchy is like installing a Formula One engine into a horse-drawn carriage. The engine isn’t the constraint. The carriage is.
The Companies Pulling Ahead Are Coordinating Better
Some of the strongest AI adopters aren’t simply automating work.
They’re redesigning how work happens.
Unilever trained more than 23,000 employees in AI by the end of 2024 and embedded hundreds of AI capabilities across forecasting, manufacturing, and supply chain operations. Rather than routing every decision upward, AI gives teams across the organization shared intelligence so they can make better decisions closer to where the work happens.
One Bayer factory line used to require 15 specialists, each handling a piece of the process and passing problems up the chain. It’s now run by a five-person team with full ownership from start to finish. Issues that once dragged through six months of escalation get resolved in days.
Even Amazon’s “two-pizza teams” anticipated this shift long before generative AI existed. Jeff Bezos’s rule was simple: no team should be bigger than a group two pizzas could feed, roughly 10 people or fewer, with full ownership of a single product or service rather than a slice of a larger one. Small autonomous teams like these consistently outperform large centralized organizations because they reduce communication overhead and accelerate learning. AI extends that principle by letting distributed teams share intelligence without recreating bureaucracy.
The companies pulling ahead aren’t necessarily using more AI.
They’re coordinating intelligence more effectively.
Networks scale. Approval queues don’t.
Distributed Doesn’t Mean Decentralized Chaos
Whenever I talk about distributed intelligence, someone inevitably asks whether organizations risk losing control.
The opposite is true.
Distributed intelligence requires stronger alignment than traditional hierarchies ever did.
- Everyone needs to understand the mission
- Everyone needs access to trusted information
- Everyone needs clear decision rights
I saw this play out inside the U.S. Navy. Leaders wanted to speed up innovation after watching cheap, off-the-shelf technology start to reshape modern warfare. The obstacle wasn’t funding or tools. It was culture. Sailors weren’t used to being trusted to act without sign-off, so when the Navy first put out a call for volunteers to test new ideas, only seven people showed up.
Leadership didn’t solve that by tightening control. They solved it by protecting people who experimented and failed, and by making intent clear enough that sailors could act on it without waiting for permission. Participation grew from seven volunteers into a network of hundreds, and the speed at which new technology reached the fleet jumped accordingly.
The military has a name for this discipline: Mission Command. Leaders communicate intent while frontline units make decisions based on rapidly changing conditions. The objective stays centralized. Execution becomes distributed.
That’s the model AI makes possible for business. Most companies still try to run AI through systems built for control, not coordination, which is exactly why the technology so often amplifies dysfunction instead of fixing it. The fix isn’t tighter oversight. It’s clearer intent, paired with people, and increasingly AI, who are trusted to act on it.
The center doesn’t disappear.
The future of leadership isn’t control. It’s coordination.
Leadership’s New Job
For generations, leadership meant collecting information and making the best possible decision.
That model assumed leaders held the most complete picture.
Increasingly, they don’t.
AI systems can synthesize vastly more information than any executive team. Employees closest to customers often spot change before headquarters does. Partners, suppliers, and even customers contribute valuable intelligence of their own.
Leadership is no longer about being the smartest person in the room.
It’s about designing a system where intelligence can emerge from everywhere.
Gallup continues to show that managers have an outsized impact on employee performance and engagement, accounting for as much as 70% of the variance in team engagement across business units. That won’t disappear with AI. If anything, Gallup’s CEO has pointed to managers as the missing link between AI adoption and actual business results, citing data showing most U.S. employees don’t believe their manager actively supports their team’s use of AI.But the manager’s role is evolving from supervising work to building the environment where people and intelligent systems make better decisions together.
The Next Organizational Revolution
The Industrial Revolution rewarded organizations that controlled capital.
The Information Age rewarded organizations that controlled data.
The AI Age will reward organizations that distribute intelligence faster than competitors distribute authority.
That’s not a prediction. It’s already visible in the examples in this piece.
It’s visible in a production line at Bayer that used to need 15 specialists passing work up and down a chain, and now needs five people who own it end to end. It’s visible in a Navy innovation program that went from seven volunteers to hundreds once leaders stopped trying to control every decision and started just being clear about intent. It’s visible in the simple fact that a sales dip can now be diagnosed in a day by a machine, yet still take a quarter to act on because a human hierarchy hasn’t caught up.
The pattern is the same every time. The constraint was never the technology. It was always the structure technology had to move through.
The organizations that thrive over the next decade won’t simply deploy better AI.
They’ll redesign themselves so intelligence can emerge anywhere, decisions can happen everywhere, and leadership becomes the architecture that connects it all rather than the chain every decision has to pass through.
Hierarchies were designed to move information upward.
Distributed intelligence moves capability outward.
That’s why hierarchies can’t keep up.
The Bottom Line
AI doesn’t just change what work gets automated.
It changes where intelligence lives in your organization.
The companies that win this decade won’t be the ones with the most AI.
They’ll be the ones that rebuild their organizations so intelligence, human and machine, can move at the speed decisions now require.
