Organizational Design for the Age of AI

I’ve been talking about the Never Normal for quite some time now, this idea that we are no longer navigating a temporary storm of volatility and disruption, but that they have become the climate itself. And the more I think about it, the more I’m convinced that one of the biggest mistakes companies are making is trying to survive this new climate with organizational structures that were designed for a completely different world. We talk endlessly about digital transformation, artificial intelligence, exponential technologies and the future of work, but strangely enough do we rarely question the design of the organization itself.

This is exactly why I was so excited to talk to Clay Parker Jones. His job consists of figuring out why organizations work, why they get stuck, and how to redesign them so they can actually thrive. He’s also the author of Hidden Patterns: A Playbook for More Human Workplaces, a fascinating book containing 75 interconnected organizational design patterns.

The central premise of his book is wonderfully simple. Many of the problems we attribute to people are actually created by the invisible systems surrounding those people: how decisions are made, how information flows, how teams are structured, how power is distributed or how meetings work. Clay doesn’t propose one perfect organizational model, because there obviously isn’t one. Instead, he approaches organizational design almost like a cookbook, a collection of 75 different patterns you can combine and adapt depending on what your organization needs.

Organizational knowledge as infrastructure

That idea becomes especially interesting now that AI is not only going to change what companies produce, but how they are organized. Leaders who understand organizational knowledge as infrastructure rather than overhead will have a profound advantage in the years ahead.

One of the most fascinating organizations I’ve encountered in this respect is Gojo, the company behind Purell. When I visited them, they explained that they essentially didn’t have any traditional roles. I remember thinking that this was incredibly strange, because I couldn’t really figure out who was in charge during our meetings. What fascinated me the most in this, though, is that Gojo wasn’t some idealistic startup experimenting with the management philosophy du jour. This was a company that has been around for a very long time, had repeatedly reinvented itself and had created extraordinary products.

The three types of org charts

Clay gave me a wonderful framework for understanding why organizations like this initially feel so alien. He says that every company has several organizational charts. There is the explicit org chart, the one somebody drew in PowerPoint: the official hierarchy with its boxes and reporting lines. Then there is the tacit chart, which is how the work actually gets done. Anyone who has spent more than six months inside a large company knows exactly what that means. The official process might tell you to go through one person, but everybody knows that if you actually want the decision made you need to talk to Janet, and if you want to understand what is really happening you need to call Jim. And then there is a third: the ideal organizational chart, the structure the company actually needs in order to achieve what it is trying to achieve.

Organizational design becomes really interesting when we start looking at those three things together and ask how we can bring the explicit, tacit and ideal organizations closer to one another.

That was exactly what fascinated me about Gojo. For them, this wasn't an ideological experiment in eliminating hierarchy; it was their secret sauce for innovation. They understood that if every idea had to navigate a traditional hierarchical structure, gather approvals, survive committees and slowly work its way upward before anything meaningful could happen, innovation would simply take too long. Instead, somebody with an idea could essentially start a movement. If enough people believed in it, the idea gathered people, momentum and ultimately the energy to become real.

From Hierarchy to Intelligence

I could not help but connect this back to General Stanley McChrystal's Team of Teams, born out of a brutally different context: hunting an adaptive, networked insurgency in Afghanistan with a rigidly hierarchical military machine that simply could not move fast enough. McChrystal's realization was essentially that “we had to become a network in order to fight a network.” He understood that the very characteristics that had made the traditional organization powerful (like control, predictability, centralized decision-making) were becoming liabilities in an environment where speed and adaptability mattered more.

Clay connected this to the recent argument from Jack Dorsey and Roelof Botha about moving “From Hierarchy to Intelligence,” and I think this is where the conversation becomes particularly relevant to the AI era. Their argument is that traditional corporate hierarchies exist partly because organizations historically needed layers of people to route information and coordinate activity at scale. Information moves upward, gets condensed and interpreted, decisions are made, and instructions move downward. We call that management, but viewed from another angle, a significant part of management has always been an information-routing architecture.

AI potentially changes the economics of that architecture. If an organization can build a continuously updated intelligence layer, a kind of organizational “world model” that understands the company, its customers, its capabilities and its context, then information no longer has to travel through the same human hierarchy before it can become useful. People closer to the problem can potentially get the context they need directly, and that means the really disruptive AI question isn't "how we can make everybody 20 or 30 percent more productive". It’s "why the company is still shaped the way it is".

I think we are still massively underestimating that question because the first wave of generative AI has conditioned us to think about tasks: “Write this email, summarize this meeting, analyze this spreadsheet, generate this code, prepare this presentation”. These are all useful but organizationally rather conservative. We are taking a technology that could transform the architecture of coordination, yet using it to make the existing architecture slightly faster. It reminds me of the early internet, when companies proudly announced that they had digitized their brochure and put it online.

Rebuilding organizations

The much bigger opportunity with AI may be to make the organization itself "legible", as Clay pointed out. Think about how much of the intelligence inside a company currently exists as tacit knowledge. Why did we make this decision three years ago? Which team tried something similar before? Why does this process exist? Enormous amounts of organizational intelligence live in people's heads, inboxes, meetings, documents and relationships. If AI allows us to turn more of that into accessible infrastructure, then suddenly more people can operate with more context, teams can coordinate without constantly pushing information up and down the management ladder, and management itself can become something performed by many people rather than something exclusively performed by the people who happen to occupy management positions.

That doesn't automatically mean eliminating managers, and it certainly doesn't mean blindly installing holacracy, which Clay rightly points out often doesn't work. It means separating the work of management from the job title of manager and asking which parts of coordination, context and decision-making can be distributed differently.

Relational wealth

That brings us to another concept of Clay I found incredibly compelling: relational wealth and its connection to role clarity. We have created workplaces that are hyperconnected and yet can be relationally incredibly poor. Imagine being a director in a technology company surrounded by twenty other directors. You know what they're working on, they know what you're working on, you sit in the same meetings, you're connected on Slack, you see one another at leadership off-sites, and yet somewhere underneath everyone is quietly competing for the next position up.

So instead of true collaboration, you spend an extraordinary amount of energy negotiating boundaries and trying to achieve role clarity. Is this mine or yours? Who owns this? Who gets the budget? Who gets the visibility? As Clay puts it, “people end up negotiating their little slice of the pie”. And yet the opposite is much more powerful: that magical feeling of being part of a genuinely great team where people stop obsessing about the boundaries and simply get down to business. There is flow, there is trust, there is joy, and there is a sense that everyone is playing for something larger than their individual territory.

Can you see your strategy in your organizational design?

There is one final point from Clay that I think every leader should pin to the wall. Organizational design isn't something done every three years by HR consultants during a restructuring. If you are a VP running a business line, you are doing organizational design every single day. Every time you decide who attends a meeting, who owns a decision, which team gets a problem, how information is shared, where a budget sits or who has permission to act, you are designing the organization.

Clay's sharpest test for whether your design is actually working borrows from strategist Roger Martin: you should be able to trace a visible line from the choices that define how your business wins directly into the shape of your organization. If your strategy says “speed is essential” but every meaningful decision requires seven approvals, your organization contradicts your strategy. If you claim innovation is your competitive advantage but new ideas need executive sponsorship before they can breathe, your organization contradicts your strategy. If you say customer intimacy is crucial but the people closest to customers have the least authority, your organization contradicts your strategy.

You should be able to look at the organization itself and see the choices you've made about where you are going to play and how you intend to win. And if you can't, Clay's advice is refreshingly simple: just start over.

No magical framework

Maybe that is the biggest organizational lesson of the Never Normal. There is no final operating model, no magical framework that we can install in 2026 and admire until 2036, because that assumes the environment will politely remain stable while the organization catches up. Instead, we need to start thinking about organizational design much more like software: something we observe, design, deploy, learn from and continuously edit. That is also why I like Clay's idea of 75 patterns rather than one grand organizational theory.

Gojo doesn't need to become Haier, Bayer doesn't need to become Gojo, and your organization doesn't need to become any of them. The objective isn't to look progressive, eliminate hierarchy or become a network because networks happen to be fashionable. The objective is to build precisely the organization your strategy and your environment require, and then retain the ability to change it when those conditions change.

Which is why I believe the next chapter of AI will be much more interesting than the productivity conversation currently dominating boardrooms. We keep asking what AI will do to jobs, which tasks it will automate, which professions it will augment and how many hours it can save us. Those are important questions, but they're still questions about putting a new technology into the old machine. The much more fascinating possibility is that AI changes the machine itself. In the Never Normal, that may turn out to be one of the most important competitive advantages of all.