From DAO to NAO

How combining human autonomy with AI agency is creating the next evolution of decentralized work

I'm launching a community entirely focused on networked agentic organizations. How they work, real examples, the challenges, what needs solving, and who's tackling it. I'm active on a new Discord server with a Discourse hub coming soon. Buy me a coffee with your Substack subscription for an exclusive invite!

---

Management is the stabilization of entropy. If nobody does it, the system breaks down. You cannot just wish it away. Someone's got to keep the system together. Why not give the work to agents?

Call this semantic nitpicking if you want—you might be right. But language shapes reality, especially in organizations. The difference between "Human Resources Department" and "People Support Team" might seem trivial until you're the one dealing with either. Same energy, different universe.

Which brings us to Decentralized Autonomous Organizations (DAOs) versus Networked Agentic Organizations (NAOs). The distinction matters—to me. Let me tell you why.

"A decentralized autonomous organization (DAO) is an organization managed in whole or in part by decentralized computer programs, with voting and finances handled through a decentralized ledger technology like a blockchain." - Wikipedia

Decentralized Beats Distributed

First, let's kill a common confusion: decentralized and distributed aren't the same thing, despite how carelessly people—me included—have been tossing them around. The difference isn't academic—it's the difference between actually changing how power works versus just shuffling the furniture.

Decentralized means spreading authority and decision-making power across multiple people, teams, or nodes instead of hoarding it at the top. No single chokepoint controls everything. Authority gets pushed to where the actual work happens, letting people closest to problems solve them without playing telephone through six layers of management.

Distributed just means spreading stuff—teams, processes, resources—across different locations. You can distribute all you want and still have some VP in a corner office making every meaningful decision. Distribution is about logistics; decentralization is about power.

You can be distributed but centralized (global teams, single decision-maker). You can be decentralized but not distributed (everyone in one building, but teams run themselves). The sweet spot is both, but if you had to pick one, decentralization wins. Power matters more than location.

Networked Trumps Decentralized

Even though I appreciate the term decentralized (the opposite of centralized), I've always preferred networked (the opposite of hierarchical). Decentralized suggests escaping central command; networked suggests flattening the hierarchy.

Networked organizations are inherently decentralized—they distribute authority and operational autonomy across nodes rather than centralizing power. But they go further. Where decentralized tells you who makes decisions, networked tells you how those decision-makers actually work together.

Think of it this way: decentralized units can still operate as isolated workers. Networked units actively collaborate, share information, and solve problems collectively. A hub-and-spokes model is decentralized but not networked—everything still flows back to one point. A true network lets any node connect to any other node, creating resilience and adaptability that hierarchies can't match.

A concrete example: for my book Human Robot Agent, I worked with a co-author, book editor, book designer, book marketer, and a bunch of proofreaders. Everyone did their work separately, but they all reported back to me. There was decentralization to some extent, but no team, no network.

Networked organizations are always decentralized, but decentralized organizations aren't always networked. The difference is connection versus mere separation.

Agency and Autonomy

Now we get to the interesting part. Autonomy has been the golden child of organizational design—one of Daniel Pink's three motivators in his book Drive, and the rallying cry of business agility advocates. I contributed my share as well. But we're living in the age of AI agents now, and agentic is the new hottest term on the block.

Autonomy is the right and capacity to choose—freedom from external control, self-direction, the ability to decide based on your own values and reasoning. It's about controlling your decision-making process, regardless of the outcome.

Agency is the capacity to act and influence—the ability to take meaningful action, set goals, and drive change in your environment. It's about being proactive, taking initiative, making things happen. Agency isn't just about choosing; it's about acting on those choices and creating impact.

Autonomy gives you the freedom to choose. Agency gives you the power to act. You can have one without the other, but humans need both for real empowerment.

This distinction becomes crucial when we consider AI agents versus human workers. Both should have agency—the ability to take meaningful action within their domain. But only humans possess true autonomy—the deeper capacity for self-determination and values-based choice.

From DAO to NAO

As I wrote before, it is time for a new term: the Networked Agentic Organization (NAO). Not a replacement for DAOs, but an evolution—a better umbrella concept that includes DAOs and much more.

A DAO becomes a NAO when:

• Work flows through networks, not simple hub-and-spokes models;

• The network includes both human agents (with autonomy) and AI agents (without autonomy),

• and blockchain infrastructure is optional, not required

Many DAOs today are governance theater—thousands voting on proposals while maybe ten people do the actual work. Those ten people? They're the real NAO. The rest is just passive participation in a decision-making ritual.

But NAOs extend far beyond blockchain-based governance:

• Haier's thousands of networked micro-enterprises

• Gig platforms enabling networked collaboration among independent workers

• Holacracy-based organizations using agents for workflow management

• Informal networks mixing humans and AI agents

You don't need tokens or blockchain to qualify as a Networked Agentic Organization. You need to organize workflows in a decentralized, networked manner while delegating significant management tasks to agents—human, AI, or both—operating with real agency.

The Management Problem Gets Solved

We're finally approaching the solution to organizational design's biggest challenge. All those inspiring examples—Buurtzorg, Morning Star, Haier, W.L. Gore, Valve—were early experiments pointing toward networked agentic organizations. But they hit a wall.

The problem was always that autonomous humans still had to do all the management work themselves. Even with decentralized authority, someone still needs to handle governance, coordination, resource allocation, and conflict resolution.

Management is the stabilization of entropy. If nobody does it, the system breaks down. You cannot just wish it away. Someone's got to keep the system together. Why not give the work to agents?

Well, it turns out most humans aren't particularly good at management tasks, which is why progress in networked organizations stalled for decades. But Google's Sergey Brin nailed it: "Management is the easiest thing to do with AI." (Note: he said management, not leadership.) Now, there's an actual way out of the mess!

What happens when AI agents handle the tedious management work while humans focus on what they do best—creative problem-solving, relationship-building, strategic thinking, and meaningful work that requires autonomy?

We will soon get truly networked agentic organizations. Networks of human and AI agents collaborating fluidly, with management overhead handled by systems designed for that purpose, leaving humans free to exercise both agency and autonomy in domains where it matters.

This isn't about replacing humans with machines. It's about finally building organizations where technology handles coordination and administration while humans do the work that actually requires human judgment, creativity, and care.

The future of work depends on Networked Agentic Organizations (NAOs) where humans and AI agents collaborate in structures that leverage the best capabilities of both.

---

I'm launching a community entirely focused on networked agentic organizations. How they work, real examples, the challenges, what needs solving, and who's tackling it. I'm active on a new Discord server with a Discourse hub coming soon. Buy me a coffee with your Substack subscription for an exclusive invite.