From Builders to Orchestrators

Earlier we established that information is no longer scarce. This essay explores the next constraint to collapse: creation itself.

To understand where institutions are going, look at what was scarce in previous eras. The architecture of the industry — and of careers — has always been defined by the Core Constraint.

In the Mainframe Era, the constraint was Access. Hardware was the bottleneck, and the customer was held captive by the physical machine. In the SaaS Era, the constraint was Engineering Capacity. It took years to build a platform and a GTM motion. That head start was the moat.

Now creation is becoming abundant. And, the constraint is shifting again.

To build anything meaningful, you needed armies of developers to build and armies of Product Managers to align everyone on what was being built. Progress was a function of Scale — how many people an institution could hire and coordinate. Every organizational form was a rational response to the scarcity of its era.

From Scale to Agency

The marginal cost of creation is collapsing toward zero. Features can be prototyped in hours. Interfaces can be spun up in a weekend. What once required a coordinated team can increasingly fit inside a single person's context window.

This is not simply about smaller teams, but a different source of leverage.

The old question was: How many people can we coordinate?

The new question is: How much leverage can each person create?

We are moving from an era of Scale to an era of Agency.

The New Firm

This forces a rethinking of the institution itself. The middle layer — the mediators who translated intent into tickets, moved information between functions, and kept the factory moving — is under pressure.

What replaces it is not just simply fewer people but a different division of work. The new firm organizes around three roles: The Builders. The Orchestrators. The Relationship Owners.

1. The Builders

For the last twenty years, software development was largely a craft and every line of code was written by a human. In the AI era, software development is increasingly becoming a production system. The Builder no longer manually creates every component. AI does more of the generation and The Builder designs the environment in which that generation can happen reliably.

Think of a modern auto factory where Humans are not on the line holding screwdrivers. They design the jigs, define the quality standards, monitor error rates, and supervise the robots. The Builder's role shifts from Writer to Supervisor.

The old metric: Velocity / Story Points
The new metric: Throughput and Quality / Evals

Builders define the acceptance criteria that tell the system whether the output is good enough. They build the architecture that allows the production system to run. They define the operating conditions and the boundaries within which it can act. They don't just build the product — they build the system that builds the product.

2. The Orchestrators

If the Builders run the production system, who decides what to build? This is where the Orchestrator comes in. In the AI era, the raw ingredients — models, compute, infrastructure, code generation — are increasingly accessible. Everyone has access to the same commodities and The differentiator therefore becomes taste.

The Orchestrator decides what deserves to exist when almost anything can be built. In the SaaS era, we separated Product — who defined the why and what — from Engineering — who defined the how. The Orchestrator reunites these decisions around the outcome — they sit close to the customer, they understand the problem, they configure the production system — and, most importantly, they decide what is worth building and what is not. They are the first and last mile between the customer and the production system.

The Orchestrator:

• Owns the outcome. They define the metrics that actually matter.
• Understands the customer. They sit close enough to understand the specific context and nuance of the market.
• Configures the system. They turn abundant production capability into something useful.
• Exercises taste. They can tell the difference between generic output and something worth acting on.

If you cannot tell the difference between generic AI output and high-value insight, you are not orchestrating.

3. The Relationship Owners

Finally, there is the Relationship Owner. In a world of abundant AI-generated output, trust becomes scarce. The Orchestrator ensures that the solution works. The Relationship Owner ensures that the customer believes in it.

They front-end demand, they preserve legitimacy, and they hold accountability when the system produces something that cannot be explained. They provide the human handshake that validates the transaction. Their role is not jut simply sales but represent the institutional trust.

They are the human expression of the Trust Budget — the people who hold the relationship when the institution cannot afford the cost of being wrong. Metric: Trust held and liability owned.

Market Structure

The same reorganization around scarcity happens across markets. When the cost of creation falls, differentiation shifts to different layer — Infrastructure consolidates and Applications fragment.

At the bottom, wholesale consolidation: the Flour Mills. At the top, fragmentation: the Bakeries. The factory becomes cheap to setup and anyone can build. The value moves from owning the means of production to knowing what to produce and for whom.

But abundance creates a second problem: when anything can be built, there will be more things to choose from.

The Agent Network

In a world of thousands of niche applications, how do you find the right one? You are not going to download ten thousand apps or visit ten thousand websites. The friction is too high.

The SaaS model of manually using software — browsing tabs, clicking buttons, filling forms — starts to break down as the number of specialized systems increases.

The application layer can fragment faster than humans can navigate it.

This is where the interface shifts. The next model looks less like a collection of applications and more like a network.

Think about your credit card. You don't sign a separate contract with every merchant. You carry one identity, and it interoperates across millions of merchants through a common set of rails.

The Agentic Era could work on similar principles.

The roles become clear:

Builders create the infrastructure that allows agents to interact.

Orchestrators create specialized agents that understand a particular market or workflow.

Relationship Owners establish the trust that allows the transaction to happen.

The interaction moves from the explicit software era to the implicit agentic era.

You don't manually navigate every system. Your agent does the connecting.

This is the architectural leap. We are leaving an era where every interaction required a manual, low-bandwidth input and entering one where software increasingly does the connecting.

The human moves further upstream — From operating the software to orchestrating what the software should accomplish.

The Organizational Transition

For decades, organizations were designed around the cost of creation — Engineering was scarce, Coordination was expensive, Scale determined advantage. That constraint is disappearing.

Creation is becoming abundant. Judgment is becoming scarce.

Organizations that continue scaling the old structure will simply produce more of what has become cheap. Organizations that redesign themselves around judgment will produce what remains valuable.

The future firm will not be defined by how many people it coordinates. It will be defined by how much leverage each person creates. That is the organizational transition of the AI era. And it changes the question at the centre of the firm.

When creation was scarce, the advantage was the ability to create. When creation becomes abundant, the advantage moves to deciding what deserves to be created. Once software itself becomes abundant, another question follows.

If everyone can create, where does value accrue? That is the question that belongs to the next essay.