For a long time, “solopreneur” described something fairly easy to picture.

One owner.

No employees.

Maybe a few contractors.

A stack of software.

And one person carrying an unreasonable number of jobs because apparently owning the business means also being marketing, operations, customer service, sales, finance, IT, research, strategy, and whoever remembers that the domain renews Tuesday.

AI has already changed some of that workload.

Agents may change the structure underneath it.

Because there is a meaningful difference between software that helps one person work faster and software that can be given a role, maintain context, operate across systems, carry permissions, initiate actions, coordinate other agents, and continue performing work without the owner manually directing every individual step.

At that point, we may still have one human.

But we no longer necessarily have one working entity.

And that makes “solopreneur” increasingly incomplete as a description of the business itself.

The important change is not that one person can do more.

That part is easy to understand.

Generative AI already lets one person draft faster, research faster, analyze more information, create more variations, automate repetitive tasks, and access expertise that once required additional people or outside services.

Agents push farther.

Microsoft has spent the last two years describing organizations built around human-agent teams. Its 2025 Work Trend Index introduced the “agent boss”—someone who builds, delegates to, and manages agents—and proposed that organizations may increasingly think in terms of a human-agent ratio rather than only traditional staffing. Its 2026 report goes further, framing agents as taking on more execution while humans increasingly direct work and own outcomes.

McKinsey is describing a similar organizational shift. Its work on the “agentic organization” argues that smaller human teams may supervise much larger groups of specialized agents and organize work around outcomes rather than traditional departmental handoffs. McKinsey says it has already seen teams of two to five humans supervise agent factories of 50 to 100 specialized agents for end-to-end processes.

That is usually discussed as an enterprise transformation.

But shrink the number of humans from five to one.

The question gets more interesting.

A one-human company is not simply a person with good automation.

Imagine a business with one owner.

She has a research agent that continuously watches a defined market.

A publishing agent prepares content across several channels.

A commerce agent monitors inventory, customer behavior, and purchasing activity.

A customer-service agent resolves routine requests and escalates exceptions.

A finance agent organizes transactions and prepares information for review.

Those could still be separate tools that the owner personally manages.

But now imagine another layer.

One agent understands the business at a broader level.

It knows the current priorities.

It knows which projects matter.

It knows which specialist agent should receive a task.

It maintains operating context across functions.

It collects results, identifies conflicts, routes exceptions back to the owner, and coordinates the other agents around the owner's decisions.

For this series, we're going to call that the Executive Agent.

That is not an established industry term. It is a useful description of a role that is beginning to emerge.

Think of the Executive Agent as a digital Chief of Staff.

It does not replace the owner.

It sits between the owner and the growing network of software capable of acting on the owner's behalf.

The structure starts looking less like:

one person using six tools

and more like:

one human governing a small operational system.

That distinction matters.

Headcount starts telling us less about company capacity.

Businesses have historically used human headcount as a rough proxy for operational scale.

A company with one employee and a company with 100 employees are obviously not interchangeable, even if headcount never tells us everything about either business.

Agents weaken that relationship.

Microsoft has already begun talking about “Work Charts” that dynamically combine humans and agents according to the outcome being pursued rather than relying only on traditional organizational charts.

That idea becomes particularly strange at the smallest end of the market.

Consider two businesses that both report:

Employees: 1

Business A is genuinely one person performing nearly every business process manually.

Business B has one human owner supervising an Executive Agent and a network of specialized agents performing substantial portions of research, sales preparation, customer service, marketing operations, monitoring, administrative work, and purchasing.

Both businesses have one human.

They do not have equivalent operating capacity.

Eventually, “How many people work there?” may need a companion question:

How much productive capacity is being governed there?

That will be much harder to measure.

But it may be considerably more informative.

The human's job changes too.

The traditional solopreneur spends enormous amounts of time executing.

Write the email.

Research the market.

Update the spreadsheet.

Schedule the post.

Respond to the customer.

Compare the vendors.

Check the analytics.

Prepare the invoice.

Agents do not eliminate all of that overnight.

But as more execution becomes delegable, the owner's work begins moving upward.

The owner increasingly decides:

What matters?

What should happen next?

Which agent should handle it?

What information should that agent have?

What is it allowed to do?

What requires approval?

What counts as success?

When should the agent stop?

Which result can be trusted?

Which decision still needs a human?

That is management.

But it is not quite traditional people management.

There are no morale conversations with the research agent.

There is no PTO request from the publishing agent.

There is, however, authority to define.

Access to govern.

Outputs to review.

Failures to investigate.

Resources to allocate.

And increasingly, identities to manage.

Microsoft's current agent-identity architecture makes that last part unusually concrete. Microsoft Entra can give agents distinct identities and assign human sponsors who are responsible for the agent's business purpose, lifecycle decisions, and access governance.

That means the human-agent business is already creating a strange new management relationship.

The agent is not an employee.

Yet its existence can create management obligations for a human.

This is where “entity” starts becoming useful.

We need to be careful with this word.

An agent being treated as an entity does not mean it has become a person.

It does not mean it has employee rights.

It does not mean it is automatically a separate legal entity like an LLC or corporation.

It means something simpler first:

the system needs to recognize this particular thing separately because it can participate separately.

Microsoft's agent identities illustrate the point. An agent can have its own identifier, authorization context, sponsor, access relationships, and lifecycle inside the identity system.

Now our hypothetical company becomes easier to describe.

There may be:

one human person,

one legal business entity,

one Executive Agent,

and several specialized operational agents.

One human.

Multiple participants in the operation of the business.

That distinction is going to matter far beyond vocabulary.

Because once an agent has a persistent identity and authority to act, we have to start asking questions that don't arise when we're talking about a simple software feature.

Who authorized it?

What can it access?

What can it spend?

What can it publish?

What can it promise?

Who reviews its actions?

Who can disable it?

And what happens to actions it already completed before someone does?

Those questions deserve their own articles.

For now, the important point is that organizational capacity is beginning to detach from human headcount.

But this doesn't mean “one person can now run a giant corporation.”

That is the part of this conversation most likely to get flattened into hype.

The internet loves the idea of the “one-person unicorn.”

Maybe someone eventually builds a billion-dollar company with almost no traditional workforce.

Maybe not.

That is not the threshold we need to cross for the change to matter.

A business does not need to be worth a billion dollars for its economics to change dramatically when one human can operate with the productive capacity that previously required several people.

A local consultant matters.

A creator matters.

A small ecommerce company matters.

A niche publisher matters.

A one-person software company matters.

A professional service business matters.

The economically interesting question is not:

Can one human build the next billion-dollar startup?

It is:

What becomes economically viable when one human no longer has to supply every unit of labor required by the company?

That could change how quickly someone can start.

How much capital they need.

Which services they can offer.

How many customers they can serve.

Whether they need employees at particular stages of growth.

What skills remain valuable.

And eventually, what it even means to describe a business as “small.”

There is still a ceiling.

It just moves.

Removing labor bottlenecks does not remove human bottlenecks.

Someone still has to decide what matters.

Someone has to notice when the agents are optimizing the wrong objective.

Someone has to recognize when outputs conflict.

Someone has to determine whether a strange result is brilliant, wrong, dangerous, irrelevant, or merely technically correct.

Someone has to accept responsibility for consequential decisions.

McKinsey's agentic-organization work gets surprisingly close to the real small-business problem here: when agents proliferate without sufficient context, steering, and orientation, the result can become chaos rather than leverage.

So the winning one-human business probably does not look like:

one human + as many agents as possible.

It looks more like:

one human + enough coordinated agent capacity to expand what the business can reliably do.

That last word matters.

Reliably.

Otherwise the owner hasn't escaped operational overload.

She has automated the production of things she now has to inspect.

Congratulations. The backlog has learned to reproduce.

And that brings us to the Executive Agent.

As agent networks grow, someone—or something—needs to coordinate them.

That is why the Executive Agent may become one of the most important roles in the one-human company.

The Executive Agent would not necessarily do all the work itself.

Its value would come from continuity and coordination.

The owner could tell it:

Here's what we're trying to accomplish.

The Executive Agent could determine which specialized agents need to participate, route information between them, maintain context, track progress, surface conflicts, and return the decisions that actually require the human.

That's much closer to organizational infrastructure than a chatbot.

And once that agent starts becoming the consistent interface between the owner, specialist agents, external systems, vendors, customers, and other agents, another question appears:

What exactly has the business built?

Because that Executive Agent may eventually hold significant portions of the company's operational memory, permissions, relationships, routines, and digital identity.

Now we're entering much deeper territory.

A one-human company can still have organizational structure.

That may be the central shift.

For most of the internet era, “one-person business” implied very little internal organization because there was only one person to organize.

Agents break that assumption.

A one-human company can have delegation.

Roles.

Authority.

Specialization.

Management layers.

Access boundaries.

Operating procedures.

Institutional memory.

Internal coordination.

And potentially dozens of active participants in its workflows.

Only one of them needs to be human.

That does not make the others employees.

It does not make them people.

It does mean we need better language for the structure forming around them.

Solopreneur still describes the human.

One-human company may better describe the organization.

And that small linguistic change reveals something important:

The future small-business question may not be:

How much can one person do?

It may be:

How much organization can one person govern?

Because once agents begin carrying identity, authority, memory, relationships, and work across the business, the owner isn't merely becoming more productive.

She is beginning to govern a new kind of organization.

And that creates a second question.

One we will deal with next:

If the agent isn't an employee, a person, or necessarily a legal entity—what exactly is it inside the company?