For decades, building a large company required something obvious: a lot of people. Growth meant hiring salespeople, accountants, analysts, marketers, assistants and managers. Artificial intelligence is beginning to loosen that relationship, creating an intriguing new possibility: companies capable of generating millions in revenue with remarkably few human employees.
The internet made distribution extraordinarily cheap. AI may be doing something potentially more consequential: reducing the cost of performing certain kinds of labor and accessing expertise.
That distinction matters. The first wave of enterprise technology largely gave employees better tools. Microsoft Excel did not replace the finance department; it made the finance department more productive. Salesforce helped salespeople sell. Photoshop helped designers design. A growing class of AI companies is pursuing a different proposition: instead of selling software that helps someone perform a job, they are increasingly selling the completed work itself.
Venture capital firm Sequoia Capital has described this emerging model as “services as software” and, more recently, “services: the new software”—businesses that use AI to deliver outcomes rather than simply tools for workers. The opportunity extends far beyond making conventional software smarter. AI allows entrepreneurs to attack portions of the enormous professional-services economy that historically depended on expensive human labor.
Consider Last Accounting Company, a Y Combinator-backed startup founded in 2026. Rather than simply selling AI software to accountants, it operates as an accounting firm, providing services such as bookkeeping and tax filings. Underneath, AI agents perform much of the work while a certified accountant provides human oversight. Y Combinator currently lists the company as having just four employees.
Then there is Rational, a two-person Y Combinator startup building AI coworkers for accounting firms. Its agents are designed to take on work ranging from client follow-ups and information gathering to month-end close and tax preparation, allowing accountants to oversee the process rather than manually execute every step.
The same model is moving into other industries. RealPact, another tiny Y Combinator-backed startup, builds AI agents for real estate brokerages. Its technology finds property records, fills out contracts, organizes documents and tracks transaction deadlines, taking on operational work that traditionally required human coordinators. Y Combinator currently lists RealPact as a three-person company.
This points toward one of AI’s most fascinating entrepreneurial possibilities: the tiny-team corporation.
Imagine a sophisticated founder supported by AI systems capable of conducting research, prospecting for customers, writing marketing materials, analyzing financial data, answering routine customer inquiries, producing designs and managing administrative workflows. Many of these tasks can already be assisted or partially automated by AI, although reliability varies and human review remains essential for consequential work. Instead of automatically hiring an employee for every function, the entrepreneur increasingly becomes the conductor of a digital workforce.
The economics could be extraordinary. Businesses that once required large teams may eventually operate with a fraction of the headcount. Companies that might once have needed ten people could conceivably be built with only a handful. Entirely new services may become viable as AI allows portions of expert workflows that were once expensive to deliver to be performed at dramatically lower marginal cost.
There is an important caveat. Human judgment, accountability, relationships and trust remain particularly valuable, especially in high-stakes professions. AI systems also remain imperfect and can require significant supervision. The winning model may therefore be less about eliminating people than radically increasing the number of customers each talented person can serve.
That distinction may ultimately be more important than the provocative idea of a one-employee company. The immediate transformation is not necessarily the disappearance of employees. It is the possibility that extraordinarily small teams can achieve what once required entire departments.
For entrepreneurs, that changes the essential question. The opportunity is no longer simply, What software can AI improve?
It is: What valuable work are people paying humans to do today that a small number of exceptional humans, supported by AI, could deliver tomorrow?
Perhaps the $10 million one-employee company has yet to arrive. But the underlying economics that could make it possible are beginning to take shape.
And the companies being built today suggest that the next generation of great businesses may be measured not only by how large they become, but by how remarkably small they can remain.