Can you actually run a company with no employees?

The distinction matters because "no employees" is doing two different jobs in that sentence. It can mean nobody is on payroll, which is achievable now. Or it can mean nobody is involved, which is not achievable and is not close.

What does a company with no employees actually look like day to day?

Software executes continuously and the owner reviews a queue. Overnight work happens without anyone present: leads scored, tickets answered, invoices reconciled. In the morning the owner works through the items that stopped — a discount above threshold, a refund, a reply that commits the business to something.

That review is the job. It is short, and it is not optional. The realistic figure is on the order of fifteen minutes on an ordinary morning, longer on a day with a real decision in it.

The work that fills those fifteen minutes is deliberately the work software should not do alone: things that are irreversible, things that cost money above a threshold, and things that state what the company charges or claims. Those are the decisions a business is actually made of.

What legally still requires a person?

Everything that binds the company. Contracts, tax filings, regulated activity, anything requiring a signature or a license. A company is a legal person and software is not; somebody has to be accountable for what the business does, and that is the owner regardless of how the work got done.

This is worth being precise about because platforms in this category sometimes describe themselves as running the company. Read the terms of any of them and you will find the same thing: the customer is responsible for what the system does on their behalf. That is not a loophole, it is how agency works.

Practically, it means a zero-employee company still has a person answering for its tax position, its customer contracts, and anything a regulator asks about. It also usually has an accountant, and often a lawyer — both contractors, not staff.

Where does a zero-employee company usually fall apart?

At the point where a customer needs a human decision and nobody is watching. The failure mode is not the software producing nothing, it is the software producing something wrong and confident while the owner is not looking. Review cadence is what prevents it, not better models.

The second failure mode is expectation. Somebody buys a platform advertised as running itself, discovers there is a morning queue, and concludes the product is broken when in fact the marketing was. This is the most common reason people leave this category, and it is entirely avoidable by telling the truth in advance.

The third is scope. A business with no staff is very good at work that repeats and very bad at work that is different every time. If the company’s core value is a relationship, a negotiation, or a judgment call that nobody has written down, that part is not going to be covered, and no amount of tooling changes it.

Do you still need an accountant or a lawyer?

Almost certainly, and neither is an employee. Contractors are how a zero-employee company gets specialist judgment without payroll, and the good ones cost less than the mistakes they prevent. The structure is about not hiring staff, not about refusing all outside help.

An accountant is the one most founders need first, because tax treatment and company filings are exactly the kind of work where being slightly wrong accumulates quietly. A lawyer matters at specific moments: your customer terms, anything involving personal data, and the first contract a customer marks up.

There is a related trap worth naming. Founders running on agents sometimes ask the software for advice in these areas because it answers instantly and confidently. Every platform in this category disclaims exactly this in its terms, and they are right to. A confident answer about your tax position is not the same as a correct one, and the consequences land on the company rather than the software.

The honest version

You can run a company with no payroll. You cannot run one with no attention. The founders who do this well are the ones who treat the morning queue as the job rather than an inconvenience, and who hire a person the moment a specific outcome genuinely needs an owner.

newc0 builds this shape of company deliberately: an application the founder owns, agents that run four functions inside it, and an approval gate that stops everything binding, irreversible or expensive. The fifteen minutes is on the homepage, not in the small print.