The One-Person Company Isn't Coming. It's Here.

Behind Coverdale One, the venture of mine that says the least about itself, sits a full operations core. Access control, permission tiers, review discipline, and the sort of internal tooling you'd expect from a firm many times its size. It's built the way you'd build for twenty people.

Nobody else has a login. I built an institution and I'm currently its only inhabitant. I know how that sounds. Let me argue that it's the most rational thing I've done.

Why operations came before people

The old sequence was simple enough. Get traction, hire people, then build structure, because people without structure is chaos. AI reorders that. The labour arrives first now: research, build, review, admin, all available on demand. And it turns out labour without structure is also chaos. An AI with broad access and no permission model isn't an employee. It's an incident waiting for a timestamp.

So all the familiar corporate machinery stopped being ceremony and became an actual safety system for a company of one person holding a lot of leverage. I don't have staff. I have staff infrastructure, and that infrastructure is what lets the leverage run while I'm not watching.

Structure used to be what you built for employees. Now it's what you build instead of them.

What "AI staff" means in practice

It doesn't mean one chat window doing everything. The useful shape is roles. A researcher that goes deep, a builder, and a reviewer whose entire job is to attack whatever the builder made. On harder problems Arthur runs several specialists at once, and they come back with genuinely different answers because they came at the thing from different angles. My job is chairing the meeting.

If that sounds like management, it is. The org chart didn't die. It changed species.

The honesty problem, and the founding-client answer

One-person companies have a marketing problem. No client list, no testimonials, no wall of logos. The industry-standard answer is to fake it politely. With Sunshower Studios, my web studio, I decided not to. No invented praise, no stock-photo "team" page. Instead the studio's own site is the pitch, and the first founding clients get founders' pricing in exchange for the case study. An empty portfolio, priced honestly, turned into an offer.

I don't yet know whether that beats the fabricated version commercially. I do know I can put my name on it. For a company that's literally my surname, that isn't a nicety. It's the asset.

What still needs a human

• Accountability. Someone has to be the person it happened to. An AI can't be sorry, or sued, or proud.
• Relationships. People fund, hire, refer and forgive other people. Not stacks.
• Taste. Somebody has to want something specific. Averaged desire produces averaged companies.
• The signature. That final pass which makes the work yours instead of merely plausible.

For whatever a founder's forecast is worth, here's mine. Within a few years, companies that look institutional and are quietly one person will be completely unremarkable. The tell won't be size. It'll be coherence, one sensibility running through everything, undiluted by committee. Some of the best products you use will have a headcount of one and a staff of dozens.

Mine already does. It's quiet about it. That's rule one.