The advice every drowning founder hears is the same: you need to hire someone. It feels true around month six, when you are answering support emails at 11pm, your roadmap has not moved in two weeks, and you cannot remember the last time you looked at your own numbers. A person would fix this. Obviously.
It usually wouldn’t. A first hire when you have no operating rhythm just gives the chaos a second host. You spend your scarce hours managing instead of building, and the things that were falling through the cracks keep falling, now with a salary attached. The fix you actually need is cheaper and less satisfying: build the chairs before you fill the seats.
You already hold five jobs
Every small company, even one with a single person in it, has roughly five jobs that have to get done. Call them the five seats:
Builder. The actual work. The product, the code, the thing customers pay for.
Seller. Pipeline and distribution. Getting the thing in front of people.
Keeper. Support and retention. Keeping the customers you already won.
Counter. Money, runway, the boring admin that quietly compounds.
Captain. Deciding. Prioritizing. Saying no to the good idea so the right one ships.
Here is what goes wrong. Most solo founders spend ninety percent of their time in the Builder seat because it is the one that feels like progress and the one they are best at. The other four seats sit empty until something breaks. A churned customer. A surprise tax bill. A competitor who out-distributed you while you were polishing a feature nobody asked for.
You are not failing at those jobs because you are bad at them. You are failing because you can only physically sit in one chair at a time, and you keep choosing the comfortable one.
A system is just a chair that holds the seat when you’re not in it
This is the whole idea. A system is not a Notion template or a fancy tool. It is whatever lets a job get done to an acceptable standard while your attention is somewhere else.
When Facebook bought Instagram in 2012 for about a billion dollars, the company had thirteen employees and tens of millions of users. Thirteen. A service that size should have needed a building full of people, and instead it ran on a team you could fit around one table. They were not superhuman. They kept the product brutally narrow, basically one thing done well, so there was less to operate in the first place. They ran on cloud infrastructure that absorbed the growth automatically instead of hiring an ops army to babysit servers. The whole operation was designed so the work did not balloon every time the user count did.
That is the lesson, just scaled down to you. Instagram filled the five seats with a dozen people, not because they outworked everyone, but because they built chairs the work could sit in. A narrow product was the chair under the Builder seat. The infrastructure was the chair under the Counter and Keeper seats. Nobody was holding the whole thing up by hand.
If that math works at thirteen people serving tens of millions, it works at one person serving five hundred. The principle does not care about your size. The bar is the same one Instagram cleared at a hundred times the scale. They defined handoffs and an operation that does not break the moment you step away from it.
Where AI changed the math, and where it didn’t
Here is why this is worth writing about now and not in 2021. For most of startup history, the only way to fill an empty seat was a human. AI agents and automation changed that. A support drafting agent can sit in the Keeper chair. A scheduled workflow can sit in the Counter chair, pulling your numbers into one view every Monday. Your Seller seat can run on automated outreach and content you set up once. A staff-light company is genuinely viable in a way it was not a few years ago.
But the new tools do not remove the work. They move it. AI is fantastic at sitting in a chair you have already designed. It is useless at deciding the chair should exist, what “good enough” looks like, or when the output is quietly wrong. Bolt an agent onto a job you have never defined and you do not get leverage. You get a faster way to produce confident garbage.
So the order matters. First define the seat: what is the job, what does done look like, what is the trigger that starts it. Then automate the parts that are mechanical. The founders getting real leverage from AI right now are not the ones who adopted the most tools. They are the ones who had clean operating scaffolding for the agent to plug into.
The one seat you can never delegate, to a person or a machine, is Captain. Deciding what not to do is the job. Everything else is a chair you can build.
The actual move this week
Do not hire. Do not buy five tools. Open a doc and write the five seats down. For each one, answer two questions: what is the job, and what currently happens when I’m not doing it. Most of your answers will be “nothing, it just breaks.” Those are your chairs to build, in order of what is bleeding the most.
A team of one that operates like a team of five is not a person doing five jobs at once. It is a person who built five chairs and now only has to sit in the one that matters today.
