AI-Native Companies
An AI-Native company is one designed so that a workforce of Digital FTEs — not headcount — is the unit of production, with evals for quality and humans accountable at the Edge Layer.
Read together, these two chapters do more than explain the AI-Native company — they prepare you to act on it. The worker-level chapters taught you to build Digital FTEs; this section gives you the company-level frame to put that skill to work three ways: founding your own AI-Native startup, advising an existing company on becoming one, or taking on freelance work building AI-Native companies and the Digital FTEs they run on. Read it and you can sit across from a business, diagnose why its AI adoption produced no advantage, and lay out what becoming AI-Native would actually take. That is the difference between an implementer and an advisor.
The AI Worker Catalog is the worker-level view: which Digital FTEs exist and how to build them. This section is the company-level view: how a whole company becomes — or is built as — one.
Both pieces here rest on the architecture in The Agent Factory Thesis. If you have not read it, start there. The Thesis defines the AI-Native company; these two chapters are the two ways you arrive at one.
The two paths in
| Chapter | The question it answers | Read it when |
|---|---|---|
| The AI-Native Transformation | How does an existing company go from using AI to being built out of it? | You run, or advise, a company that has already adopted AI but seen no advantage from it. |
| Founding an AI-Native Startup | How do you build an AI-Native company from nothing, around a workforce of Digital FTEs? | You are starting from zero and want the company to be AI-Native from day one. |
How they fit together
The two chapters reach the same destination from opposite ends. The AI-Native Transformation is the incumbent's path: it starts from a working company and rebuilds it, and it reads naturally as the motivation behind the Thesis. Founding an AI-Native Startup is the founder's path: it assumes you already know the machinery (the Agent Factory, Digital FTEs, the Two-Layer Model, evals, the system of record) and shows how to assemble a company out of it. It is the more advanced read, closer to a capstone than a starting point.
What unites them is a single discipline: a company written down precisely enough that an AI workforce can run on it, judged by evals, and held accountable by humans at the Edge Layer. The transformation chapter argues why that discipline matters; the founding chapter shows what it looks like when a company is born with it.