The companies that thrive in the next decade won't be the ones with the most people. They'll be the ones with the best architecture for multiplying a few good people through governed AI execution. This is the case for why — and the blueprint for how.
AI labour is not oil. A barrel burned in Tokyo can't be burned in Riyadh — but an agent, once built, replicates at near-zero marginal cost and never gets used up. That breaks the old economics of work. The jobs that died in the backlog for lack of hands — the follow-up never sent, the reconciliation never done, the customer never called back — all of it becomes affordable at once.
Every company spends 60–80% of its headcount on general functions — finance, HR, compliance, support — that generate no revenue. The AI-hybrid organisation restructures that: a human nucleus holds the strategy and the signature; governed agent pools do the execution. The result is the operational capacity of a 500-person company with a core team of thirty to fifty — or, for an SME, the capacity of a fifty-person firm with three to ten.
Billion-dollar companies built by a dozen people are no longer science fiction — AI has become the operating system of venture building. The advantage no longer flows to whoever can hire the most. It flows to whoever architects a few good people multiplied by governed AI execution, and keeps a hand on every consequential action. That is a small giant. We wrote the blueprint for becoming one.
Two foundational works: the economics of AI labour, and the blueprint for the organisation built on it. Leave an email and we'll send you both.
If everyone builds AI to export, who imports? A theory of how AI-labour trade actually unfolds: non-rivalrous supply, the domestic-consumption trap, and the four roles — who builds, operates, hosts, and consumes — distributed across nations for the same unit of work.
A theory and blueprint for lean, self-evolving organisations where humans lead and agents execute. Borrows the biology of the cell — nucleus, membrane, mitochondria, organelles — to design roles, governance, and scaling, and closes with a universal prompt template for any org.
Beyond our own thesis, we keep a curated library of the best outside work on AI labour, venture building, and the economics of small teams — papers, presentations, and links we actually recommend. A taste below; the full repository lives in the Knowledge Centre.
The playbook for AI-first venture building: innovation cycles, productivity, and velocity — with human expertise at the centre.
Why output-per-employee is being rewritten, and what it means for how you staff and bid.
The sources we return to — labour-gap data, agent orchestration, and governance. Growing every month.
The blueprint becomes your business in ninety days. It starts with a Workforce Audit.