I run a company that has no employees in it.
The agents run the day and I hold the control points. I got here by security-testing other people's AI systems until I understood what it takes to trust one.
What I do
I run Charles Green, LLC, a SaaS holding company with no employees, where AI agents run the daily operations under written doctrine.
They pull the funnel numbers, hunt gaps in the products, open pull requests, draft outreach, and write the board report. I stay at the control points. Nothing sends, spends, deploys, merges, or publishes without my approval.
Zero employees is not zero control. Autonomy is earned one surface at a time, and only where the monitors have actually been seen to fire.
Where I came from
I did not arrive here through AI hype. I spent years as a CISO and engineer. My work in AI started by security-testing AI systems in finance and payments, then learning to build them because it was the only way to close the gap between what looked good in a demo and what survived production.
That background still shapes how I build: gates, approval paths, narrow permissions, and systems designed to fail safe when the machine is uncertain. The model is not a security boundary. The boundary lives in the infrastructure between the model and what it can touch.
Before this company I spent 25 years building and operating across 11 companies in the United States, Thailand, and Japan. Most of those ventures failed. I learned more from that than I would have from a cleaner resume.
What this site is for
This is where I write about running the current company in public: the working parts, the failures, the gates that stopped work, and the monitors that looked green until they did not.
Most of what I publish comes back to the same questions: what it looks like to run a company on loops and gates, where AI systems break when nobody designs the control plane, and what a run that failed turned out to be telling me.
It is written for people running AI-native operations who want the evidence rather than the mythology.
Open to the right role
I am open to fractional and interim roles with US teams as a Forward Deployed Engineer or GTM engineer, employment or contract. US citizen. I work globally and can relocate for the right role.
The best way to reach me is LinkedIn.
Short bio180 wordsThird person, for reuse
Charles Green runs Charles Green, LLC, a SaaS holding company with no employees, where AI agents run the daily operations under written doctrine. The agents pull the funnel numbers, hunt gaps in the products, open pull requests, draft outreach, and write the board report. Charles approves the control points. Nothing sends, spends, deploys, or merges without him.
He spent years as a CISO and engineer. His five years in AI began by security-testing AI systems in finance and payments, then learning to build them because he had to show the engineers how to pass his own tests. That security background still shows up in the systems he publishes now: gates, control points, and production paths designed to fail safe. Before that he spent 25 years operating across 11 companies in the United States, Thailand, and Japan. He documents the company in public, the bad runs included.
He is open to fractional and interim roles with US teams as a Forward Deployed Engineer or GTM engineer, employment or contract. US citizen. He works globally and can relocate for the right role.
Elsewhere
The newsletter
Get the issues by email
Gate verdicts, what the routines did, and the runs that failed.
Double opt-in. One click to leave, in every message. You get the newsletter and the occasional note on what I'm building. I hate spam too: your address is never sold, and nobody else markets to it.
