CMO by title, operator by instinct. Instead of managing a team of marketers, I build AI agents that launch campaigns, read the funnel and write the copy — and I supervise the system, not the tasks.
How I operate
I don't bolt AI onto an old workflow. I rebuild the workflow around the agent — and keep the human where judgment actually matters.
Ship in public, let the numbers decide, calibrate. Opinions are cheap; the funnel doesn't lie.
One operator with a fleet of specialised agents outruns a department. I'd rather build the machine than staff it.
What I'm building
Ad platforms are dashboards built for humans clicking buttons. Give a write-capable agent the keys instead — let it operate the account directly, faster and more disciplined than any media-buyer.
An agent with write access to the ad API. It ships combinatorial creatives (~100 variants per ad), reads the funnel every morning, and enforces hard stop-rules on a cron. I supervise; it operates.
308% ROAS in its first 40 days — running the account daily, unattended.
Every AI benchmark rots once published — models train on it and the test becomes a memory. The one thing they can't memorise is what hasn't happened yet. So grade them on the future.
A public arena where AI models submit locked predictions before each Formula 1 weekend; reality scores them at the flag. Points, not money. Every call logged, season over season.
8 models on the grid, predictions graded by reality every race weekend — a public, auditable track record of machine judgment.