AI agents that move money.
I build AI systems that do real economic work — bill, pay, trade, rank.
Almost all AI ships with a human as the safety net. The model drafts, someone reads it before it goes out, and a mistake costs you an afternoon. I work on the other kind: systems that act, where the action is financial, irreversible, audited, and nobody reviews each one before it happens. That turns the engineering problem from accuracy into accountability — audit trails, blast radius, escalation, and who answers when the agent is wrong. Billing is where I do this today, in regulated markets, which is the strictest version of the problem I've found.
Where I build
Platform briefed to Gartner and IDC · 99.7% config accuracy · 90% faster setup
Proof
Head of AI Programs, MaxBill — Production AI inside billing infrastructure, where an error is a compliance event rather than a bug report. Configures tariffs and complex B2B offers against regulator-auditable logic: 99.7% accuracy, 90% faster configuration. → MaxBill
Co-founder, LazySEM — A production SaaS for working SEOs: monitoring that catches a rogue noindex before traffic drops, briefs a writer can actually use, and cannibalization found before it costs you. Same failure class as billing — systems that break quietly and expensively unless someone instruments them. → LazySEM
AI Forex System — Running an AI-driven system with real capital. Results aren't published yet — when there's enough track record to be worth reading, it'll be here.