From the factory floor to AI agents.

The factory years

I started managing people in 2000, in the graphic arts industry in Athens. For nearly 20 years I was General Manager of production at Text, Mega Press and Artion. At Artion we produced 300,000 to 400,000 boxes a day, with 30 to 60 people on the floor.

I owned suppliers, purchasing, pricing and subcontractors. I negotiated about 30% off our premium carton board, about 15% off inks on a spend of roughly €50K a month, and about 20% off laminating materials. When a machine stops or a big order is late, you learn fast that systems beat heroics.

The US years

Since 2013 I’ve worked with US companies, first in paid growth, then running operations. I’ve managed over $7 million in ad spend hands-on. As COO and CMO at Driving Force, I rebuilt our SOPs, review workflow and QA so a two-person team could run 30 client accounts instead of 15, with payroll down 40%.

Later I led growth and operations for US agencies: SOP libraries for Meta, Google and TikTok, hiring processes, onboarding templates, a meeting leadership standard, and a governed migration of 45,000+ files.

How I use AI

I build my own agents and copilots with Claude, ChatGPT and Codex. I’m an operator, not a software engineer. Every tool I build has written rules: what it may do, what it must check, and when a person steps in. AI should make your team faster, not make decisions nobody can explain.

How I work

I map how work really flows, fix the roles and SOPs first, build tools only where they save real time, and hand everything over with documentation. Bounded scope, clear decision rights, and a weekly check-in.

I love a messy problem. I’m calm in a crisis, I look at the numbers first, and I fix the real cause, not the symptom.