The gap between knowing what AI can do and knowing what to make it do — that's where I've spent the last three years.

For two decades I was inside the machine — managing partner at a digital agency in Atlanta, building content systems and production infrastructure for brands that needed to stay current in a medium that never stopped changing.

That work wasn't glamorous in the way agency portfolios tend to make it look. It was translating between languages — client language, design language, production language, business language — and finding the system underneath all of it that could actually scale. I was a maker and a manager, a strategist and a builder. That combination was the job, even when the job didn't have a name for it yet.

What I built in those years wasn't just campaigns and content. It was judgment. About what works, what doesn't, why most digital projects fail, and how to structure the ones that don't.

In 2023 something changed. Not for everyone at once — just quietly, for anyone paying close enough attention. The tools crossed a threshold. AI video went from a novelty to a genuine production medium inside of about eighteen months, and the question stopped being can it work and started being who knows how to work it.

I started building. Not experimenting — building. Studying the tools the way I'd once studied craft: what each model excels at, where the seams show, how to write to a generator the way a director talks to a cinematographer. Prompt architecture as creative direction. AI tool stacks as production infrastructure.

The short film Riviera Breakup came out of that period — a French New Wave pastiche assembled entirely through AI generation, edited with the same editorial sensibility I'd spent twenty years developing. It was proof of something I'd been arguing in theory: that taste and judgment are the scarce resources in this era, not technical fluency.

I run an AI Video Strategy and Production practice. That means different things for different clients. Sometimes it's creative direction and production — concept through delivery, built on AI-native tools. Sometimes it's strategy and systems work — helping teams figure out what to build, what to stop doing, and how to integrate AI into their existing production infrastructure without creating chaos. Sometimes it's fractional embedded work, two or three days a week inside an organization that needs senior-level guidance without the overhead of a full-time hire.

The common thread is always the same: the work requires someone who understands both sides. The craft and the strategy. The tool and the outcome. The frame and what goes inside it.

That's a rare combination. Twenty years of agency experience doesn't automatically produce it, and neither does being early to AI. It comes from having spent long enough doing the work — the original work, and the new work — to know the difference.