Scale revenue with sales architecture that holds.

The company had the science — consumer neuroscience research used to predict how audiences actually respond to ads and content — but no product. No roadmap, no OKRs, no go-to-market plan, nothing scoped. I'd spent years shipping features inside products someone else had already built. This was the first time the question wasn't "what should we build next" — it was "should this be a product at all, and if so, what."

Building multiple AI-native products from zero, without a traditional team — and no existing audience to launch any of them into. Most products build in private and hope an audience shows up at launch; this started with no guarantee anyone would be there to see it.

A quiz run is one question at a time: a countdown, a 2×2 answer grid, then a breakdown of score by difficulty, streak and speed. Underneath it sits a six-stat card — players, clubs, nations, competitions, tactics, eras — averaged into an overall rating that moves with your match history. You can play the daily deck, chase the weekly, or open a room with a six-character code and take turns against someone, with the server calling every answer.