I didn't set out to build a gaming company. I started experimenting with Polsia to answer a question I've been increasingly curious about: how far can AI actually go in building and operating a real business? That experiment created Roundhall. And the problem it chose to tackle got my attention. Live-service games generate an enormous amount of player data, but turning that data into action still requires work across analytics, LiveOps, creative, localization and other teams. Roundhall is built around a simple idea: shorten the distance between what players are doing and what you do about it. Its AI agents continuously analyze player and performance data, identify opportunities across engagement, retention and monetization, and help turn those signals into actions across events, rewards and creative. Instead of: data → analysis → meeting → decision → creative → launch → measurement We're exploring how much of that loop can become: observe → decide → act → measure → learn → repeat continuously. The other experiment is Roundhall itself. This business was created through Polsia, so in a strange way we're using AI to build a company whose product uses AI to help operate games. I don't know yet how far either experiment can go. That's part of why I'm launching it. I'd love feedback from anyone working in LiveOps, game production, retention or monetization—especially on where you think this approach could be genuinely useful and where we're getting it wrong. https://roundhall-2.polsia.app/

Roundhall

Turn behavior into action!

Turn behavior into action!