An AI esports-scouting venture I ran end to end. It taught me what product-market fit actually costs — and gave me the confidence to keep going.
The whole motion was pure go-to-market. I took one alumni connection and kept pulling the thread — from an amateur player, to a semi-pro, to a retired coach, and finally to the manager-level gateway of China's top-three clubs, where we reached a preliminary partnership.
Then I cold-reached out to Fnatic — one of the most storied organizations in esports — and got a reply straight from their general manager. We went on to sign a technical-cooperation agreement with Fnatic's CS2 team. No warm intro, no fund behind me. Just the thread, pulled patiently, all the way to the top.
The product was a bet on VLM technology: treat a professional player's style of play as a kind of fingerprint, then scan ordinary players at scale to surface the rare ones whose play matches a pro's. Talent discovery by match-degree — finding the next prodigy hiding in the ladder.
It didn't survive contact with reality. Technically, we couldn't get our hands on enough professional match demos to train the model — and there was no clean way to pull ordinary players' demo files in real time to screen them. Even with the data, the precision the idea demanded was hard to reach.
The vision was grand. Reality broke it one constraint at a time — which, I've learned, is exactly how grand visions tend to go.
Product-market fit is a balance of three things — and a venture dies when any one of them gives way.
What the technology can actually do — not what you fantasize it can. What the market on the ground actually allows — is the data even there, and can you get it. And how big the prize actually is. We walked out with real product insight CS2 teams wanted — but it didn't transfer, and a single-title CS2 market was too small to carry a company. So we killed it.
It was the first time I really dealt with the world outside a classroom — and I came away surprised by how open most people are to new technology and new business. The alumni thread held the whole way up; the conversations with Fnatic went smoothly. That gave me real confidence. And the passion of building something with like-minded brothers deeply shaped the choices I made after graduation.
It also taught me the discipline that now runs underneath everything I build: kill the things that won't carry a company, and go looking for the ones with real ground under them. That search — for something with landability, not just a grand story — is what led to Iterant.