How to run an AI company like a AAA live-service game.
Every few weeks an AI company ships, and within the hour the internet has decided how it feels. I ran that exact problem for 15-million-player live games. It has a playbook, it just does not have a name in AI yet.
Let’s build it together
The same signal that grows a game community grows the crowd around a model.
At Ubisoft I ran the community for Ghost Recon, held it at high positive sentiment through launches and rough patches, then built the first global fan advocacy program for its next release and unveiled it live on the E3 stage. AI products now work the same way.
Where I ran it- 15Mplayers in the live community I managed
- 85%positive sentiment held through launches
- E3stage unveil of a first-of-its-kind program
Where fan-led growth pays off for AI
- 01 · Distribution
Recommended by the models
When someone asks an AI what to use, you want to be the answer. Models learn from what real users write on Reddit, in forums, reviews and threads, so earning that advocacy is now distribution.
- 02 · Sentiment
Sentiment you can move
The point is not a dashboard. Listen to the sentiment, read it against your product roadmap and business goals, turn it into usable recommendations that improve the product, then communicate the changes back to the community the right way.
- 03 · Feedback
A feedback corps from your power users
Find the power users already inside your userbase and turn them into a trained feedback corps. They pressure-test releases early and surface issues before they spread, so you get signal you can act on.
- 04 · Operations
Run it like a live service
Manage your AI community the way I ran games with millions of players: real-time, close to the product, ready before sentiment turns.
I do not just advise on this. I build with it.
This whole site
I designed and built this site end to end, myself. What you are reading is the proof.
Tools, agents and skills
I build my own AI tools, agents and skills to do the work, not slideware about them.
Building it for real
Part-time founding role at an AI-driven startup for fan engagement.
None of this is a cost.
It is growth.
The AI companies that treat the crowd around their model as a growth engine, not a cost, will pull ahead in a way the others cannot buy back.
I have run this at the scale of a live game: a 15-million-player community, read in real time and held at 85% positive through launches and rough patches. The opportunity now is to build that for AI, inside a team shipping models.
Three ways in, in the order most AI companies start.
- 01
AI Sentiment SOS
When your community turns on a release, a price change or a deprecation, I read what they are actually saying, find the signal, and hand you a build-ready plan across product, comms and community. The fastest way to stop the bleed, and the work I have done longest.
- 02
The Fan-Led Growth Engine, for AI
Once the fire is out, I build the thing that stops it recurring: the developer and power-user community, the advocacy and ambassador programs, the sentiment defence, all instrumented to adoption and retention.
- 03
Fan Moments
A model release or a developer-conference moment your community actually feels. Creator programs and ambassador cohorts around it, so the people with their own audiences turn up early and bring theirs. The UGC that comes out of it is what your next users read, and what the models read too.
I’m taking on a small number of founding AI partners.
I have run this playbook at the largest scale in gaming. I have not yet run it inside an AI company, and I am not going to pretend otherwise. So I am taking on a few founding partners at pioneer terms: you get senior, hands-on fan-led growth you could not hire fast enough, and we build the case study together.
If you are an AI company between Series A and C, with a passionate user base and no senior community or advocacy leader yet, this is built for you.
Talk about a founding-partner pilot