Wouldn’t the AI be on the server by default or are you referring to a dedicated AI Client? Are you developing Server-to-Server communications to support Sharding, Zoning, Cloning, Instancing to handle massive concurrent users? I would be interested in a discussion on this.