Right now the limit for AI invocations are organization based, we would need to be able to set consumption/invocation limits per user