no problem, it should be easy to know if the model is waiting on an API call to return. Logically/generally, you cannot process another proposal until it completes, though some part of the application state can be parallelized in specific scenarios. All you have to do is queue the proposal and present it again in NAP. You can even implement some behavior where only the latest proposal is processed or the entire queue (one after another). Hopefully, it's never a big queue.