I've just started looking at the DX API functionality today and I've watched a number of presentations about it. I completely understand the benefit of using this capability for when I want to create a case from a JS framework front end, or where I want to update a case or process an assignment. 

What I'm wondering about is how Pega can take a front end user through a multi-step process - if the front end pulls an assignment will Pega (on submission of that assignment) make a number of decisions in order to arrive at the next applicable assignment and then communicate that back to the front end or is there an expectation that the front-end will once again call the api to pull the next assignment?