A widely observed phenomenon in interacting systems of many particles is a phase transition from a chaotic to an integrable/non-chaotic phase and vice versa. The mechanism which controls such a transition is strongly model-dependent. It is therefore necessary to develop a model of such chaotic transitions which picks up the model-independent features and explains the nature of the transition. I will present and discuss two potential mechanisms which describe such transitions. The first mechanism, titled "network classes" addresses integrable-to-chaotic transitions and separates systems into universality classes. Inspired from classical dynamics, I will extend its' definition to quantum systems. In the process, I will also introduce quantum analogues of known classical chaos markers. I will discuss where such a description works and where it fails through a quantum—classical correspondence. Then I will present the next mechanism, titled "scale separation" which will address chaotic-to-integrable transitions. I will discuss it purely in a quantum setup by constructing a (semi=solvable) model which explicitly realizes this mechanism. I will present the various properties of this model show how such a transition may occur. I will conclude with comments and future directions on developing better descriptions of such transitions.