As a Roblox developer, it is currently tedeious to determine viable abstract steps (i.e. data sent from client to server and data recieved) to create usable unit testing for systems directly from a playtesting session for use with the new open cloid engine api for executing luau.
I feel that there is a potential valid use case for machine learning or a similar approach alongside supporting systems to suggest potential abstractions for this purpose based on user inputs and event executions in studio playtesting sessions - and potentially to explain what they do. Although such a system may not be of mich use for highly experienced developers, it could feasibly serve quite well as a way for those who have never worked with CI/CD workflows to have an easier time learning the core concepts of the practice - hopefully allowing for more wide-spread adoption of such workflows across the roblox developer community over time.