I would suggest that using a platform that already has the users/students/learners on it is a good starting point. That sorts out access and serving content where they are experiencing life.
Apart from trying to bring education into games, those games are already delivering on missions and milestones. We (teachers) are saddled with a burden of pedagogy that has largely fallen behind the curve of delivering learning content (I would contend that more people learn on Youtube than any other platform in existence). We are also missing out on the attained skills, knowledge and motivations learners/students already have by being on these gaming platforms. I liken this to the phrase "Digital Citizens" where we have skills in the digital realm that allows us to communicate and work globally. It is possible to quantify these skills and show people how skilled they really are.
In the same sense, students/learners on these gaming platforms have quantifiable skills that we are simply not acknowledging. And saying then that one wants to bring educational lessons INTO games, is perhaps using the wrong end of the stick?
The games are already delivering educational lessons. We are just not recognising them for what they are. They are delivering skills, which we are not measuring.
If we lack the insight into how we tap into this worldwide phenomenon, then we first need to upskill/reskill to identify what is really happening. I suspect once that is done, tweaking the existing missions and milestones are much simpler.
I think it is harder to try and imagine it when one imposes an "older" learning pedagogy onto a newer format? Then one attempts to mold the newer platform to display the older paradigm.
Much more sensible to take only the essence of the educational content, and contemplate how to let that flow within what people are already doing.