The most important thing to understand about ML is that it is incredibly accessible. The terminology and the concepts that underlie Machine Learning can be extremely complex and hard to understand, but here are so many open source frameworks that make these algorithms extremely easy to apply. Like many other things in life, with Machine Learning — you really can fake it until you make it.
Even if you have a very basic understanding of how the vast array of algorithms work, you can systematically create algorithms that do really powerful things.
Another thing that I’d mention for PM’s that are wondering how this sort of thing is deployed is that these Machine Learning concepts pair extremely well with horizontally scaling cluster computing technology like Apache Spark — combining them can create incredible insights at scale.