The industrial utilization of machine learning (ML) technology is still in its infancy. This chapter provides empirical insights on how ML has been deployed in three firms and which forces are at work in this transformation. It is clear that two complementary advancements are needed to make ML generally useful: while ML technology thrives on access to big and varied datasets, the first advance is a reduction in the laborious work of manually cleaning, sorting and labelling the data, which defines how knowledge creation, technology and organization are interrelated. The second advance is to find sensible collaborative modes of data access and sharing, which challenges the very boundaries and interdependence of firms since the value of data for training ML algorithms depends on access to others’ data.