Commodity smart phones and initial Internet-of-Things (IoT) devices have made the visions of ubiquitous computing common place. However, despite the rapid adoption of this technology, we have not progressed much past the use of location-based systems in commercial ubicomp systems. The Ubicomp Lab at Carnegie Mellon University has been using peoples interactions with such devices to better understand human behaviour and routines. In this talk, I will describe the value of leveraging human behaviors and routines in interactive mobile and IoT-based systems. I will demonstrate how the ability to opportunistically extract these routines (and anomalies) can dramatically enhance interactive systems in a number of domains, including healthcare, driving, education, and sustainability, as they can be used to create effective interventions and supporting systems.