We live in an information age. In this new world, educational and even personal interactions are often digitally mediated. When they are, they leave traces: data which reflect what students are doing, when, how much, how well, and with whom. The ability to observe and analyse both the process and product of education has fueled explosive growth in a new field called learning analytics which is helping us to better understand how student behaviors relate to learning and to personalize educational experiences at scale.

Built on tools originally developed for digital health coaching, Michigan's ECoach tool allows us to provide students with personalized feedback, encouragement, and advice even in courses which enroll thousands of students. ECoach uses everything we know about each student's background, interest, goals, and current state to tailor the content they receive: changing what we say, when and how we deliver the message, and even who delivers the message. This last is done by providing students with testimonials from identity salient peers. Tailored coaching tools like this hold great promise for student support at scale. In this work, we draw on decades of behavior change research: nudging students at key moments, providing interventions designed to affirm values, build a sense of belonging, and encourage a growth mindset. While these tools have so far been applied with an academic focus, they hold great promise for providing advice and basic counseling at scale.