AI-supported digital tools offer great promise for improving college students’ academic engagement and progress, yet little research has focused on whether and how the character of tool implementation might influence tool uptake in specific campus contexts. Our study documents variation in implementation of the same digital tool across four public university campuses in a large state system. Scaled computational observation across campuses reveals substantial variation in magnitude and style of tool use. Site visits and qualitative interviews with university staff reveal parallel variation in how the tool is embedded in technical infrastructures and inter-office divisions of labor and how data are used to inform tool use. Findings suggest that the uptake of digital tools, and consequently their sustained adoption, is partly dependent on how tools are embedded in socio-technical systems that intertwine technologies, organizational conditions, and interpersonal relationships in context- specific ways.