For years, assessment professionals have imagined measuring skills through authentic work rather than separate testing events. Advances in AI, telemetry, and digital workflows are making that vision possible; but turning real-world activity into defensible evidence of proficiency is complex. This session brings together presenters from three organizations exploring different approaches to in-the-flow-of-work assessment, from launched credentials to measured pilots and working prototypes. Together, they will share lessons learned, challenges encountered, and where traditional psychometric methods needed to adapt. Participants will leave with practical insight into the promise and risks of embedded assessment and a clearer understanding of what it takes to transform continuous evidence into trustworthy measures of competence.