The technology has never been better. Sensors are faster, denser, and more affordable than ever. So why do so many reality capture programs produce data no one uses six months after closeout? After decades in AEC across North America's most complex infrastructure, the answer isn't the hardware. It's the strategy gap between what gets captured and what asset owners need. Geometry without intent isn't intelligence. A point cloud without a data model isn't a digital twin. A BIM deliverable not designed for operations is an expensive archive. This session examines patterns that undermine reality capture ROI: misaligned scope, data structures serving geometry over function, AI layered onto poor foundations, and robotics optimized for coverage over insight. We'll identify the decision points where trajectory gets set before the first scan. Not a product pitch. An unfiltered view from someone who's seen these programs succeed and fail at scale, and believes the industry can do better.