This presentation examines how artificial intelligence is being applied to health care quality measurement and evaluates the differences between generalist models and specialized models in this context. The discussion is grounded in prior work published in Harvard Business Review with John Glaser, and extends that analysis using observations from real-world quality measurement deployments. The session focuses on how specialized artificial intelligence models, designed around explicit quality measurement logic, perform when applied to digital abstraction and member classification. It reviews how awareness of denominator criteria, exclusions, timing windows, and evidence hierarchies affects consistency, validation, and audit defensibility, and contrasts these outcomes with those produced by more generalized approaches. Rather than making prescriptive claims, the presentation shares measured observations from applied use, including differences in stability across reporting cycles, alignment with certified human reviewers, and downstream operational rework. The goal of the session is to provide a clear, factual examination of how specialized models behave in quality measurement workflows and what this implies for organizations responsible for accreditation-ready reporting and oversight.