AstraZeneca said it is leaning on AI to reduce risk and speed filings in late-stage development, even as it acknowledged late-stage setbacks in parts of its pipeline. CEO Pascal Soriot said AZ’s AI efforts have not yet produced clinical-stage candidates, but could improve success rates later in development. In a separate update, AstraZeneca reported disappointing data in two late-stage programs: an antibody-drug conjugate in stomach cancer and Ultomiris in a rare disease. Together, the messages show a pharma recalibration—using AI as a systems-level risk tool while managing clinical program attrition.