Prioritizing Artificial Intelligence Opportunities for Food as Health in the U.S. Agrifood Value Chain

Food and health are converging across the agrifood value chain, and artificial intelligence (AI) is widely promoted to operationalize the link, yet firms cannot tell which food-as-health opportunities current AI can address. We identify the agrifood industry's highest-priority food-as-health opportunities in the United States and assess, against the peer-reviewed AI literature, how far current AI can address each. A PRISMA 2020 systematic review of 34 U.S.-focused studies establishes the problem space; a Jobs-to-Be-Done map organizes ninety opportunity areas; a blind discovery forum with 37 senior U.S. practitioners gives an independent reading; a convergence analysis joins the map and the forum; and a separate assessment grades each opportunity against the AI literature. Practitioner priorities converge on twelve opportunities in the two segments that physically determine food's health attributes, agricultural production and food manufacturing. These resolve into five AI capability families. The evidence is strong for sensing, traceability, design, and personalization, and more qualified for evidence synthesis and regulatory intelligence, where reliability limits require human oversight. One core need lies beyond AI, a preliminary finding of a symmetric long-term-contract gap that appears relational. We grade an agenda by how far current AI reaches, with an explicit boundary on what it cannot address.