Artificial intelligence is increasingly used to inform or mediate consequential organizational decisions, yet accountability for those decisions remains human. In practice, human judgment in AI-mediated decision-making is often implicit, informal, or undocumented, complicating accountability, oversight, and post-hoc explanation. This paper reframes human judgment not as an individual intuition, but as an organizational asset—one that must be deliberately defined, exercised, preserved, and governed when decisions are influenced by AI systems. It introduces Judgment Assurance, an organizational governance framework designed to help institutions define, record, own, and guard human judgment in AI-supported decision-making. Judgment Assurance provides institutions with a structured means to ensure that, when AI influences outcomes, a human can clearly, contemporaneously, and defensibly explain why an AI output was followed, modified, or rejected. Judgment Assurance is intentionally technology-agnostic and scalable. It can be applied narrowly to specific high-risk or high-impact decisions or integrated into existing governance, risk, and compliance structures, and it incorporates a feedback loop by capturing reasoned decisions for review, accountability, and continuous institutional learning. The framework does not replace existing legal, ethical, or technical AI governance frameworks; rather, it complements them by addressing a persistent gap: the governance of human judgment itself. Human judgment is an institutional inheritance. As organizations increasingly rely on AI systems, they owe it to themselves, and to those they serve, to ensure that judgment is exercised deliberately, retained visibly, and not allowed to atrophy through delegation to machines.