Most hardware engineering organizations run design review as a checkpoint rather than a decision-making process. Designs get approved or rejected, but the reasoning behind those decisions — why a tolerance was changed, why a supplier was asked to revise, why a feature was flagged — rarely gets recorded. That rationale lives in someone's memory or disappears into email threads. The next project team starts without it, repeating mistakes that were already caught and corrected on previous programs. AI design review tools address this by capturing feedback in context and making decisions and their rationale searchable, so institutional knowledge compounds over time rather than being lost between programs.