A McKinsey report argues that biopharma companies will not fully capture AI benefits without a “structural redesign” of R&D operations. The consultancy says AI can make individual steps more efficient, but won’t create compounding learning unless companies shift to a closed-loop model where each pivotal decision generates data that informs the next. McKinsey proposes five connected decision points spanning patient understanding and disease biology through clinical trial decisions to improving impact of approved therapies, with continuous feedback across stage gates. The report contrasts traditional linear development with “closed-loop” techbio platforms that already operate as systems, including multi-agent hypothesis-and-experiment cycles. For operators and investors, the immediate implication is organizational: scaling AI in drug discovery may require rebuilding how teams share outputs and refine decisions, not simply adding tools.