Biopharma leaders are revisiting how to operationalize AI in discovery and development after McKinsey argued that AI value won’t “compound learning” unless companies redesign R&D around closed-loop decision points. In its report, McKinsey said step-level efficiencies can increase within stage gates, but systematic feedback across decisions is needed to create continuous learning. McKinsey proposed five connected decision points spanning patient/disease biology understanding through improving impact of approved therapies. The consultancy pointed to multi-agent and loop-based systems—such as FutureHouse’s Robin and DeepMind’s Co-Scientist—as examples of closed-loop cycles generating hypotheses and experiments. For established biopharma, the immediate question is how to integrate AI into existing workflows while changing governance so data from pivotal decisions feeds subsequent decisions.