Unlocking deadlock: Why AI isn’t business as usual (yet) in healthcare

AI in healthcare is full of pilots, policy signals and promise. It is not yet embedded as business as usual across European health systems. Some could even say it’s in a deadlock. Vendor exhaustion, the lack of clinical-grade evidence, high transaction costs and outdated risk management are some reasons the health systems are failing to adopt and scale AI. But that’s not to say it isn’t getting there. Live tools are in place, trials are ongoing, and patients are seeing the benefit. What is AI in healthcare actually delivering today, beyond pilots and proofs of concept? At what point does experimentation become business as usual, and who decides when that threshold has been met? How should health systems respond to vendor exhaustion in an overcrowded and often indistinguishable AI market?