Challenges in Commercial Deployment of AI: Insights from The Rise and Fall of IBM Watson's AI Medical System

When IBM set about commercializing its artificial intelligence-driven Watson AI in the healthcare market, its early successes were widely publicized. Senior managers and the media claimed that its diagnostic features would soon surpass those of the sharpest doctors. The case describes the large gap between what was promised and what happened in practice, offering insider insights on why IBM's projects failed. As the corporate commitment to AI escalated in response to successful lab results, cognitive dissonance arose between managers' expectations and what they could actually deliver. How could that have happened? Three reasons for Watson's downfall are explored: 1) The tendency for societal expectations to exceed the actual technical capabilities, leading to a gap in perception between AI in the lab and AI in the field. 2) Overselling of the economic benefits of AI by the salesforce; 3) Failure to secure the cooperation of key stakeholders, notably doctors who were asked to improve the performance of AI but were undermined by claims that AI could outperform them.