The vision of Ambient Intelligence (AmI) was first developed in the late 1990s. It describes new worlds, economies and paradigms that emphasize the centrality of human experience, however, distinguished from related visions such as ubiquitous and pervasive computing. A key feature of the AmI vision are the seamless intelligent environments and gadgets, capable of anticipating people’s needs and motivations, and acting autonomously on their behalf. So what can be gleaned from exploring the conditions under which this innovation domain evolves over time and how it adapts to various criticisms and technical challenges? The AmI vision not only represents possible futures but actively creates the worlds in which AmI applications appear to be possible. Visionaries and research leaders build expectations, marshal resources and align key stakeholders. Promises and progressions toward realizing AmI have performative and generative features but the original promise of intelligence has largely failed. This outcome points to a two-sided problem. The definitional looseness of intelligence is permissive of what can be expected of the role and scope of artificial reasoning in AmI interaction paradigms, while ordinary human reasoning and knowing what people actually want and need remains persistently elusive. Grappling still with the problem of what the intelligence in Ambient Intelligence can stand for, research and development has shifted its focus toward the design of practical win-win solutions, coined synergetic prosperity.