Press

There is a sense of euphoria at the start, and it is usually huge. Expectations reach unimaginable heights before an AI project begins, then disillusionment sets in as the weeks and months pass by. The end result is not the hoped-for revolutionary improvement to your own processes or brilliant new service idea, but a proof of concept (PoC). This evaluation shows what is possible and quickly reaches a state in which many AI initiatives remain forever: stuck in the PoC nightmare. The project fails as it moves from “this is possible” to “this is productive and earns money” – not because the developed AI solution does not work, but because those responsible have paid too little attention to some peculiarities and challenges typical of AI. In order to avoid the PoC nightmare, the IT service provider adesso believes companies should pay attention to these ten points: