AI – and in particular its new star, generative AI – is today a central theme in company boards of directors, in management discussions and in informal exchanges between employees wishing to increase their productivity . Unfortunately, behind the ambitious headlines and tantalizing potential lies a sad reality: most AI projects fail. Some estimates put the failure rate as high as 80%, almost double the failure rate for enterprise IT projects a decade ago. However, there are approaches to increase the chances of success. Companies can significantly reduce their risk of failure by carefully navigating the five critical stages every AI project goes through to become a product: selection, development, evaluation, adoption and management.