Many digitization projects in the industry, launched with great expectations, fail. They drag on endlessly, have to deal with skyrocketing costs or fail to deliver the expected benefits – not even a trace of greater resilience to production downtime and more efficient processes that are supposed to lead to higher quality production with lower costs and increased sustainability. Experience shows that projects that involve artificial intelligence or machine learning – both technologies that are truly state-of-the-art and extremely versatile – are particularly often affected. So what is going wrong in companies?