Do you know when your customers will pay their bills? Depending on your business this can be a tough question. Maybe for some of them, but can you also write it down for every one of your hundreds or thousands or even more customers, so it can be automatically estimated? Why should I, you may ask. The reason why companies do not get bust, is because they can pay their liabilities. So the financial accounting has to know what is going in and what is going out in the next few weeks to ensure enough liquidity to keep the business running – but you probably also do not want to raise liquidity if not needed as it can be expensive. The missing key here is that you as a company only have very little influence on when your customers will pay as it is their decision and out of your control. Of course, you agreed on payment terms, but these may cover several weeks between different payment targets which leads to inaccurate cash forecasts if you simply assume the net due date as the date of payment receipt. We know the payment terms and when previous bills were paid. So, how can we improve cash forecasting? Following the Cross-Industry Standard Process for Data Mining (CRISP-DM) different time-to-event and regression models were evaluated and deployed. We will discuss the challenges of delivering these new-found insights to the business department and integrating it into the system landscape and the cash forecasting process.