If we look at current IT infrastructure model we clearly see that traditional way of management and administration are no longer sufficient. Imagine setting millions of thresholds in a monitoring software for thousands of virtual servers. That simply will not work. It is time to engage mathematical statistics into our analytical tasks. Robotics and automation are the current tools. We developed a deeply mathematical module called „Artificial Intelligence” that allows us to clasify and predict on any sort of data. We bring regression algorithms along with Neural Networks. The user can train multilayer neural network what is typical for his set of data without defining any threasholds. Anything that breaks the rule is a anomaly, very easy to detect. At the bottom we present historical data where the top graph shows prediction build with implemented algorithms of „Artificial algorithms”. Nowi t is very easy to combine current data set with its prediction. We can be informed when our prediction reaches the limit even if at the current stage we cannot identify all symptoms.