Artificial intelligence has progressed a long way from the early days of rule-based expert systems. For building diagnosis and machine health monitoring systems, there are many very powerful techniques available. The most important of which are qualitative reasoning, qualitative simulation and the use of deep models. Recently, online expert systems, that is systems which are permanently coupled to their test and measurement systems, are becoming increasingly important. In the paper the author examines several of the fundamental techniques which are employed in systems such as these today.