Let's consider a complex system such as an industrial furnace where temperature control is critical, and achieved through commands like adjusting burner flow rates. If you're looking to implement a new controller or improve an existing one, the first step is to conduct an automation audit. This involves using periodic sensor and command values to create a mixed physical and statistical model. With this model, we can determine the predictability of temperature and evaluate the performance of any existing automation. Depending on the audit results and your needs, we can provide a prediction model for you to use (for example, to send warnings), or we can develop and commission an optimal controller based on the process model.