We describe the validation of a software solution that automatically detects changes in downhole drilling conditions. Based on early warnings provided by the system, the drilling team can take pre-emptive action when conditions start to deteriorate. However, we have found that the decision maker takes time to build confidence that the reported symptoms, if ignored, will lead to a drilling incident.To illustrate this problem, a particular BHA run, where the system has been used, is analyzed in detail. A leak-off test should have been performed after drilling out a 300m long cement plug. While drilling the cement, the system gave several warnings that the cuttings transport was poor. Nevertheless the decision was made to continue drilling. Close to the end of the plug, the hole started to pack off and shortly after, mud losses were experienced compromising the planned leak-off test. While pulling out of hole, the system triggered new warnings because of the possible presence of a cuttings bed. The decision was made to clean the hole thoroughly and at the second attempt, the subsequent trip out of hole was uneventful.The principle of the system is to compare results from calibrated physical models of the well with surface and downhole measurements. The system analyses the following indicators continuously: sliding and rotational friction, downhole and pump pressure, free rotating weight deviations and pit volume variations. Experience has shown that using multiple symptoms to detect abnormal drilling conditions is essential because similar problems do not always present the same pattern of indicators.The system has regularly shown its ability to provide early warning signs of changing downhole conditions well before an eventual incident. This consistency has helped build confidence amongst the decision makers who are then ready to take prompt actions in response to warnings.