For the owner of a car, it is important that all quality scales are met, especially on a premium brand. One such scale is the acoustic situation inside the cabin: a new car should not have any squeak, rattle, or other unwanted noises perceptible by the passengers. The root causes of such acoustic anomalies are manifold. Currently, the readily assembled cars are driven on an indoor roller dyno or an outdoor test track to have a human driver detect acoustic anomalies. To shift this subjective quality estimation to a reproducible and more objective level, the goal is to develop an automated test procedure for squeak & rattle integrated in the test facility of the production line. This is where the formerly proposed AI-based automated squeak & rattle detection comes into the play. We explain the boundary conditions for such a procedure and describe a possible approach, before we take a closer look at a pre-test setup in a lab with a suitable excitation of the whole vehicle body, as well as the measurement setup of the microphones. The initial results of a measurement campaign are discussed, highlighting the precision of the anomaly detection and – of course – the limits. The paper ends with an outlook on next steps.