One major trend may be observed in the automotive industry: built-toorder. This means reducing the mass production of cars to a limited-lot-production. Emphasis for optimization issues moves then from the production step to earlier steps as the collaboration of suppliers and manufacturer in development and delivering. Thus knowledge has to be shared between different organizations and departments in early development processes. In this paper we describe a project in the automotive industry where ontologies have two main purposes: (i) representing and sharing knowledge to optimize business processes for the testing of cars and (ii) integration of life data into this optimization process. A test car configuration assistant (semantic guide) is built on top of an inference engine equipped with an ontology containing information about parts and configuration rules. The ontology is attached to the legacy systems of the manufacturer and thus accesses and integrates up-to-date information. This semantic guide accelerates the configuration of test cars and thus reduces time to market.