Model-based testing has been successfully applied for test case generation in practice. Its underlying idea is to utilize models of the system for obtaining system inputs and their corresponding expected outputs. In this paper, we report on experiences gained when using a different methodology relying on models, i.e., ontology-based testing, for generating test suites in practice. Instead of modeling the system’s behavior, ontology-based testing relies on models of the system’s environment, i.e., an environmental ontology. Test cases are generated from ontologies converting them into an input model for combinatorial testing, and using a combinatorial testing algorithm for finally computing the test cases. We show how ontology-based testing can be applied in three different application domains, i.e., testing autonomous driving functionality, security testing, and compiler testing, discuss issues arising and indicate future research objectives.