During the construction of service-oriented architectures, service modelers concern themselves with the characteristics of good services and how such services can be designed. For instance, they look for advice regarding interface granularity and criteria to assess whether existing software assets are fit for reuse in service-oriented environments. There are no straightforward answers to such questions – service identification, specification and realization techniques are required. Service identification and specification are well covered by existing methodologies; for service realization, architectural decision models can be leveraged. At present, the construction of architectural decision models is an education- and labor-intensive undertaking; if such models exist at all, they often are isolated from other artifacts. In this paper, we propose a new engineering approach to service modeling that leverages reusable architectural decision models as its central service realization concept. We outline a multi-level decision tree and position it as a prescriptive service realization methodology for three engagement types observed in practice. The benefits of service engineering with reusable architectural decision models are semiautomatic decision identification in analysis models, improved decision making quality, and better decision enforcement and risk mitigation capabilities.