The demand for customized high-quality products is growing continuously. For companies to be able to offer these products, they have to rely on innovative manufacturing technologies. One of these is the hybrid process of wire arc additive manufacturing (WAAM) and machining post-processing. To ensure that quality requirements are met, especially in an early manufacturing phase, Computer-Aided Manufacturing (CAM) is commonly used in the industry for process planning. CAM planning is, however, an iterative process, which is influenced by the expertise of the respective programmer. Assessing the quality achieved through CAM planning as well as identifying suitable process parameters are challenges, especially for WAAM parts, which often feature inhomogeneous geometrical and mechanical properties. In this article, a conceptual data-driven approach is proposed that allows the identification of quality-oriented process parameters in CAM planning. To assess the results of CAM planning, key performance indicators (KPIs) are identified. Economical KPIs, such as the machining time as well as process-related KPIs (e.g., the resulting surface quality of the machined surfaces), are considered. To determine suitable process parameters for milling, a reinforcement learning framework and an artificial neural network for predicting the expected machining outcome based on experimental data are proposed. The methodology aims to enable the data-driven identification of quality-oriented process parameters for post-processing WAAM parts with inhomogeneous properties.