In highly configurable information systems such as SaaS information systems, business process variability management is an important issue. The variability model, which is often called the configurable process model (CPM), can be reused to configure a family of processes each serving a separate purpose or customer. If not already present, these business process variability models have to be “extracted” based on event logs residing in the databases of the target enterprise(s). Such extraction is costly to carry out manually. In this study, inspired by Software Product Line Engineering concepts, we propose a novel automated process-mining-based method by extending the “Alpha” algorithm for process discovery as a preliminary solution. The proposed method takes a set of event logs as input,and in three phases, outputs a CPM in terms of a model called “BPFM”. To evaluate the method, we used the Goal-Question-Metric approach in a case study on 10 cases. For this purpose, input event logs were artificially extracted from the cases’ existing BPFM models and were fed as input to the proposed method. Then, we observed if the output models of the method were similar to the preliminary existing ones. The results showed that the method was promising in identifying the CPMs,since the extracted models involved activities that were 97. 5% identical to what was expected. Moreover, a structural precision of 98% and a structural recall of 97. 3% were obtained. The set of configurations derivable from the output models was 100% similar to and provided 100% coverage over the expected configurations.