Modern buildings are equipped with various complex automation systems that produce vast amounts of data. The operation of these systems has significant influence on energy demands and user comfort in the premises. Therefore, an optimal configuration of the systems to increase energy efficiency is desired by the operators of the buildings. An automated decision-making process that evaluates and analyzes the bulk of available data can help to achieve these objectives. Hence, this work introduces a framework for the automatic generation of recommended configuration changes to increase energy efficiency and user comfort. The approach makes extensive use of semantic building information modeling as basis for the selection of potentially useful measures as well as for their assessment. Expected consequences of configuration changes are forecasted via simulation. Additionally, current results of an application of the proposed solution in an office building are discussed.