The transformation of manufacturing companies towards a carbon-neutral economy requires energy transparency, energy analyses and the implementation of energy efficiency measures. Given the continuing skills shortage, the need for automated analysis methods to gain insights from measurement data is increasing. Expert systems that combine the knowledge of multiple experts, analyze load profiles, and derive energy efficiency measures are one approach to tackle this challenge. This paper presents an expert system that quantifies energy efficiency potentials based on the detection of machining cycles and derives promising measures. For this purpose, a new algorithm for the detection of machining cycles is introduced, which shows an accuracy between 76.7% and 94.3% on a representative production day for electrical load profiles of different types of production machines. Since the detected machining cycles are in a form impractical for further processing, information is extracted as energy performance indicators. The expert system utilizes this aggregated information to identify energetic hotspots and derive appropriate energy efficiency measures. The machining cycle detection based expert system is demonstrated on a typical production chain for the metalworking industry within the ETA research factory at the Technical University of Darmstadt.