Given current shortages of skilled labour in the construction industry, this paper presents a study on the feasibility and application of an image-based, automated approach for construction site monitoring and documentation using machine learning methods. The study concentrates on object detection based on images of a specific construction site, taken multiple times a day periodically over the course of a year, that have been evaluated using the YOLOv8 technology, thus enabling progress monitoring for selected elements. Training and validation data have been created from annotated images for the object detection, which was accompanied by an evaluation of the chosen hardware and the observation viewpoint for future reference in the data acquisition. Further, a ground truth for the construction progress was generated manually to allow comparison with the results achieved by the machine learning approach. This study demonstrated, that the expected results were achieved without the need for writing a single line of code, which is meaningful given the aforementioned labour shortages in the construction industry and highlights the fast-paced nature of the machine learning field.