Automated and robust retrieval of three-dimensional (3D) Computer-Aided Design (CAD) objects from laser scanned data would have many potentially valuable applications in construction engineering and management. For example, it would enable automated progress assessment for effortless productivity tracking, automated 3D image database searching for forensic and legal analysis, and real-time local modeling for automated equipment control and safety. After reviewing and analyzing previous research in the field of automated object recognition, this paper presents a new approach for robust automated recognition/retrieval of 3D CAD objects in range point clouds in the Architectural/Engineering/ Construction & Facility Management (AEC-FM) context. This approach is validated in laboratory experiments. A first experiment demonstrates that this new approach can efficiently and robustly automatically retrieve 3D CAD model objects in construction laser scanned data. A second experiment demonstrates how this approach can be used for efficiently assessing construction progress. The results presented here are preliminary but conclusive for proof of concept. More extensive field experiments in this and other application areas will follow to characterize performance trade-offs in practice.