Automated verification of as-built building elements is increasingly vital for quality control, safety, and compliance. Scan-vs-BIM methods that compare BIM models with terrestrial laser scanning (TLS) point clouds have advanced geometric inspection, but many studies underuse open formats such as Industry Foundation Classes (IFC) and omit implementation details, hindering reproducibility. This work develops and documents an automated, implementation-ready pipeline for geometric verification of rectangular columns with explicit IFC integration. IFC models provide as-planned geometry—shapes, corner coordinates, and centers of lateral faces. Filtering removes elements with undefined or invalid geometry. The analysis focuses on the lateral faces: an iterative principal component analysis (PCA) refines plane parameters in the point cloud, while DBSCAN clustering isolates dense, relevant points from background data. For each column, outputs include a summary of as-planned geometry; tables of deviations between as-built and as-planned models; and residuals from segmented points to their best-fit regression plane. Two 2D color maps visualize spatial differences, enabling diagnosis of potential deformations or misalignments. The approach is validated on a building with 28 reinforced-concrete columns using TLS data from a Trimble X9 scanner. Results demonstrate effective integration of BIM and point-cloud processing for automated monitoring and quality control, with strong potential to improve structural assessment and ensure compliance with design standards.