Currently, in the design and construction of large industrial enterprises, three-dimensional BIM models of buildings and structures are being introduced. This article describes the practical application of the Leica RTC360 terrestrial laser scanner at the industrial site of PJSC Uralkali, Berezniki, Perm Territory, under construction. The purpose of the survey is geodetic monitoring of construction and the implementation of as-built documentation for the sections of heating, ventilation, water supply, sewerage and heating networks of the shaft building No. 2. To bind point clouds to the construction coordinate system and heights, linear-angular networks are created inside existing industrial enterprises using an electronic total station to determine the coordinates and heights of marks for a ground-based laser scanner. The article considers an automatic method and method for registering the best approximation for combining point clouds obtained using NLS with BIM models. As a result, the automatic method is used when there is a building coordinate system on the object under study. The method of registering the best approximation is used when there is only a conditional coordinate system, for example, for control assemblies of metal structures. It is proposed to perform automatic classification of point clouds, both inside industrial facilities and outside. The classification process is based on a machine learning algorithm that relies on calculations using a video card. Ground-based laser scanning was used directly at a real mining enterprise under construction, which allows further geodetic monitoring, creation of as-built documentation and timely detection of deviations from the project using BIM models.