Quantification of modal mineralogy in drill-core samples is crucial for understanding the geology and metal deportment in a mining operation. This study assesses conventional procedures to quantify modal mineralogy, that includes an initial drill-core logging, followed by petrographic descriptions and SEM-based automated mineralogy analyses performed in selected regions of interest, against a novel approach using laser-induced breakdown spectroscopy (LIBS). Our proposed methodology aims to quantify the modal mineralogy directly in a drill-core sample, avoiding previous stages of selection and preparation of samples. The novelty of our methodology lies in the simultaneous selection of spectral signals corresponding to a group of elements that are interrelated within a mineral species. The resulting signal combination is strongly correlated with a mineral found in the sample. Our proof of concept combines previously described mineralogy with a detailed spectroscopic and principal component analysis. The selected spectral signals are defined as “mineralogical patterns”, which are processed using supervised chemometrics methods, such as artificial neural networks, to enable an automated mineral classification. We implemented our workflow in three molybdenite-bearing drill-core samples, yielding results comparable to operational characterization, based on petrographic studies, and validated by QEMSCAN analyses, for a suite of ore and gangue minerals, including molybdenite, pyrite, hematite/magnetite, quartz, and aluminosilicates. In brief, we demonstrate how the LIBS-ANN technique can perform automated mineral quantification directly in selected drill-core regions of interest, minimizing previous sample preparation and without expert judgment.