Polymetallic sulfide ores are often not amenable to cyanide leaching due to the presence of several elements and minerals capable of interfering with this process. Thus, various strategies, such as chemical pretreatments, are often studied to improve the efficiency of cyanidation. Beyond the results of such strategies, it is important to understand the changes occurring on the mineral samples during these pretreatments. Herein, an alkaline pretreatment was applied to a silver concentrate (∼8 kg Ag/t) composed of polymetallic sulfides (Fe-Pb-Mn), which increased the silver extraction during subsequent cyanidation from 40% to 80% and decreased the cyanide consumption in half (from approximately 60 to 30 kg NaCN/t). X-ray diffraction (XRD) and ICP-MS indicated that the pretreatment could remove significant amounts of elemental sulfur, which is a known cyanicidal agent. The dissolution of significant amounts of sulfur was confirmed by chemical analysis, which also demonstrated that the dissolution of iron, lead, manganese, and silver were negligible during pretreatment. At surface level, X-ray photoelectron spectroscopy (XPS) demonstrated that the pretreatment exposes fresh sulfide surfaces (e. g. pyrite). In addition, the XPS spectra indicated that the pretreatment facilitated the exposure of clean mineral surfaces. The presence of cleaner surfaces suggested a more uniform and less hindered diffusion of leaching agents through the mineral. Indeed, fitting the extraction data to the shrinking core model showed that pretreated samples featured a nearly ideal diffusion-controlled process, while in the case of untreated samples this fitting was less adequate. During cyanidation of both untreated and pretreated samples, lead build-up was detected on the surface (readsorption), which suggested that this phenomenon does not affect the efficiency of a leaching process. This study highlights the importance of combining bulk analytical methods with surface-sensitive techniques to obtain a more complete understanding of leaching processes.