IGO Ltd’s (IGO’s) Nova-Bollinger Deposit (Nova-Bollinger) is a magmatic nickel-copper-cobalt (Ni_x0002_Cu-Co) sulfide deposit found 160 km east-north-east of the town of Norseman in Western Australia (WA). The deposit is hosted by the Mesoproterozoic rocks of the Fraser Zone, which is part of the Albany Fraser Orogen (AFO). Since mining commenced in 2016, IGO’s Nova Operation (Nova) has mined and processed to 31 December 2020, 5.63 Mt of ore grading 2.04 per cent Ni, 0.86 per cent Cu and 0.07 per cent Co. Nova’s mine geology team (MGT) has used commercially available implicit modelling (IM) and general-purpose resource modelling (RM) industry software systems, to prepare the 31 December 2020 (CY20) update of Nova-Bollinger’s JORC Code reportable Mineral Resource estimate (MRE). This CY20 MRE update was predominantly based on ~386 km of diamond core drilling, with the drill holes having a nominal grid spacing of 12.5 by 12.5 m throughout the mineralised zones of the deposit. Using IM tools, Nova’s MGT interpreted and modelled 22 separate domains from the drill hole logging, assaying results, and the MGT’s underground mapping. These IM closed volumes were then used to control the estimation of MRE variables into conventional industry digital block models using RM and other ancillary software systems. During 2021, a commercial machine learning (ML) software was trialled by Nova’s MGT for preparing block models of Nova-Bollinger’s MRE domains. The motivation for the trial was that the ML process could expedite modelling through the direct creation of a domain coded block model from the drill hole logging. Compared to the industry standard approaches of wireframing or IM, where the modeller usually makes many subjective choices to produce the domain model, ML offers a more objective ‘hands-off’ process for determining the connectivity of domains defined in drill holes. Additionally, the ML process provides a quantitative confidence metric associated with its output that may have utility in mine production and resource classification. In this paper, the results of the Nova-Bollinger ML modelling trial are compared to the MGT’s current workflow for MRE work. The MGT’s conclusions as to the advantages and disadvantages of the IM and ML methods for preparation of MRE domains are discussed. The ML software was also tested as a method of rapidly interpreting geology outside the principal zones of MRE interest. These testing results are also presented in this paper.