Background Preoperative 3D reconstruction in hepatobiliary surgery has reported benefits in surgical planning and outcomes; however, most studies examine the impact by expert users with routine institutional use. In Australia, this is not the case. We present the results of a trial of 3D software, comparing anatomical impressions and surgical plans obtained pre and post 3D reconstruction, in a cohort of liver surgeons that did not have routine access to software. Methods The trial was administered at a national liver surgery conference following a pre-trial email survey. Participants consisted of opportunistically recruited liver surgeons, freely allocated to one of four computers uploaded with CT data from real-world liver tumour cases. Participants reviewed 2D images and recorded their tumour/segmental involvement and surgical plan. An interactive 3D model was then automatically generated using a proprietary software; participants re-recorded their responses following review of the 3D model. Paired responses were qualitatively assessed and coded for magnitude and directionality of change. Results Low national rate of routine 3D software use (9%) is confirmed in pre-trial survey. 37 surgeons provided 48 responses. Univariate analysis demonstrated overall significant rates of response changes for tumour segmental impressions (54.2%, p<0.001), surgical plans (57.1%, p<0.001) and structures targeted for division (56.3% p<0.01). In one case, mis-segmentation or ‘hallucination’ of a major portal branch went unrecognised by 7 of 9 participants. Conclusion We demonstrate that 3D visualisation can impact surgical planning in a population of non-routine users, however the dangers of fully automated 3D segmentation was highlighted as well. While beneficial, routine use of 3D technology in surgical planning mandates high level anatomical familiarity and verification of model accuracy by multiple users, with additional human annotation required during the introductory phase.