Purpose The purpose of this study was to quantify accuracy of a recently FDA-approved robotic-assisted device. Methods Thirty-seven patients underwent TKA with the Robotic Surgical Assistant (ROSA) by the same operating surgeon and team over the course of 3 months. Intra-operative mechanical axis measurements, composed of alpha (α), beta (β), gamma (γ), and delta (δ) angles, and the hip-knee-ankle angle (HKA) were calculated by the ROSA. Post-operative mechanical implant angles were taken from 36″ stitched post-op films and measured in the PACS imaging system. Accuracy was assessed by comparing the percentage of postoperative long length films within 2° and 3° of the ROSA intra-operative plan. Results The ROSA system accurately calculated the HKA, α, and β angles (95% CI), but was inaccurate in calculating both γ and δ angles. Using a window of ± 3° accuracy, the HKA, α and β angles were accurate at levels of 89, 100 and 92% respectively. In contrast, the sagittal relationships were considerably less accurate at 77 and 74% for the γ and δ angles respectively. Subsequently, the proportion of cases within 2 and 3 degrees of the intra-operative plan for resection angles was considered accurate for HKA (73% within 2°, 89% within 3°), α (92% within 2°, 100% within 3°), and β (76% within 2°, 92% within 3°) angles, but considered inaccurate for γ (51% within 2°, 77% within 3°) and δ angles (57% within 2°, 74% within 3°). Conclusions This study demonstrated that while the ROSA system seems to accurately predict coronal plane resections in TKA, it falls short in the sagittal plane. Further research in these deficiencies can provide insight into the overall efficacy of robotic assisted surgery in TKA. Level of Evidence Level III Therapeutic Study.