Introduction The purpose of this study was to evaluate robotic-assisted cortical bone trajectory (CBT) screw placement. Early, middle, and late phases of robotic-assisted CBT screw placement were compared for accuracy and screw insertion time by comparing time and accuracy in every phase. Methods A retrospective review was conducted on the initial 40 patients who underwent spinal fusion using CBT screws in one institution from September 2021 to September 2022 utilizing a spine surgery robot system (Mazor X Stealth Edition, Medtronic Inc., Dublin, Ireland). The inclusion criterion was one- or two-level posterior lumbar interbody fusion (PLIF). Exclusion criteria were 1) patients who underwent posterior-lateral fusion in other segments, 2) patients who underwent additional decompression in other segments, 3) patients who underwent reoperation, and 4) patients with spondylolysis. The deviation of the CBT screw was evaluated on computed tomography (CT) one week after surgery using the Gertzbein-Robbins grade system. The rate of Grade A was considered the perfect accuracy rate, and the rate of penetration of 2 mm or more (Grades C, D, and E) was calculated as the deviation rate. To assess the learning curve, patients were divided into three groups. The first 10 cases were in the early phase group, the subsequent 10 cases were in the middle phase group, and the last 10 cases were in the late phase group. We compared the perfect accuracy rate, deviation rate, operative time, operative time per segment, intraoperative blood loss, registration time, and screw insertion time among the three groups. Results Thirty patients met the criteria. Overall, the perfect accuracy (Grade A) rate of the screw was 95.3% and the deviation rate was 1.4%. The perfect accuracy rate was 90.4% in the early phase, 95.5% in the middle phase, and 100% in the late phase. The deviation rate was 3.8% in the early phase, 0% in the middle phase, and 0% in the late phase, and there was no statistically significant difference between the three groups. Among the three groups, the operative time, the operative time per segment, the intraoperative blood loss, and the registration time were not significantly different. There was no significant difference in the screw insertion time among the three groups, but it decreased with experience (early phase: 156.9 ± 54.7 sec, middle phase: 139.9 ± 41.6 sec, and late phase: 106.4 ± 39.9 sec, p=0.060). The screw insertion time of the late phase tended to be shorter than that of the early phase (p=0.052). Conclusions The deviation rate of robotic-assisted CBT screw placement with one- or two-level PLIF was 1.4%, which was highly accurate. The deviation rate was 3.8% in the early phase, 0% in the middle phase, and 0% in the late phase. Although the deviation rate was low even in the early period, the screw insertion time in the early 10 cases tended to be longer than that in the late 10 cases. After passing the experience of 10 cases, this study concluded that robotic-assisted CBT screw placement was proficient.