This study investigated the influence of layer height, outer boss diameter, and taper angle on the axial pull-out performance of commercial M3 brass heat-set inserts installed in FDM-manufactured Z-PETG bosses. A three-level full factorial design with replicated centre points resulted in 31 pull-out tests. Maximum pull-out force was evaluated using complete quadratic response-surface regression and analysis of variance. Recorded crosshead displacement at failure and visually observed failure modes supported the mechanical interpretation. Outer boss diameter was the only statistically significant individual model term, and larger diameters increased the fitted maximum pull-out force within the investigated range. Layer height showed a negative tendency but did not reach statistical significance. Neither taper angle nor the quadratic and two-factor interaction terms were statistically significant. The complete quadratic model was not statistically significant and showed weak predictive performance. It was therefore used only to describe experimental tendencies within the investigated domain. Failure analysis indicated that cracking at the boss base and complete boss separation were the dominant failure modes, whereas direct insert pull-out produced a more progressive response. The results show that boss design should consider maximum pull-out force, displacement capacity, and failure mode together.