A prospective cohort of 116 IVF/ICSI cycles compared a 25-point clinical/laboratory KPI score with an AI blastocyst-image algorithm (Embryo/Aid-MIM Fertility) for predicting implantation and live birth. AI outperformed the KPI score (AUC 0.80 vs 0.66; accuracy 73% vs 63%), though the single-centre design and small sample limit the conclusions.