Current genomics-driven precision oncology identifies actionable mutations in < 10% of cancer patients. Pediatric cancer is especially challenging due to limited mutations and fewer genomics-guided options. Functional precision medicine (FPM) addresses this by integrating genomic profiling with rapid, high-throughput functional ex vivo drug testing on live patient-derived cells. However, there is lack of FPM prospective data showing clinical utility in pediatric cancers. In this prospective, non-randomized, single-arm study (NCT03860376), we investigated feasibility and impact of FPM in pediatric/adolescent with refractory/relapsed solid and hematologic cancers. Of 25 patients, 19 (76%) had FPM data reviewed by the FPM tumor board within four weeks (FPMTB), meeting the primary outcome of the study. Additionally, six patients received FPM-guided treatment. Among these 6 patients, 83% (5 patients) experienced a greater than 1.3-fold improved progression-free survival compared to their previous therapy, and together demonstrated a significant increase in progression-free survival and objective response rate versus physician’s choice-treated patients (8 patients). Post-hoc analysis showed that patients with the same subtype of cancer do not cluster together, reinforcing the concept of optimizing cancer treatments one patient at a time (n-of-1 approach). Additionally, our study used a novel artificial intelligence/machine learning (AI/ML) platform that leveraged drug responses and sequencing data to identify novel biomarkers of drug efficacy and gain potential mechanistic insights within specific subsets of pediatric cancer patients. The findings from our proof-of-principle study illustrate the impact of FPM for relapsed/refractory pediatric/adolescent cancer patients, highlight future integrations of FPM and AI/ML, and support ongoing patient cohort expansion (NCT05857969).