Molecular karyotyping has moved from bench to bedside for the genetic screening of patients with mental retardation and/or congenital anomalies. The commercial availability of high-resolution microarray platforms has significantly facilitated this process. However, the notion that copy number variants are also abundantly present in the general population challenges the interpretation of the clinical significance of detected copy number variants (CNVs) in these patients. Moreover, the awareness of incomplete penetrance and variable expression, exemplified by the inheritance of causal CNVs from apparently unaffected parents, has further blurred the boundary between benign and pathogenic variation. We analyzed 1001 patients using a large insert clone array (298 patients) and an oligonucleotide-based (703 patients) platform. In this cohort we encountered several examples of causal imbalances that could have been easily interpreted as benign variants when relying on established paradigms. Based on our experience and the pitfalls we encountered, we suggest a decision tree that can be used as a guideline in clinical diagnostics. Using this workflow, we detected 106 clinically significant CNVs in 100 patients, giving a diagnostic yield of at least 10%. Of these imbalances, 58 occurred de novo, 22 were inherited and 26 of unknown inheritance. This underscores that inherited CNVs should not be automatically disregarded as benign variants. Among the clinically relevant CNVs were 11 single-gene aberrations, highlighting the power of high-resolution molecular karyotyping to identify causal genes.