Static pile load tests have traditionally been considered the gold standard test and, if well executed, provide the reference load-movement response of the pile. Setting aside any difficulties with proper execution of static pile load tests, their primary deficiency is in the statistically insignificant rates at which they are undertaken — typically 0.5% to 2.0%. Furthermore, static pile load tests cannot be directly related to installation parameters and are therefore not well suited to development of driven pile acceptance criteria. If well executed, dynamic pile tests provide a rapid and generally reasonable estimate of pile load-movement response. The primary issue is that the static response is inferred from a dynamic response using simplistic models of complex dynamic pile-soil behaviour. However the advantages of dynamic testing are that it is generally performed on a statistically meaningful sample size – 5% to 15% in many cases – and it is concurrent with installation, which allows dynamic testing to be the basis for construction control and development of pile acceptance criteria. The remaining 85% to 95% of piles are necessarily installed using simple set criteria, or dynamic formula approaches which of themselves have significant deficiencies and represent project risk. Given that the foundation system will only be as good as the pile installed with the least confidence, improvements in foundation quality will be most effectively achieved by improvements in the monitoring and assessment of untested piles. This paper discusses a state-of-the art approach to reduction of overall foundation risk.