Safety is of utmost importance in urban road traffic. Many studies reveal that speed is the major factor contributing to road accidents. In the Netherlands, 50% of the speed offences and 65% of the accidents occur on urban roads. Research by the Dutch Ministry of Transport indicates that the elimination of speeding would reduce accidents and fatalities by 21% and hospitalized victims by 15%, and provide benefits to the Dutch economy of 300 million euros per year. Implementation of Intelligent Speed Adaptation (ISA) seems to be a promising way to meet these targets. Although ISA is a tested and proven technology, its implementation is impeded by many uncertainties, including its acceptance by drivers, its effects on driving behavior, and its effectiveness in real world driving conditions (as opposed to 'laboratory' conditions). As a result, policymakers need a way to implement ISA so that its safety benefits can be obtained despite the existence of uncertainties that cannot be reduced in advance. This paper proposes an innovative, adaptive approach for implementing ISA that would permit implementation to proceed while knowledge was gained about the uncertainties over time (i.e., the level of uncertainty was reduced). Under the adaptive approach, actions to deal with the certain vulnerabilities of the policy are implemented at the same time as the policy, and uncertain vulnerabilities are monitored and actions taken as required. This approach allows adaptations to the policy over time as knowledge about ISA is gained and critical events for ISA implementation take place. In particular we propose to begin by implementing ISA (Dynamic Voluntary) for young car drivers between 18-34 years of age and on unsafe roads i.e. urban roads (30 to 80km/h) and to prepare to adapt the policy over time.