Road safety is high on the agenda of transport policymakers. When looking for policies to improve road safety, more focus should be placed on urban roads, since the accident risk is greater for urban roads than for other types of roads. One of the major contributing factors to urban road accidents is speed. Intelligent Speed Adaptation (ISA) uses an in-vehicle device to monitor local speed restrictions and support the driver in adjusting his driving speed to the posted maximum speed. The technological feasibility and potential safety impacts of ISA have been documented. Policymakers and others are now exploring possibilities for implementing ISA on a large scale. Policymaking, however, is restrained by uncertainties about the real-world impacts of ISA implementation. Concerns about acceptance among drivers, technological failures, unexpected costs, etc., are causing delays in ISA implementation. In this paper, we describe an innovative policy analysis approach that can take account of these uncertainties. In particular, using this approach, the various possible uncertainties surrounding ISA implementation are modelled and explored in-depth. Using minimization of maximum regret as a criterion, the effectiveness of six alternative ISA policies is shown across a wide spectrum of plausible futures. Our initial findings offer some counter-intuitive results concerning uncertainties related to the effectiveness of ISA policies. For example, policies to have the ISA system installed in more vehicles do not necessarily lead to a higher reduction in accident risk. We also show how to deal with uncertainty about the relationship between speed and accident risk.