This chapter develops a framework for algorithmic governance, including considerations of the nature and consequences of the decision through processes of impact assessment. It analyzes ex ante AI design issues, such as mechanisms for sourcing technology through departmental or agency development or procurement processes, the calibration of error rates, level of human oversight, and participatory processes. Following this, the chapter considers the implementation of models of internal and external algorithmic auditing. The chapter then canvasses the trend towards centralized coordination of AI policies and principles across government, both through horizontal mechanisms of central agencies and regulators and vertical arrangements at the supranational and international organizational level imposed upon nation states. It also discusses transparency measures, including central publication of algorithmic tools, and individual notification and explanation of automated decisions.