Sensors in the smart grid provide a way for utilities to obtain measurements that can be used to assess grid health. However, the vast majority of physical grid assets do not have sensors attached to them. For example, a utility may have hundreds of thousands of poles, many of which are part of an aging physical infrastructure. The utility, however, does have a wealth of information concerning pole material composition, age and location among other attributes as well as domain knowledge about pole operation and power network connectivity. By integrating all of these different sources of information in a systematic way, we show how to make effective and efficient decisions about poles. In this paper, we describe a decision support framework to identify utility poles that may require maintenance or may be at risk. First, we process attribute data from GIS and inspection records and classify poles into 'at-risk' categories. Then, we derive a health score or failure probability of a pole based on policies and constraints. Finally, the failure probabilities are inputs to consequence analysis, where we calculate individual pole impact, which is the expected lost value if a pole fails. We argue that this approach to risk analysis is holistic, capturing interdependencies across the power network, and that it scales across asset classes. The uses for risk management of assets are many, ranging from inspection and servicing to capital planning and rate case preparation.