According to some embodiments, a risk relationship
data store may contain
electronic records, each electronic
record representing a risk relationship between an enterprise and a risk relationship provider (e.g., an insurer), and including, for each risk relationship, an electronic
record identifier and a set of estimated usage attribute values. A back-end application computer
server may receive, from a current
usage data source, current usage information for the enterprise (e.g., financial information, utility information, IoT information, etc.). Based on the current usage information, the computer
server may infer a likely actual current usage for the enterprise. The computer
server may then compare the likely actual current usage with the predicted usage attribute value to determine a risk difference result and adjust a risk relationship parameter based on the risk difference result.