Abstract. Change design is one of the key steps within the IT change management process and involves defining the set of activities required for the implementation of a change. Despite its importance, existing approaches for automating this step disregard the impact that actions will cause on the affected elements of the IT infrastructure. As a consequence, activities that compose the change plan may not be executable, for example, due to runtime constraints that emerge during the change plan execution (e.g., lack of disk space and memory exhaustion). In order to address this issue, we propose a solution for the automated refinement of runtime constraint-aware change plans, built upon the concept of incremental change snapshots of the target IT environment. The potential benefits of our approach are (i) the generation of accurate, workable change plans, composed of activities that do not hinder the execution of subsequent ones, and (ii) a decrease in the occurrence of service-delivery disruptions caused by failed changes. The experimental evaluation carried out in our investigation shows the feasibility of the proposed solution, being able to generate plans less prone to be prematurely aborted due to resource constraints.