Abstract
Predictive Process Monitoring (PPM) forecasts the future behavior of ongoing business process executions to support proactive decision-making. Although state-of-the-art deep learning approaches achieve high predictive accuracy, they often act as black boxes, without explaining the execution steps that lead to a given outcome, thus limiting their applicability in scenarios requiring transparent what-if analysis. In this paper, we propose a planning-based framework for situation prediction, which anticipates not only whether a specific outcome is achievable but also the exact sequence of activities and data updates required to reach it. We introduce an automated pipeline that transforms event logs into PDDL planning tasks by combining process discovery with decision mining, allowing classical planners to generate compliant, goal-oriented predictions. Experiments on real-world logs show that our framework provides high solvability and predictive performance while offering transparency.