What is it about?
Organizations often discover that a business process has gone wrong only after it has been completed. This paper presents FIREFOX, an AI-based approach that monitors processes while they are still running and predicts whether they are likely to end with an undesirable outcome. FIREFOX does more than issue an early warning. It generates counterfactual recommendations that describe how upcoming actions could be changed to steer the process towards a better outcome. The approach was evaluated on event logs from several business processes to explore its ability to support timely and informed interventions.
Featured Image
Photo by Campaign Creators on Unsplash
Why is it important?
Most predictive process-monitoring systems answer the question “What is likely to happen?”, but provide limited guidance on what people should do next. FIREFOX connects prediction with actionable recommendations. By identifying possible corrective actions before a process is completed, it could help organizations reduce delays, errors, unsuccessful outcomes and unnecessary costs. It also makes AI-supported decisions easier to understand by showing how alternative future actions could affect the predicted result. The initial results demonstrate the potential of the approach, while further studies in real operational environments will be important to assess its broader applicability.
Perspectives
Our goal was to move beyond simply detecting that a business process is at risk. We wanted to provide people with practical and understandable guidance on how that risk might be reduced. Counterfactual explanations make predictions actionable: instead of only saying that something may go wrong, they indicate which feasible future actions could lead to a better outcome. We are particularly honoured that this contribution received the Best Paper Award at the 28th International Symposium on Methodologies for Intelligent Systems — ISMIS 2026. Key takeaways Business process deviations can be predicted before a process is completed. FIREFOX combines early risk detection with counterfactual recommendations. The recommendations indicate how future actions could be changed to improve the expected outcome. The approach bridges predictive process monitoring, explainable AI and decision support. The paper received the Best Paper Award at ISMIS 2026.
Prof. Donato Malerba
Universita degli Studi di Bari Aldo Moro
Read the Original
This page is a summary of: Counterfactuals to Manage Ongoing Business Process Deviances, July 2026, Springer Science + Business Media,
DOI: 10.1007/978-3-032-32643-0_12.
You can read the full text:
Contributors
The following have contributed to this page
