Paper Number
ECIS2026-1602
Paper Type
SP
Abstract
Prescriptive process monitoring (PSPM) aims at recommending interventions that improve business process performance, yet the vast majority of existing approaches are correlation-based and unable to define the actions that actually cause better outcomes. This limitation is especially critical in IT service management (ITSM), where effective resource allocation is crucial for timely resolutions. Using a DSR approach, we develop a PSPM artifact that utilizes causal inference and double machine learning to provide KPI-oriented resource reallocation recommendations during runtime through conditional average treatment effect (CATE) estimation. The artifact combines offline causal modeling with online preprocessing to generate resource reallocation suggestions. A preliminary evaluation using real-world ITSM data from a company in the medical technology field shows initial promising results for causal resource reallocation. We outline a two-stage evaluation plan combining performance metrics, expert validation, and scenario-based KPI comparison to assess the system’s effectiveness.
Recommended Citation
Linkohr, Lisa-Marie; Bourguiba, Ahmed Rayen; and Liessmann, Annina, "Towards Causal Resource Reallocation In Prescriptive Process Monitoring" (2026). ECIS 2026 Proceedings. 4.
https://aisel.aisnet.org/ecis2026/bus_analytics/bus_analytics/4
Towards Causal Resource Reallocation In Prescriptive Process Monitoring
Prescriptive process monitoring (PSPM) aims at recommending interventions that improve business process performance, yet the vast majority of existing approaches are correlation-based and unable to define the actions that actually cause better outcomes. This limitation is especially critical in IT service management (ITSM), where effective resource allocation is crucial for timely resolutions. Using a DSR approach, we develop a PSPM artifact that utilizes causal inference and double machine learning to provide KPI-oriented resource reallocation recommendations during runtime through conditional average treatment effect (CATE) estimation. The artifact combines offline causal modeling with online preprocessing to generate resource reallocation suggestions. A preliminary evaluation using real-world ITSM data from a company in the medical technology field shows initial promising results for causal resource reallocation. We outline a two-stage evaluation plan combining performance metrics, expert validation, and scenario-based KPI comparison to assess the system’s effectiveness.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.