AI Pipeline Transforms ITSM Data into Actionable Intelligence.

IT Service Management (ITSM) systems typically generate vast amounts of diverse ticket data, which often proves challenging for sales and executive teams to translate into meaningful, actionable insights. To address this, researchers have designed and evaluated a sociotechnical AI pipeline aimed at transforming raw ITSM exports into decision-support artifacts at multiple levels. The pipeline integrates several AI techniques: LLM-based schema normalization, HDBSCAN for sub-topic clustering, and hierarchical agglomerative clustering. This combination allows it to generate both high-level executive-facing Main-topics and more granular Sub-topics. A stakeholder evaluation involving sales engineering and customer success roles across various artifacts showed that key decision-support metrics—interpretability, actionability, trust, and likelihood of use—all averaged above 4.0 out of 5.0, with trust being the most consistent positive signal. The findings underscore ITSM analytics as an information systems challenge requiring transformation, abstraction, and human-centered design.