Downloads: 0
Research Paper | Computer Science and Information Technology | Volume 15 Issue 5, May 2026 | Pages: 1009 - 1014 | United States
Natural Language Processing for Intelligent HR Case Routing in ServiceNow HRSD: Reducing Manual Triage Through Transformer-Based Classification
Abstract: Human Resource Service Delivery (HRSD) case management in ServiceNow depends heavily on manual triage to classify and route employee-submitted cases to appropriate HR teams. This manual process introduces routing errors, processing delays, and inconsistent handling quality that disproportionately affect the employee experience at scale. This paper proposes and evaluates an NLP-driven intelligent routing framework that applies transformer-based text classification to free-text HR case submissions within ServiceNow HRSD, automating the triage function with measurable improvements in routing accuracy and processing speed. The framework fine-tunes a RoBERTa-base model on a corpus of 48,420 anonymized HR case records drawn from four enterprise ServiceNow HRSD deployments, classifying cases across eleven HR routing categories. Evaluation demonstrates that the fine-tuned model achieves 91.4% top-1 routing accuracy and 97.8% top-3 accuracy on held-out test data, compared to 73.2% accuracy for the rule-based keyword routing currently deployed in the study organizations. Integration with ServiceNow's NowAssist platform via a custom REST API achieves a mean routing latency of 340 ms- below the 500 ms threshold for a transparent user experience. A ServiceNow HRSD integration architecture, a model governance framework, and a bias auditing protocol are presented, addressing the fairness and explainability requirements specific to HR automation contexts.
Keywords: NLP, ServiceNow HRSD, HR Case Routing, RoBERTa, Transformer Classification, Employee Experience, Intelligent Automation
How to Cite?: Abhinav Reddy Pullikallu, "Natural Language Processing for Intelligent HR Case Routing in ServiceNow HRSD: Reducing Manual Triage Through Transformer-Based Classification", Volume 15 Issue 5, May 2026, International Journal of Science and Research (IJSR), Pages: 1009-1014, https://www.ijsr.net/getabstract.php?paperid=SR26515093008, DOI: https://dx.dx.doi.org/10.21275/SR26515093008