Generative Assistants for Handling Incidents in IT Service Management

This short paper analyzes how a generative AI assistant based on Retrieval Augmented Generation (RAG) and vector databases can improve incident management in IT service management (ITSM). The goal is to reduce escalation rates and to improve the efficiency of information retrieval in ITSM. It employs a Design Science Research framework to develop and evaluate an artifact that utilizes these technologies to enhance access to IT documentation. The evaluation included quantitative metrics (escalation rate, self-service resolution rate, first contact resolution rate, average resolution time) and preliminary testing with Microsoft Knowledge Base showed a 90.91% accuracy in identifying relevant documentation. These findings contribute to the understanding of AI applications in ITSM, advancing both academic knowledge and practical solutions for reducing incident escalations. By balancing theoretical rigor with practical relevance, we address significant business needs while enhancing the discourse on AI-assisted ITSM.