KPMG Switzerland observed that many companies are not fully utilizing AI's capabilities. The selection of spare parts by service technicians, often based on personal experience and vague customer descriptions, frequently results in inaccurate estimates, wasted time, increased costs, and reduced customer satisfaction. Integrating AI offers companies the chance to significantly enhance the accuracy and efficiency of planning and executing service tasks. By leveraging real-time data and advanced analytics, AI can streamline operations, reduce errors, and improve service quality.

The project followed a systematic approach, beginning with problem identification in field service management. An AI-driven system for inventory management was developed iteratively using prototyping, guided by Design Science Research methodology. The theoretical foundation combined Design Science Research and Machine Learning, ensuring alignment with real-world needs. Simulations tested the prototype’s ability to handle real-time data and complex decisions. The evaluation focused on functionality, usability, and performance to ensure the prototype met predefined criteria.

The prototype development achieved key milestones in integrating machine learning (ML) into SAP-based service order management. The project demonstrated the feasibility of using supervised ML models to predict spare part requirements from historical service data, enhancing inventory accuracy and improving technician preparedness, efficiency, and customer satisfaction. The results emphasized the need for robust data management practices to maintain model accuracy and reliability. Expanding the system's capabilities to handle more diverse and complex service conditions could further enhance its effectiveness, making it a versatile tool across various operational scenarios. Additionally, continuous optimization of algorithms and processes will be crucial for sustaining long-term benefits. Despite some limitations, the project laid a solid foundation for future scalability and enhancements, offering valuable insights into integrating ML within business processes. The outcomes highlighted both strengths and challenges, providing a roadmap for future projects to develop even more robust solutions adaptable across various industries.