Generative AI-driven BoM analysis for multi-criteria supplier selection

This thesis presents the development of a generative artificial intelligence system designed to automate supplier selection within the context of Bill of Materials (BoM) optimization. The proposed framework addresses multi‐criteria evaluation of industrial components by implementing the Analytic Hierarchy Process (AHP). For each BoM item, eight critical parameters are assessed: cost, delivery time, material quality, environmental impact, obsolescence risk, technical compatibility, volume, and weight. The system architecture comprises three layers: the user interaction layer, the generative AI layer, and the data management layer. When user uploads the BoM, critical components are automatically identified by the application of AHP. For each identified component, AI web scraping extracts supplier data; missing information triggers automated email requests to suppliers. When all data is available, user defines the weights of parameters and the AI perform a comparative analysis, yielding a final report that details trade‐offs and recommends the optimal supplier for each critical component. Key innovation lies in the integration of web mining, automated communication and multi‐criteria decision analysis into a unified framework. Experimental results on an industrial case study demonstrate significant reductions in analysis time and improved procurement decision quality.