AI-Enabled IT Procurement Categorization Framework

This white paper discusses the application of Artificial Intelligence (AI) in automating the categorization of IT procurement data using the Technology Business Management (TBM) taxonomy. It outlines the challenges faced by federal agencies in aligning procurement records with TBM categories, which often involve extensive manual analysis and can lead to inconsistencies. The paper presents a solution based on the BERT machine learning model, which captures the complex context of procurement data to enhance accuracy and efficiency. The BERT model demonstrated a 97% accuracy rate in classifying procurement data into TBM Sub-Towers, significantly reducing the time and effort required for manual classification. The document details the training process of the model, its performance evaluation, and the overall business impact of improved IT cost transparency. By leveraging AI, the proposed solution aims to streamline procurement processes, reduce errors, and ultimately save the government substantial amounts in IT expenditures.