ERP system implementation is complex and costly, with software modifications tailored to organizational needs being one of the most challenging stages. The scope and complexity of these modifications depend on the type of project-whether it is a first-time implementation, an upgrade, a roll-out, or a reimplementation. One of the key directions in the development of ERP systems is the integration of AI solutions, both in the context of the company’s day-to-day operations and in the automation of repetitive tasks within the implementation process itself. However, there is still a lack of tools that effectively support the automation of implementation activities. The aim of the presented research is to increase the effectiveness of ERP implementation projects through the use of AI-based tools. A solution that leverages NLP techniques to enable the reuse of modification documentation (HLS/DLS) by performing similarity comparisons and searches is proposed. The paper also includes a systematization of ERP project types, an analysis of modification similarities, and a review of implementation methodologies used by various ERP vendors (SAP, Microsoft, IFS), with particular emphasis on the impact of project type on the scope of modifications.