With the increasing volume of data across various sources, Entity Matching (EM) has become crucial for integrating data sources to support decision-making and other applications. Despite technological advancements, current EM approaches offer only partial solutions, resulting in high manual effort and the need for specialized knowledge to apply these in practice. Furthermore, most approaches utilize benchmark datasets that do not comprehensively reflect real-world data heterogeneity, thereby affecting model robustness. This paper proposes a research design proposal for the development of an end-to-end EM solution for real-world entities, with a focus on product data. By addressing all EM process steps with customized and pre-configured models, this research aims to ensure more accurate and robust integration results while significantly reducing the need for manual effort and enhancing process automation. We discuss current challenges and highlight the research gap for an automated and tailored end-to-end solution for EM.