Manual order-picking systems (OPS) often suffer from inefficiencies that manifest as the seven forms of lean waste, including motion, defects, over-processing, and waiting. Consequently, order-picking accounts for approximately 55% of warehouse operational costs and 30–40% of total operational time. Automation emerges as a promising solution, yet its broad scope and diverse trade-offs leave businesses uncertain about where to begin or how to determine the right level of investment. This study addresses these challenges by classifying the levels of automation (LOA) in OPS into three categories: 1) paperless picking, 2) AGV/AMR-assisted picking, and 3) fully autonomous picking. Each category is explored in depth, highlighting the technologies and their applications. Furthermore, a decision-making framework is proposed based on the Full-Consistency Method (FUCOM) to guide businesses in selecting the optimal LOA tailored to their specific requirements. This framework is demonstrated through a case study on Calo, a direct-to-consumer foodtech startup, where criteria such as scalability and infrastructure flexibility were prioritized, resulting in the selection of AGV/AMR-assisted picking. The main contribution of this study is the clear classification of automation levels in OPS and a practical decision-making framework to guide businesses in optimizing their operations. This research concludes by recommending lean automation as an area for future research, emphasizing its importance in ensuring the selected LOA will not introduce waste during implementation.