無人商店技術及應用之成熟度研究

Abstract

近年來無人商店在全世界便掀起了一波浪潮，雖然無人商店創造了良好的消費體驗，但其科技應用組合複雜，且代表性案例及相關研究尚少，企業打造無人商店時該應用何種的科技組合以達目的便成了一大課題。\n 本研究之目的便是透過成熟度模型，建構商店無人化科技應用的具體途徑，使企業在打造無人商店時對於科技應用有一條清楚的道路可以參考。本研究採用文獻分析法，蒐集包含台灣超商龍頭、無人商店先驅等十個超商及超市的代表性案例，並以四大科技應用構面－消費者身份辨識、商品辨識與結帳、數據蒐集與分析、商店日常維運，闡述各案例在無人化時的科技應用樣貌，以歸納無人化程度不同的商店在科技應用上的差異。而後，依據上述案例資料統整四大科技應用構面之發展現況及未來方向，並作為建立成熟度模型的依據。最後，統合上述兩者建立無人化科技應用程度由低至高之成熟度模型，並包含各成熟度階層之科技應用樣貌與通往下一階層之方式。\n 本研究結果將商店無人化科技應用成熟度劃分為Level 0至Level 4的五個階層，其中Level 0在商店營運流程中完全依賴人力；Level 1在營運上雖主要仍依賴人力，但導入POS、會員制度等科技減輕店員負擔；Level 2則在大部分營運環節達到無人化，但因AI及演算法不成熟導致消費流程及數據分析仍有不足；Level 3在AI及演算法等科技已達成熟，消費流程也鮮有錯誤發生，但在日常維運如補貨上架仍須依賴人力；Level 4則除了在各項科技及流程爐火純青外，更達到上架補貨無人化，進而使商店所有營運完全無人化。\n 從研究結果發現商店無人化的過程中，通往下一階層需具備特定的條件，如較低階層的關鍵為消費者身份辨識，而高階層則為AI及演算法等科技，本研究之模型可以幫助企業檢視所需補強的科技，避免無效科技應用的發生。

Unmanned stores have been a trend in the world in recent years. Although it creates great consumer experience, the approaches of integrating technologies are dynamic, and there are few successful cases and related research. What kind of technologies should be used when enterprises build unmanned stores in different circumstances has become a research topic.\n The purpose of this research is to develop a maturity model of unmanned store technologies application as well as building a path way in the model. So, enterprises can have a clear reference when building unmanned stores. This study adopted qualitative analysis approach and collected successful cases of ten convenient stores and supermarkets including Taiwan`s convenient stores leader and unmanned stores pioneers, and clarifies the applications of technologies in the unmanned case by using four technologies applications facets: 1.Consumer Identification, 2.Product Identification and Checkout, 3.Data Collection and Analysis, and 4.Daily Maintenance, to summarize the differences in the applications of technology with different degrees of unmannedness. Then, the four technologies applications facets as shown above were used as the basis for establishing the maturity model. Finally, we established a low-to-high maturity model of unmanned stores technologies application, including the descriptions of the companies in each level and the way to the next level.\n The results of this research divide the maturity model into five levels from Level 0 to Level 4. The company in Level 0 relies on manpower in the store operation process entirely; The company in Level 1 still relies on manpower in its operation, but imports technology such as POS and membership system to reduce the loading of clerks; The company in Level 2 becomes unmanned in most of the operations, but the consumption process and data analysis are still insufficient due to the immature of AI and algorithm; The company in Level 3 has matured AI and smart algorithm. There are only few mistakes in the operation process, but in daily maintenance, such as replenishment, is still based on manpower; The company in Level 4 is not only high degree mature in technologies and process, but also achieves unmanned in daily maintenance, thus making the store fully operational unmanned.\n The results of this research found that in the process of unmanned stores maturity model, specific conditions are required for the next level. For example, the company in Level 0 and Level 1 required the adoption of consumer identification technologies to promote to the next level , while the company in Level 2 and Level 3 required more mature AI and algorithm to do so. The maturity model of this research can help enterprises view the technologies needs to reinforce and avoid the inefficiency of technologies applications.