The thesis addresses the implementation challenges of Machine Learning (ML) for merchandisers in the scenario of digitalization of retailing, and proposes a product-service design as the solution. The digitalization of retailing is defined as an on-going process to integrate Internet-connected digital technologies into interfaces between retailers and consumers. The researcher collaborates with Bloomreach, who provides an ML-powered merchandising tool called Bloomreach Search & Merchandising (brSM), and uses the context as an example of digitalization of retailing with ML. brSM helps merchandisers to improve the search and category experiences by optimizing the ranking of products, improving search results and curating recommendations on e-commerce platforms. The project presents a comprehensive analysis of the product, service, and merchandiser. In the product analysis, it is suggested that brSM doesn’t facilitate the interaction between the merchandisers and algorithms. Due to the knowledge gap, merchandisers have difficulties to align the expectation of the product at the beginning. Furthermore, the product doesn’t provide proactive feedback that improves the supervision of the user. In the service analysis, the misalignment of the internal feature communication leads to the confusing implementation service for the merchandisers. Specifically, the internal workflow and communication during the new feature introduction are confusing internally and externally. In the merchandiser analysis, it identifies two personas of merchandisers during the implementation of ML due to different business contexts and product characteristics. It thus is suggested to provide customized implementation supports according to their different needs.To address these challenges, the design solution aims to improve the (new) feature communication by adopting a use-case oriented approach for merchandisers and internal stakeholders with supportive tools. Based on the implementation framework of service design, the solution will be addressed on three levels, experience, service and strategy. At the experience level, brXtrategy family, supportive tools that provide merchandising inspirations, is introduced. It provides customized implementation information according to merchandisers’ business context. Also, it simplifies the product information by the adopting use-case oriented approach, which provides example-based explanations. Moreover, it improves the interaction between the merchandisers and algorithms by an interactive education tool and proactive notification of algorithmic performance. The front-stage and back-stage services are illustrated by the user journey map and service blueprint, which specify methods to improve the intra-company collaboration and the customer services in the critical moments like new feature introduction, onboard, and re-training. On the strategy front, a roadmap and a transitional workflow are introduced to facilitate the product strategy and the solution implementation. The workflow, called Use-case oriented development workflow, bridges the gaps of product/merchandiser understandings between the field teams and the R&D teams during feature developments. With comprehensive research and three aspects of the design solution, the thesis contributes to the company and academic domain. It contributes to a better understanding of merchandisers in the process of digitalization of retailing. Also, the solution improves brSM’s services and facilitates the implementation of ML. Last but not least, it demonstrates a design approach that designers can perform that improves ML-powered products.