In this paper we investigate the use of human movement qualities and the design of intelligent products. Our future products and systems are envisioned to become context-aware and adaptive. The design of these adaptive products brings new opportunities to the design of interactive products. Self-adaptivity of products depends on their ability to learn through interaction with the user. We explored a research-through-design process that revolves around a product which is able to interpret human movement qualities. In our approach we integrated three fields: Laban Movement Analysis, neural learning and interactive product design. In this paper, we explain our approach to design adaptive interactive products, and describe the resulting walk-in closet research platform. We present the choices and findings, show results of initial user-testing of the prototype, discuss the open questions that this innovative design approach raised, and further research possibilities.