<p>The digitalization of agriculture is bringing about far-reaching socio-technical changes. This article analyzes these changes by looking at farming habits. Transactional theory of learning is introduced as an analytical perspective for investigating farmers’ consideration processes and experimentation with potential habit changes related to the use of digital technologies. The analytical perspective is applied to a case study of robotic weed control in sugar beet cultivation in northeastern Germany. The study shows how prevalent habits are crucial anchor points in farmers’ careful considerations of whether to use a field robot: habitual beliefs such as an excitement for robots, as well as professional and private habits and resulting free and committed capacities are included in these considerations. Experiences with technology use and weather-related uncertainties, furthermore, led to a habitual risk assessment and anticipatory solution seeking as an overarching element in the formation of new habits. In addition to the empirical study, the article aims to make a methodological and conceptual contribution to advance much-needed research on change of (agricultural) practices and the role of technologies in it. In this regard, the use of transactional theory of learning is discussed by reflecting on the kind of knowledge that can be produced with this type of analysis and how it can benefit research on practice change and broader (agricultural) transition processes.</p>