Generative Artificial Intelligence (GenAI), by enabling the transfer of tasks associated with human cognition to machines, is taking centre stage in public debate and corporate strategies. However, empirical studies conducted in real-work settings remain limited, with findings largely focused on quantitative indicators. In order to fill this gap in empirical research on the use of GenAI in work contexts, we conducted a study within an in-house trial of a GenAI tool in a major French company. Our objective was to explore how the integration of a GenAI-based tool within an existing sociotechnical system reconfigures employees’ work practices. Based on 9 in-depth re-situating interviews with material traces (self-documented logbooks), our results highlight (1) how workers adapt to the new interaction modality, (2) the varied way they integrate GenAI in their work and (3) the challenges and new tasks associated with the use of the system.