Purpose <p>Digital agriculture technologies (DAT) could be used to reduce the environmental impacts of agriculture. However, their production, use and end-of-life outcomes generate their own environmental impacts. This paper evaluates the environmental impacts associated with the DATs, as well as the changes in farming practices they induce, to provide an integrated assessment of selected cases.</p> Methods <p>Three distinct DAT use cases were selected based on two criteria: their technological level (costs, complexity and user capability) and their transformative level (changes to the agricultural system). The case studies concern (i) irrigating an apple orchard using a soil moisture probe, (ii) mechanical weeding in a vineyard using an autosteering system and (iii) sowing and weeding in a green bean field using a multitask robot. Comparative life cycle assessments (LCAs) were conducted to evaluate the environmental impacts of these three digitalisation cases. System boundaries and functional units were defined according to the extent of the change in farming practices (e.g. changes in yields or not). Experimentations and interviews with farmers and DAT providers served as primary data sources, whereas LCA databases provided secondary data.</p> Results and discussion <p>Results reported that the technological level minimally influenced environmental impacts, as DAT life cycle contributions were negligible except in specific categories like mineral resource use for the autosteering system. Conversely, the transformative level significantly affected LCA outcomes. Non-transformative DATs (probe and autosteering system) optimised agricultural practices with minimal to moderate environmental impact changes. Transformative DATs (robot) amplified impacts due to yield losses resulting from operational limitations at the early adoption stage and more inputs per kilogramme of beans produced, indicating an incomplete appropriation process and the need for prospective LCA. Findings were sensitive to methodological assumptions, particularly system boundaries and functional units. Uncertainty analysis identified changes in farming practices as the primary source of variability in environmental impacts.</p> Conclusions <p>This paper concludes that the transformative level of DATs is the primary driver of environmental impacts, while the technological level plays a minor role. Additionally, these results underscore the importance of defining system boundaries broadly to encompass all changes brought about by digitalisation. Excluding these changes in farming practices could lead to a misrepresentation of the overall environmental impacts. Future research might benefit from adopting prospective approaches to examine the adoption of DATs over time.</p>