Researchers have recently started to investigate how artificial intelligence (AI) and learning algorithms are impacting on work, organization and professional knowledge. Many analyses are focused on the deployment phase and practitioners’ reactions and challenges of integrating the new technologies into their existing practices. However, fewer studies have investigated the development and pre-implementation phase and how data analysts and computer engineers work to construct reliable systems and communicate their products to users and domain experts. This paper reports findings from an ongoing study on the development and implementation of AI in shipping. We illustrate the work done by technology providers in framing the system to intended user and how the relationship- and trust building of developers reduced users’ propensity of asking critical questions regarding the opacity of the system and its implications for organizational change and work.