In this paper I explore the production of expertise and its relationship to power, especially by focusing on everyday experiences in the Abadan refinery in the southwest Iranian province of Khuzistan. I argue that workers’ agency in the Abadan refinery, despite in practice being important to the running of operations, was simultaneously distinguished from and obfuscated by a process of knowledge production that constantly (re)produced a separate domain of technical expertise. I demonstrate that although workers were able to contest management’s definitions of ‘technical’ work (which was juxtaposed to ‘labour’), they were systematically denied a place in the daily production of the refinery as a space of expertise. Thus, we see a process that others such as Scott and Mitchell have identified in other technocratic projects: the subordination of localised agencies in favour of more abstract, universalising forms of knowledge. In addition, I highlight how this was especially visible in the Anglo-Iranian Oil Company’s PR, which, in depicting the refinery through its technological edifices, represented it as a self-regulating system devoid of the subjectivities of human error. I then follow the effects of such representation as they played out in two spheres. Firstly, I show how the Abadan refinery became situated in the realm of established knowledge predominating within the global oil industry, figuring as a mobile assemblage capable of transportation to any locality. Secondly, I point to the perception of the Abadan refinery in public discussion within Iran itself in order to help explain knowledge production in the Iranian oil industry post-nationalisation. My paper will have a strong historical methodological focus, making use of subaltern and microhistorical approaches to address how historians can uncover quotidian agency from below from mainstream representations of the refinery and official company documents – my two main types of sources. However in my theoretical framework I also use both critical geography concerning the production of space, as well as the sociology and ontology of Bruno Latour in order to understand how human and non-human agencies are involved in the construction of expertise as a bounded, objective domain.