Computer-assisted interpreting (CAI)—the use of support tools powered by automatic speech recognition and artificial intelligence during simultaneous interpreting—may be regarded as the epitome of ‘skill of the future’ for interpreters. CAI tools bear the potential to increase output quality and productivity and help interpreters cope with the demands of a fast-changing and increasingly technologized profession. Despite a growing demand for CAI training, no research has been conducted to specify the necessary skills and instruction. The present dissertation aims to identify crucial CAI skills and develop research-based recommendations for the design of CAI training. A Cognitive Task Analysis (CTA) was conducted on data gathered in two empirical cycles exploring experienced simultaneous interpreters’ performance, behaviour, and perception in a CAI test. The key outcomes of the CTA are a model of the CAI task, an inventory of CAI skills specifying their cognitive nature, and the analysis of the knowledge and attitudes required to successfully acquire and perform those skills. Starting from CTA results, recommendations for CAI training were developed through the application of the Four-Component Instructional Design Model (4C/ID). The study informs CAI training practice, expands the current scientific understanding of this complex task, and provides methodological inputs for future educational research on interpreting technology and beyond.