Expert performers in many domains exhibit an uncanny ability to size up dynamic situations and to make good, fast, and largely unconscious decisions. These performers' cognitive processes can appear, to the performers as well as to observers, to be intuitive. This paper describes the instructional design theory of Expertise-Based Training (XBT) as a way to systematically train intuitive decision-making. XBT takes laboratory methods used in expert-novice research and repurposes them as computer-based training. In particular, XBT targets the recognition component of Gary Klein's Recognition-Primed Decision-Making (RPD) model. Recognition skills such as selective attention, pattern recognition, and situation awareness can be trained as a way of accelerating expertise in performance contexts that require quick decisions and immediate actions, including: emergency response, security screening, use-of-force, vehicle operation, interrogation, negotiation, medical and equipment diagnosis, and many others. XBT focuses solely on the recognition component of RPD, separate from skill execution, and can therefore be delivered on laptop computers and mobile devices. Despite being aimed at the still mysterious cognitive processes of intuitive decision-making XBT is in the tradition of Gagne and systematic design of instruction. XBT empowers an agile force by accelerating acquisition of the intuitive decision-making aspect of expertise that is increasingly acknowledged as both valuable and trainable.