Summary: The intention-superiority effect refers to the finding that intentions are more accessible than other memory contents. Our primary goal was to test for automatic processing in this effect, testing three features of automaticity: unintentionality, effortlessness, and lack of awareness. We used a postponed-intention paradigm with short action scripts. The intention-superiority effect was defined as greater accessibility in a lexical decision task (LDT) for words from to-be-performed scripts than to-beremembered scripts. Working memory load was experimentally manipulated to assess automatic processing. A general intentionsuperiority effect was found, demonstrating the automatic feature of unintentionality, and it was not diminished by a high load, demonstrating the automatic feature of effortlessness. Also, participants who reported that they lacked awareness of the link between the LDT and encoded scripts showed a larger intention-superiority effect than participants who were aware. Therefore, this study demonstrated an implicit intention-superiority effect, which was actually larger than the explicit effect. Copyright © 2012 John Wiley & Sons, Ltd. Consider the following common prospective memory scenario. During the work day, you form an intention to obtain something—such as cash from the cash machine—on your drive home. Although your attention is focused on other demanding tasks in the intervening period, when you later encounter a task-relevant cue (e.g. a bank), you retrieve the intention and successfully perform the task. In the current study, our purpose was to investigate one small piece of the prospective memory task, the representational state of intentions during the retention interval [i.e. after intention formation and before the performance interval (Ellis, 1996; Kliegel, Martin, McDaniel, & Einstein, 2002)]. More specifically, we investigated the intention-superiority effect (ISE), which refers to the finding that pending intentions to perform actions are more accessible than other memory contents (e.g. Goschke & Kuhl, 1993). This effect seems related to the Zeigarnik effect, the finding that participants have better recall for tasks that had been interrupted than for tasks that had been completed (e.g. Zeigarnik, 1939). Recent research