Event-based prospective memory (PM) is the ability to remember to perform an intention in response to an environmental cue. Recent microstructure models postulate four distinguishable stages of successful event-based PM fulfillment. That is, (a) the event must be noticed, (b) the intention must be retrieved, (c) the context must be verified, and (d) the intended action must be coordinated with the demands of any currently ongoing task (e.g., Marsh, Hicks, & Watson, 2002b). Whereas the cognitive processes of Stages 1, 2, and 3 have been studied more or less extensively, little is known about the processes of Stage 4 so far. To fill this gap, the authors manipulated the magnitude of response overlap between the ongoing task and the PM task to isolate Stage-4 processes. Results demonstrate that PM performance improves in the presence versus absence of a response overlap, independent of cue saliency (Experiment 1) and of demands from currently ongoing tasks (Experiment 2). Furthermore, working-memory capacity is associated with PM performance, especially when there is little response overlap (Experiments 2 and 3). Finally, PM performance benefits only from strong response overlap, that is, only when the appropriate ongoing-task and PM response keys were identical (Experiment 4). They conclude that coordinating ongoing-task and PM actions puts cognitive demands on the individual which are distinguishable from the demands imposed by cue-detection and intention-retrieval processes.