The present study implemented response time distribution modeling to better characterize context-specific attention dynamics underlying task interference due to possessing a prospective memory intention. During a three-phase paradigm in which prospective memory cues appeared only in the final phase, prospective memory performance was better when participants were informed at encoding of the context in which cues were to appear than when participants were not informed. Additionally, task interference increased during the third phase when the cue context was previously specified. Ex-Gaussian parameter estimates revealed that task interference during the third phase was due to a greater relative frequency of longer latencies, rather than an overall increase in latencies across all trials, suggesting that participants relied primarily on transient, rather than continuous, monitoring processes to support cue detection. Functionally, variability in transient and continuous monitoring profiles was predictive of prospective memory cue detection. More generally, the results from the present study suggest that ex-Gaussian parameter estimation procedures may provide a fruitful avenue for better understanding how attention is differentially allocated to ongoing tasks, what processes might underlie monitoring behavior, and how this behavior is related to eventual intention fulfillment.