In prospective memory (PM) research, costs (slowed responding to the ongoing task when a PM task is present relative to when it is not) have typically been interpreted as implicating an attentionally demanding monitoring process. To inform this interpretation, Heathcote, Loft, and Remington (2015), using an accumulator model, found that PM-related costs were associated with changes in a decision threshold parameter. This pattern was interpreted as disfavoring a monitoring process and supporting a non-capacity-consuming delayed responding strategy. The present study combined both behavioral and modeling techniques, as well as embedded parameter validation, to better illuminate the underlying processes involved in PM. We encouraged participants to use either a delayed responding or a monitoring strategy and used these conditions as anchor points for comparing a standard PM condition (with no strategy instructions). The monitoring strategy benefited PM more than did a delayed responding strategy. Most importantly, behaviors and modeling parameters associated with the standard PM instructions more closely reflected footprints of monitoring. Further, we found no individual model parameter that directly implicates monitoring behavior.