This article is about augmenting a learning agent with “what if” thinking ability. Before effectively doing an action during exploration, the agent tries to predict what may be eventually observed. Thus, possible disastrous consequences can be avoided. Building on the potential of quantum computing, we explore how an agent can be augmented with predictive modeling. We introduce a quantum predictive modeling approach leveraging an extended Grover’s algorithm. The approach is illustrated using the metaphor of a cat and a mouse evolving in a grid environment.