The methods, apparatus, and systems described herein facilitate dynamic occupancy routing decisions. The methods include predicting a likelihood, based on day or time of day, that a customer having a retrieved or predicted profile is expected to initiate a customer communication; receiving, by one or more processors, the customer communication; providing a list of currently available agents and expected available agents, wherein the currently available agents are selected by excluding agents that (i) exceed a predetermined work threshold; and (ii) have exceeded a predetermined occupancy level; providing a routing recommendation to a communication distributor based on the predicted likelihood that the customer will initiate the customer communication on a day or at a time of day, and currently available agents' and expected available agents' proficiency at handling customers with the retrieved or predicted profile; and routing the customer communication via the communication distributor to an agent based on the recommendation.