Patent attributes
Methods and systems for determining cadences are described. Behavior information between a user and a potential customer is stored, where the behavior information comprises actions between the user and the potential customer. The behavior information is analyzed to determine an engagement score to represent time effort the potential customer spends in interacting with the user, where the analyzing utilizes model-free reinforcement learning in a generalized Markovian process. A cadence is determined utilizing the engagement score, where the cadence comprises scheduled action items for the user for future interactions with the potential customer.