System and method for managing routing of leads

A processor-based method comprising:

• executing, by a processor, a predictive machine learning model configured to determine an assigned agent cluster and an individual assigned agent from the assigned agent cluster for each lead record of a plurality of lead records stored in a lead model database, by applying a combination of an unsupervised clustering analysis, a supervised clustering, and an aggregate conversion model, to lead data representative of leads and agent model data representative of agents, the predictive machine learning model outputting a plurality of agent clusters and sales conversion scores, wherein each of the sales conversion scores is representative of a likelihood that a respective lead will accept an offer to purchase a product from a respective agent from the assigned agent cluster within the plurality of agent clusters;
• wherein execution of the aggregate conversion model using the agent model data as input generates a score for each of the agents and for each lead as output, the scores provided as input to the supervised clustering to generate the plurality of agent clusters for the agents,
• wherein the unsupervised clustering analysis determines a clustering vector for each of the agents from the agent model data and clusters each of the agents into one of the plurality of agent clusters based on degree of similarity of the clustering vector, and
• wherein assignment of each of the leads to one of the agents comprises execution of one of a greedy model that determines the one of the agents that has a highest conversion score for that lead, or Bayesian Bandits heuristics that assigns each of the leads to the one of the agents in order to maximize cumulative expected conversion scores;
• and
• running, by the processor, the predictive machine learning model on demand to update and display, by a display device in operative communication with the processor, a graphical user interface (GUI) including information representative of the plurality of agent clusters and assignments of respective leads to respective agent clusters within the plurality of agent clusters and to respective individual assigned agents within the respective agent clusters for contacts with the respective leads.

A system comprising:

• non-transitory machine-readable memory that stores lead model records for a plurality of leads, sales records for a plurality of sales, and agent model records for a plurality of agents, wherein the agent model records comprise a flattened data structure comprising a single record for each of the plurality of agents;
• a predictive machine learning model configured to determine an assigned agent cluster for each lead model record for the plurality of leads of an enterprise by applying a combination of an unsupervised clustering analysis, a supervised clustering, and an aggregate conversion model;
• and
• a processor in operative communication with a display, wherein the processor in communication with the non-transitory, machine-readable memory and the predictive machine learning model executes a set of instructions instructing the processor, for each agent model record, to:
• apply the unsupervised clustering analysis to lead model data, the sales data, and agent model data to determine a plurality of agent clusters by determining a clustering vector for each of the agents in the agent model data and including each of the agents in one of the plurality of agent clusters based on degree of similarity of the clustering vector, wherein each assigned agent cluster is representative of a respective group of two or more of the plurality of agents assigned for contacts with a lead represented by each lead model record;
• apply the supervised clustering to the aggregate conversion model to assign each of the leads of the enterprise to one of the plurality of agent clusters determined by the clustering analysis;
• output the plurality of agent clusters, the agents included in each of the plurality of agent clusters, and the assignments of each of the leads of the enterprise to one of the plurality of agent clusters wherein execution of the aggregate conversion model using the agent model data as input generates a score for each of the agents and for each lead as output, the scores provided as input to the supervised clustering to generate the plurality of agent clusters for the agents;
• and
• run the predictive machine learning model on demand to update and display, by a display device in operative communication with the processor, a graphical user interface including information representative of the plurality of agent clusters, the agents included in each of the plurality of agent clusters, and the assignments of each of the leads of the enterprise to one of the plurality of agent clusters for the contacts with the respective lead.