Dynamic analytics and forecasting for messaging staff

Abstract

Systems and methods are provided for dynamic generation of staff analytics and forecasts based on skill and service level. Dynamic forecasting allows for forecast generation in real-time and may be based on historical data regarding skills and results, as well as data science to identify patterns and make predictions. The resulting staffing forecast may therefore provide for efficient management of messaging staff costs while preserving the desired service quality. The staffing forecast may include a volume forecast that is tailored to the unique nature of asynchronous messaging, as well as the unique messaging needs of the entity so as to efficiently manage messaging operations and make data-driven staffing decisions that take service level into account. An exemplary embodiment may include dynamic analytics tools that may use specified target and/or resource numbers (e.g., desired service level) for an existing messaging operation and get a detailed per-skill staffing forecast.

1 . (canceled) 
     
     
         2 . A computer-implemented method comprising:
 obtaining historical messaging data, wherein the historical messaging data includes skills exercised and customer satisfaction levels for a plurality of messaging conversations;   filtering the historical messaging data based on a particular skill to obtained filtered historical messaging data;   training a learning model using the filtered historical messaging data to obtain a trained learning model, wherein the trained learning model is trained to generate a staffing forecast for the particular skill at a target service level corresponding to a particular customer satisfaction level;   receiving a request for an updated staffing forecast corresponding to the particular skill and a desired target service level;   generating a prediction using the trained learning model, wherein the prediction corresponds to a projected staffing level for the particular skill and the desired target service level; and   providing the projected staffing level as the updated staffing forecast.   
     
     
         3 . The computer-implemented method of  claim 2 , further comprising:
 obtaining, from a messaging communication in real-time, a skill exercised during the messaging communication and a customer satisfaction level corresponding to the messaging communication, wherein the skill exercised is the particular skill.   
     
     
         4 . The computer-implemented method of  claim 2 , further comprising:
 adding a skill corresponding to the particular skill and a customer satisfaction level obtained from a messaging conversation in real-time to the filtered historical messaging data to obtain updated filtered historical messaging data; and   re-training the trained learning model using the updated filtered historical messaging data to obtain a re-trained learning model, wherein the re-trained learning model is used to generate projected service levels for future requests for updated staffing forecasts.   
     
     
         5 . The computer-implemented method of  claim 2 , wherein the historical messaging data corresponds to messages related to a particular entity. 
     
     
         6 . The computer-implemented method of  claim 2 , wherein the historical messaging data corresponds to messages related to an aggregation of entities. 
     
     
         7 . The computer-implemented method of  claim 2 , wherein generating the prediction using the trained learning model includes analyzing an arrival rate parameter. 
     
     
         8 . The computer-implemented method of  claim 2 , wherein generating the prediction using the trained learning model includes analyzing an agent net handle time parameter. 
     
     
         9 . The computer-implemented method of  claim 2 , wherein generating the prediction using the trained learning model includes analyzing a particular time period, and wherein the projected staffing level corresponds to the particular time period. 
     
     
         10 . The computer-implemented method of  claim 2 , wherein the request is automatically triggered by a messaging volume exceeding a messaging volume threshold. 
     
     
         11 . A system comprising:
 one or more processors; and   one or more non-transitory machine-readable storage media containing instructions that, when executed on the one or more processors, cause the one or more processors to perform operations including:
 obtaining historical messaging data, wherein the historical messaging data includes skills exercised and customer satisfaction levels for a plurality of messaging conversations; 
 filtering the historical messaging data based on a particular skill to obtained filtered historical messaging data; 
 training a learning model using the filtered historical messaging data to obtain a trained learning model, wherein the trained learning model is trained to generate a staffing forecast for the particular skill at a target service level corresponding to a particular customer satisfaction level; 
 receiving a request for an updated staffing forecast corresponding to the particular skill and a desired target service level; 
 generating a prediction using the trained learning model, wherein the prediction corresponds to a projected staffing level for the particular skill and the desired target service level; and 
 providing the projected staffing level as the updated staffing forecast. 
   
     
     
         12 . The system of  claim 11 , wherein the operations further include:
 obtaining, from a messaging communication in real-time, a skill exercised during the messaging communication and a customer satisfaction level corresponding to the messaging communication, wherein the skill exercised is the particular skill.   
     
     
         13 . The system of  claim 11 , wherein the operations further include:
 adding a skill corresponding to the particular skill and a customer satisfaction level obtained from a messaging conversation in real-time to the filtered historical messaging data to obtain updated filtered historical messaging data; and   re-training the trained learning model using the updated filtered historical messaging data to obtain a re-trained learning model, wherein the re-trained learning model is used to generate projected service levels for future requests for updated staffing forecasts.   
     
     
         14 . The system of  claim 11 , wherein the historical messaging data corresponds to messages related to a particular entity. 
     
     
         15 . The system of  claim 11 , wherein the historical messaging data corresponds to messages related to an aggregation of entities. 
     
     
         16 . The system of  claim 11 , wherein generating the prediction using the trained learning model includes analyzing an arrival rate parameter. 
     
     
         17 . The system of  claim 11 , wherein generating the prediction using the trained learning model includes analyzing an agent net handle time parameter. 
     
     
         18 . The system of  claim 11 , wherein generating the prediction using the trained learning model includes analyzing a particular time period, and wherein the projected staffing level corresponds to the particular time period. 
     
     
         19 . The system of  claim 11 , wherein the request is automatically triggered by a messaging volume exceeding a messaging volume threshold. 
     
     
         20 . A non-transitory machine-readable storage medium, including instructions embodied thereon, the instructions executable by one or more processors to perform operations including:
 obtaining historical messaging data, wherein the historical messaging data includes skills exercised and customer satisfaction levels for a plurality of messaging conversations;   filtering the historical messaging data based on a particular skill to obtained filtered historical messaging data;   training a learning model using the filtered historical messaging data to obtain a trained learning model, wherein the trained learning model is trained to generate a staffing forecast for the particular skill at a target service level corresponding to a particular customer satisfaction level;   receiving a request for an updated staffing forecast corresponding to the particular skill and a desired target service level;   generating a prediction using the trained learning model, wherein the prediction corresponds to a projected staffing level for the particular skill and the desired target service level; and   providing the projected staffing level as the updated staffing forecast.   
     
     
         21 . The non-transitory machine-readable storage medium of  claim 20 , wherein the operations further include:
 obtaining, from a messaging communication in real-time, a skill exercised during the messaging communication and a customer satisfaction level corresponding to the messaging communication, wherein the skill exercised is the particular skill.   
     
     
         22 . The non-transitory machine-readable storage medium of  claim 20 , wherein the operations further include:
 adding a skill corresponding to the particular skill and a customer satisfaction level obtained from a messaging conversation in real-time to the filtered historical messaging data to obtain updated filtered historical messaging data; and   re-training the trained learning model using the updated filtered historical messaging data to obtain a re-trained learning model, wherein the re-trained learning model is used to generate projected service levels for future requests for updated staffing forecasts.   
     
     
         23 . The non-transitory machine-readable storage medium of  claim 20 , wherein the historical messaging data corresponds to messages related to a particular entity. 
     
     
         24 . The non-transitory machine-readable storage medium of  claim 20 , wherein the historical messaging data corresponds to messages related to an aggregation of entities. 
     
     
         25 . The non-transitory machine-readable storage medium of  claim 20 , wherein generating the prediction using the trained learning model includes analyzing an arrival rate parameter. 
     
     
         26 . The non-transitory machine-readable storage medium of  claim 20 , wherein generating the prediction using the trained learning model includes analyzing an agent net handle time parameter. 
     
     
         27 . The non-transitory machine-readable storage medium of  claim 20 , wherein generating the prediction using the trained learning model includes analyzing a particular time period, and wherein the projected staffing level corresponds to the particular time period. 
     
     
         28 . The non-transitory machine-readable storage medium of  claim 20 , wherein the request is automatically triggered by a messaging volume exceeding a messaging volume threshold.