Apparatus and method for call center service level prediction and action

• US 9,378,475 B2
• Filed: 11/28/2014
• Issued: 06/28/2016
• Est. Priority Date: 11/28/2014
• Status: Active Grant

• Current Assignee: Conduent Business Services, LLC (Conduent, Inc.)
• Original Assignee: Xerox Corporation (Xerox Holdings Corp.)
• Inventors: Zheng, Bichen, Gnanasambandam, Shanmuga-Nathan
• Primary Examiner(s): Deane, Jr., William
• Application Number:
  US14/555,988
• Publication Number:
  US 20160155080A1
• Time in Patent Office: 578 Days
• Field of Search: 37926501-26514, 37926601-2661, 379/309
• US Class Current: 1/1
• CPC Class Codes: G06Q 10/063112 Skill-based matching of a person or a group to a task
• CPC Class Codes: H04M 3/5175 Call or contact centers supervision arrangements
• CPC Class Codes: H04M 3/5238 with waiting time or load prediction arrangements

Abstract

An apparatus, method and non-transitory computer readable medium for predicting a service level of a call center are disclosed. The method performs operations for predicting a service level of a call center. The operations include collecting call data, agent topic skill data, agent skill level data, agent schedule data and agent attrition rate data, building a discrete event simulation model based on the call data, the agent topic skill data, the agent skill level data, the agent schedule data and the agent attrition rate data, predicting the service level of the call center at a future time based on the discrete event simulation model and recommending a number of agents and a skill mix of agents for each queue in the call center at the future time based on the service level that is predicted based on the discrete event simulation model to achieve a call center service objective.

Claims

1. An application server for predicting a service level of a call center, comprising:

a processor; and

a non-transitory computer-readable medium storing a plurality of instructions, which when executed by the processor, cause the processor to perform operations, the operations comprising;

collecting call data, agent topic skill data, agent skill level data, agent schedule data and agent attrition rate data;

building a discrete event simulation model based on the call data, the agent topic skill data, the agent skill level data, the agent schedule data, the agent attrition rate data and an overhead of each agent of the call center, wherein the overhead comprises a percentage of an idle time between consecutive calls;

executing the discrete event simulation model to predict the service level of the call center at a future time, wherein the service level that is predicted comprises a prediction of the service level for each one of a plurality of different topic queues;

generating a graph that displays a call volume per topic queue of the plurality of different topic queues that is predicted; and

transmitting a recommendation to the call center, wherein the recommendation comprises a number of agents and a skill mix of agents for each queue in the call center at the future time based on the service level that is predicted based on the graph generated by the discrete event simulation model to achieve a call center service objective.

2. The application server of claim 1, wherein the call data for each call comprises a number of calls, a call arrival time, a caller waiting time and a call handling time.

3. The application server of claim 2, further comprising:

creating a time series of the number of calls that is cleaned up by including an average of the number of calls for missing values within the time series.

4. The application server of claim 1, wherein the agent skill level data is determined based on an average service time of the agent relative to average service times of all other agents.

5. The application server of claim 1, wherein the agent attrition rate data comprises a percentage of agents that will quit at a particular time based on a number of agents that quit after a certain period of time.

6. The application server of claim 1, wherein the discrete event simulation model comprises a replay mode, a prediction mode and a generation mode.

7. A method for predicting a service level of a call center, comprising:

collecting, by a processor of an application server, call data, agent topic skill data, agent skill level data, agent schedule data and agent attrition rate data;

building, by the processor of the application server, a discrete event simulation model based on the call data, the agent topic skill data, the agent skill level data, the agent schedule data and the agent attrition rate data and an overhead of each agent of the call center, wherein the overhead comprises a percentage of an idle time between consecutive calls;

executing, by the processor of the application server, the discrete event simulation model to predict the service level of the call center at a future time, wherein the service level that is predicted comprises a prediction of the service level for each one of a plurality of different topic queues;

generating a graph that displays a call volume per topic queue of the plurality of different topic queues that is predicted; and

transmitting, by the processor of the application server, a recommendation to the call center, wherein the recommendation comprises a number of agents and a skill mix of agents for each queue in the call center at the future time based on the service level that is predicted based on the graph generated by the discrete event simulation model to achieve a call center service objective.

8. The method of claim 7, wherein the call data for each call comprises a number of calls, a call arrival time, a caller waiting time and a call handling time.

9. The method of claim 8, further comprising:

creating a time series of the number of calls that is cleaned up by including an average of the number of calls for missing values within the time series.

10. The method of claim 7, wherein the agent skill level data is determined based on an average service time of the agent relative to average service times of all other agents.

11. The method of claim 7, wherein the agent attrition rate data comprises a percentage of agents that will quit at a particular time based on a number of agents that quit after a certain period of time.

12. The method of claim 7, wherein the discrete event simulation model comprises a replay mode, a prediction mode and a generation mode.

13. A method for predicting a service level of a call center, comprising:

generating, by a processor of an application server, a time series of a number of calls in each one of a plurality of different time slots;

calculating, by the processor of the application server, a skill level of each agent of a plurality of agents;

calculating, by the processor of the application server, an attrition rate of the plurality of agents;

fitting, by the processor of the application server, a distribution to an inter-arrival time between calls, a call-handling time and an overhead, wherein the overhead comprises a percentage of an idle time between consecutive calls;

calculating, by the processor of the application server, an expected call-handling time based on each topic of a plurality of different topic queues;

calculating, by the processor of the application server, a frequency of the each topic of a plurality of topics of the plurality of different topic queues;

building, by the processor of the application server, a discrete event simulation model based on the time series of the number of calls, the skill level of the each agent of the plurality of agents, the attrition rate, the distribution of the inter-arrival time between calls, the call-handling time and the overhead, the expected call-handling time based on the each topic of the plurality of topics and the frequency of the each topic of the plurality of topics;

executing, by the processor of the application server, the discrete event simulation model to predict the service level of the call center at a future time, wherein the service level that is predicted comprises a prediction of the service level for each one of the plurality of different topic queues;

generating a graph that displays a call volume per topic queue of the plurality of different topic queues that is predicted; and

transmitting, by the processor of the application server, a recommendation to the call center, wherein the recommendation comprises a number of agents and a skill mix of agents for each topic queue of the plurality of different topic queues in the call center at the future time based on the service level that is predicted based on the graph generated by the discrete event simulation model to achieve a call center service objective.

14. The method of claim 13, wherein the discrete event simulation model comprises a replay mode, a prediction mode and a generation mode.