Techniques for behavioral pairing in a task assignment system with an external pairing system
There are provided a method for behavioral pairing in a task assignment system with an external pairing system. The method includes transmitting, to the external pairing system over an application programming interface, a control flag and a benchmark flag and receiving, from the external pairing system, a task-agent pairing based on one of a first pairing strategy and a second pairing strategy. The task-agent pairing is based at least in part on the control flag and the benchmark flag.
1 . A method for operating a task assignment system comprising:
an external pairing system; an internal pairing system; and a switch communicatively coupled to the internal pairing system and the external pairing system, the method comprising:
the external pairing system, which is communicatively coupled to the switch, using a first pairing strategy to pair a first task with an agent, wherein, after the first task is paired with an agent, there is an interaction between the first task and the agent paired with the first task;
recording first outcome data indicating an outcome of the interaction between the first task and the agent paired with the first task;
recording a first benchmark flag, the first benchmark flag being associated, directly or indirectly, with the first outcome data and indicating that the first outcome data indicates an outcome that was produced by an interaction between a task and an agent where the task and the agent were paired using the first pairing strategy;
the external pairing system using a second pairing strategy to pair a second task with an agent, wherein, after the second task is paired with an agent, there is an interaction between the second task and the agent paired with the second task;
recording second outcome data indicating an outcome of the interaction between the second task and the agent paired with the second task;
recording a second benchmark flag, the second benchmark flag being associated, directly or indirectly, with the second outcome data and indicating that the second outcome data indicates an outcome that was produced by an interaction between a task and an agent where the task and the agent were paired using the second pairing strategy;
recording a third benchmark flag associated, directly or indirectly, with third outcome data, the third outcome data indicating an outcome of an interaction between a third task and an agent paired with the third task, wherein the third benchmark flag indicates: i) the first pairing strategy or the second pairing strategy was used to pair the third task with an agent, ii) that the internal pairing system paired the third task, and/or iii) that the third outcome data should not be used to benchmark the performance of the first and second pairing strategies; and
benchmarking performance of the first and second pairing strategies, wherein
the benchmarking comprises:
retrieving the first outcome data;
retrieving the second outcome data;
using the first outcome data to determine a performance of the first pairing strategy;
using the second outcome data to determine a performance of the second pairing; and
based on the determined performance of the first pairing strategy and the determined performance of the second pairing strategy, determining a performance of the first pairing strategy relative to the performance of the second pairing strategy.
2 . The method of claim 1 , wherein
benchmarking the performance of the first and second pairing strategies further comprises:
retrieving the third benchmark flag associated with the third outcome data;
determining, based on the third benchmark flag, whether or not to use the third outcome data to benchmark the performance of the first and second pairing strategies; and
using the third outcome data to determine the performance of i) the first pairing strategy or ii) the second pairing strategy.
3 . The method of claim 1 , wherein
benchmarking the performance of the first and second pairing strategies further comprises:
retrieving the third benchmark flag associated with the third outcome data;
determining, based on the third benchmark flag, whether or not the third outcome data should be used to benchmark the performance of the first and second pairing strategies; and
as a result of determining that the third outcome data should not be used to benchmark the performance of the first and second pairing strategies, refraining from using the third outcome data to determine the performance of either the first pairing strategy or the second pairing strategy.
4 . The method of claim 3 , wherein the third benchmark flag indicates that an error had occurred in the process of paring of the third task with an agent.
5 . The method of claim 1 , wherein the third benchmark flag indicates that the third task was paired by the internal pairing system.
6 . The method of claim 5 , wherein the third benchmark flag indicates that the third outcome data should be used to update a pairing model.
7 . A system comprising one or more computer processors for operating a task assignment system, the task assignment system comprising:
an internal pairing system;
an external pairing system; and
a switch that is communicatively coupled to the internal pairing system, wherein the task assignment system is executed by the one or more computer processors to perform a method comprising:
the external pairing system, which is communicatively coupled to the switch, using a first pairing strategy to pair a first task with an agent, wherein, after the first task is paired with an agent, there is an interaction between the first task and the agent paired with the first task;
recording first outcome data indicating an outcome of the interaction between the first task and the agent paired with the first task;
recording a first benchmark flag, the first benchmark flag being associated, directly or indirectly, with the first outcome data and indicating that the first outcome data indicates an outcome that was produced by an interaction between a task and an agent where the task and the agent were paired using the first pairing strategy;
the external pairing system using a second pairing strategy to pair a second task with an agent, wherein, after the second task is paired with an agent, there is an interaction between the second task and the agent paired with the second task;
recording second outcome data indicating an outcome of the interaction between the second task and the agent paired with the second task;
recording a second benchmark flag, the second benchmark flag being associated, directly or indirectly, with the second outcome data and indicating that the second outcome data indicates an outcome that was produced by an interaction between a task and an agent where the task and the agent were paired using the second pairing strategy;
recording a third benchmark flag associated, directly or indirectly, with third outcome data, the third outcome data indicating an outcome of an interaction between a third task and an agent paired with the third task, wherein the third benchmark flag indicates: i) the first pairing strategy or the second pairing strategy was used to pair the third task with an agent, ii) that the internal pairing system paired the third task, and/or iii) that the third outcome data should not be used to benchmark the performance of the first and second pairing strategies; and
benchmarking performance of the first and second pairing strategies, wherein
the benchmarking comprises:
retrieving the first outcome data;
retrieving the second outcome data;
using the first outcome data to determine a performance of the first pairing strategy;
using the second outcome data to determine a performance of the second pairing; and
based on the determined performance of the first pairing strategy and the determined performance of the second pairing strategy, determining a performance of the first pairing strategy relative to the performance of the second pairing strategy.
8 . The task assignment system of claim 7 , wherein
benchmarking the performance of the first and second pairing strategies further comprises:
retrieving the third benchmark flag associated with the third outcome data;
determining, based on the third benchmark flag, whether or not to use the third outcome data to benchmark the performance of the first and second pairing strategies; and
using the third outcome data to determine the performance of i) the first pairing strategy or ii) the second pairing strategy.
9 . The task assignment system of claim 7 , wherein
benchmarking the performance of the first and second pairing strategies further comprises:
retrieving the third benchmark flag associated with the third outcome data;
determining, based on the third benchmark flag, whether or not the third outcome data should be used to benchmark the performance of the first and second pairing strategies; and
as a result of determining that the third outcome data should not be used to benchmark the performance of the first and second pairing strategies, refraining from using the third outcome data to determine the performance of either the first pairing strategy or the second pairing strategy.
10 . The task assignment system of claim 9 , wherein
the third benchmark flag indicates that an error had occurred in the process of paring of the third task with an agent.
11 . The task assignment system of claim 7 , wherein
the third benchmark flag indicates that the third task was paired by the internal pairing system.
12 . The task assignment system of claim 11 , wherein
the third benchmark flag indicates that the third outcome data should be used to update a pairing model.
13 . An article of manufacture comprising:
a non-transitory computer processor readable medium; and
instructions stored on the non-transitory computer processor readable medium, wherein the instructions are configured to configure a task assignment system comprising an external pairing system; an internal pairing system; and a switch communicatively coupled to the internal pairing system and the external pairing system to perform a method comprising:
the external pairing system, which is communicatively coupled to the switch, using a first pairing strategy to pair a first task with an agent, wherein, after the first task is paired with an agent, there is an interaction between the first task and the agent paired with the first task;
recording first outcome data indicating an outcome of the interaction between the first task and the agent paired with the first task;
recording a first benchmark flag, the first benchmark flag being associated, directly or indirectly, with the first outcome data and indicating that the first outcome data indicates an outcome that was produced by an interaction between a task and an agent where the task and the agent were paired using the first pairing strategy;
the external pairing system using a second pairing strategy to pair a second task with an agent, wherein, after the second task is paired with an agent, there is an interaction between the second task and the agent paired with the second task;
recording second outcome data indicating an outcome of the interaction between the second task and the agent paired with the second task;
recording a second benchmark flag, the second benchmark flag being associated, directly or indirectly, with the second outcome data and indicating that the second outcome data indicates an outcome that was produced by an interaction between a task and an agent where the task and the agent were paired using the second pairing strategy;
recording a third benchmark flag associated, directly or indirectly, with third outcome data, the third outcome data indicating an outcome of an interaction between a third task and an agent paired with the third task, wherein the third benchmark flag indicates: i) the first pairing strategy or the second pairing strategy was used to pair the third task with an agent, ii) that the internal pairing system paired the third task, and/or iii) that the third outcome data should not be used to benchmark the performance of the first and second pairing strategies; and
benchmarking performance of the first and second pairing strategies, wherein
the benchmarking comprises:
retrieving the first outcome data;
retrieving the second outcome data;
using the first outcome data to determine a performance of the first pairing strategy;
using the second outcome data to determine a performance of the second pairing; and
based on the determined performance of the first pairing strategy and the determined performance of the second pairing strategy, determining a performance of the first pairing strategy relative to the performance of the second pairing strategy.
14 . The article of manufacture of claim 13 , wherein
benchmarking the performance of the first and second pairing strategies further comprises:
retrieving the third benchmark flag associated with the third outcome data;
determining, based on the third benchmark flag, whether or not to use the third outcome data to benchmark the performance of the first and second pairing strategies; and
using the third outcome data to determine the performance of i) the first pairing strategy or ii) the second pairing strategy.
15 . The article of manufacture of claim 13 , wherein
benchmarking the performance of the first and second pairing strategies further comprises:
retrieving the third benchmark flag associated with the third outcome data;
determining, based on the third benchmark flag, whether or not the third outcome data should be used to benchmark the performance of the first and second pairing strategies; and
as a result of determining that the third outcome data should not be used to benchmark the performance of the first and second pairing strategies, refraining from using the third outcome data to determine the performance of either the first pairing strategy or the second pairing strategy.
16 . The article of manufacture of claim 15 , wherein
the third benchmark flag indicates that an error had occurred in the process of paring of the third task with an agent.
17 . The article of manufacture of claim 13 , wherein
the third benchmark flag indicates that the third task was paired by the internal pairing system.
18 . The article of manufacture of claim 17 , wherein
the third benchmark flag indicates that the third outcome data should be used to update a pairing model.