HK40007886B - Techniques for hybrid behavioral pairing in a contact center system - Google Patents

Description

Techniques for hybrid behavioral pairing in a contact center system

The application is a divisional application of a patent application with the international application number 201680070037.9(PCT/IB2016/001762) filed on 30/5/2018 and 2016, 11/2016 and 22/2016, entitled "technology for hybrid behavioral pairing in a contact center system". Cross Reference to Related Applications

This patent application claims priority to U.S. application No. 14/956,086 filed on 12/1/2015, which is a partial continuation application of U.S. patent application No. 14/871,658 filed on 9/30/2015 (U.S. patent No. 9,300,802 issued on 3/29/2016), which is a partial continuation application of U.S. patent application No. 12/021,251 filed on 28/2008/10/2014, and a partial continuation application of U.S. patent application No. 14/530,058 filed on 31/2014 (U.S. patent No. 9,277,055 issued on 3/1/2016), which is a continuation application of U.S. patent application No. 13/843,724 filed on 3/15/2013/2014 (U.S. patent No. 8,879,715 issued on 11/4/2014), which claims U.S. provisional patent application No. 61/615,788 filed on 26/2012, U.S. provisional patent application No. 61/615,779 filed on 26/2012, and U.S. provisional patent application No. 61/615,772 filed on 3/2012/26/2012 The antecedent, each of which is incorporated by reference herein in its entirety as if fully set forth herein.

Technical Field

The present disclosure relates generally to contact centers and, more particularly, to techniques for hybrid behavioral pairing in contact center systems.

Background

A typical contact center algorithmically distributes contacts arriving at the contact center to agents available for handling the contacts. Sometimes, a contact center may have agents available and waiting to be assigned to inbound or outbound contacts (e.g., phone calls, internet chat sessions, email) or outbound contacts. At other times, the contact center may have contacts waiting in one or more queues for agents to become available for distribution.

In some typical contact centers, contacts are distributed to agents ordered based on time of arrival, and agents receive contacts ordered based on the time at which those agents become available. This strategy may be referred to as a "first-in-first-out", "FIFO", or "polling" strategy. In some contact centers, contacts or agents are assigned to different "skill groups" or "queues" before a FIFO allocation strategy is applied within each such skill group or queue. These "skill sets" may also incorporate policies for prioritizing individual contacts or agents within the baseline FIFO ordering. For example, a high priority contact may be given a queue position before other contacts that arrive at an earlier time, or a high performance agent may be ordered before other agents that have been waiting longer for their next call. Despite such variations in the one or more queues forming the caller or one or more orderings of available agents, the contact center typically applies a FIFO to the queues or other orderings. Once such a FIFO policy has been established, the assignment of contacts to agents is automatic, with the contact center assigning the first contact in the sequence to the next available agent, or assigning the first agent in the sequence to the next arriving contact. In the contact center industry, the process of contact and agent distribution between skill queues, prioritization and ordering of skill queues, and subsequent FIFO allocation of contacts to agents is managed by a system known as an "automatic call distributor" ("ACD").

Some contact centers may use "performance-based routing" or "PBR" methods to order the queues of available agents. For example, when a contact arrives at a contact center with multiple available agents, the rankings available for distribution to the contact will be preceded by the highest performing available agent (e.g., the highest sales conversion rate, the highest customer satisfaction score, the shortest average processing time, the highest performing agent for a particular contact profile, the highest customer retention rate, the lowest customer retention cost, the highest one-call resolution rate). PBR ordering strategies attempt to maximize the desired outcome of each contact agent interaction, but do so generally do not consider uniformly utilizing agents in the contact center. Thus, a higher performing agent may receive significantly more contacts and feel overworked, while a lower performing agent may receive fewer contacts and sit longer, potentially reducing its chances for training and improvement and potentially reducing their rewards.

In view of the foregoing, it can be appreciated that there is a need for a system that attempts to improve contact center performance more evenly than PBRs, utilizing agents while overriding what FIFO strategies present.

Disclosure of Invention

Techniques for hybrid behavioral pairing in a contact center system are disclosed. In one particular embodiment, the techniques may be realized as a method for hybrid behavioral pairing in a contact center system, comprising: ordering the contacts; ordering the plurality of agents; applying, by at least one processor, a hybridization function to the ordering of the plurality of agents to bias a first policy for pairing toward a second policy for pairing; comparing, by the at least one processor and based on the hybridization function, a first difference in ordering between the contact and the first agent in the first pair with a second difference in ordering between the contact and a second agent different from the first agent in the second pair; and selecting, by the at least one processor, the first pair or the second pair for connection based on the comparison.

In accordance with other aspects of this particular embodiment, the first policy may comprise a behavior pairing policy and the second policy may comprise a performance-based routing policy.

According to other aspects of this particular embodiment, the blending function based comparison may further include applying, by the at least one processor, a curved diagonal policy to the ordering.

In accordance with other aspects of this particular embodiment, the ordering of the contacts or the ordering of the plurality of agents may be expressed as a percentage or a range of percentages.

According to other aspects of this particular embodiment, applying the hybridization function may also include adjusting a percentage or range of percentages of the plurality of agents.

In accordance with other aspects of this particular embodiment, the adjusted percentage of each of the plurality of agents may include a midpoint of the bandwidth of each of the plurality of agents.

In accordance with other aspects of this particular embodiment, ordering contacts may further include determining, by the at least one processor, a bandwidth for a contact type of the contact to be proportional to a frequency with which contacts of the contact type become available for distribution.

According to other aspects of this particular embodiment, a hybridization function may enable controllably targeting unbalanced agent utilization by at least one processor.

According to other aspects of this particular embodiment, applying the hybridization function may further include determining a parameter corresponding to a degree of bias toward the second policy.

According to other aspects of this particular embodiment, the method may further comprise using parameters within the continuous differentiable function.

According to other aspects of this particular embodiment, the method may further comprise using parameters within the discontinuous differentiable function.

In accordance with other aspects of this particular embodiment, applying the hybridization function may further include determining, by the at least one processor, a disproportionate bandwidth for each of the one or more agents.

According to other aspects of this particular embodiment, the selected agent of the selected pair may not be any of: agents that lag in a fairness metric, agents that rank highest in a performance metric for a particular contact type, agents that were previously assigned to contacts of a selected pair, order-marked agents, or randomly selected agents.

In other aspects of this particular embodiment, the selected one of the first and second pairs may include a worse desired immediate result than the other of the first and second pairs.

In accordance with other aspects of this particular embodiment, each successively higher ranked agent of the plurality of agents may be more likely to be selected than a corresponding lower ranked agent.

In accordance with other aspects of this particular embodiment, each successively higher ranked agent of the plurality of agents may be targeted to have a lower average latency than a corresponding lower ranked agent.

In another particular embodiment, the techniques may be realized as a method for hybrid behavioral pairing in a contact center system, comprising: ordering, by at least one processor, the contacts; determining, by at least one processor, a first ordering of a plurality of agents according to a first policy for pairing; determining, by the at least one processor, a second ordering of the plurality of agents according to a second policy for pairing; applying, by at least one processor, a hybridization function to combine the first ordering with the second ordering; comparing, by the at least one processor and based on the hybridization function, a first difference in ordering between the contact and the first agent in the first pair with a second difference in ordering between the contact and a second agent different from the first agent in the second pair; and selecting, by the at least one processor, the first pair or the second pair for connection based on the comparison.

In another particular embodiment, the techniques may be implemented as a system for hybrid behavioral pairing in a contact center system, comprising: at least one processor, wherein the at least one processor is configured to: ordering the contacts; ordering the plurality of agents; applying a hybridization function to the ordering of the plurality of agents to bias a first policy for pairing toward a second policy for pairing; comparing a first difference in ordering between a contact in the first pair and a first agent to a second difference in ordering between a contact in the second pair and a second agent different from the first agent based on a hybridization function; and based on the comparison, selecting either the first pair or the second pair for connection.

According to other aspects of this particular embodiment, the at least one processor may be further configured to controllably target unbalanced agent utilization using a hybridization function.

In accordance with other aspects of this particular embodiment, the at least one processor may be further configured to determine a disproportionate bandwidth for each of the plurality of agents.

The present disclosure will now be described in more detail with reference to specific embodiments thereof as illustrated in the accompanying drawings. While the present disclosure is described below with reference to specific embodiments, it should be understood that the disclosure is not limited thereto. Those of ordinary skill in the art having access to the teachings herein will recognize additional implementations, modifications, and embodiments, as well as other fields of use, which are within the scope of the present disclosure as described herein, and with respect to which the present disclosure may have significant utility.

Drawings

In order to facilitate a more complete understanding of the present disclosure, reference is now made to the accompanying drawings, in which like elements are referenced with like numerals. These drawings should not be construed as limiting the present disclosure, but are intended to be illustrative only.

Figure 1 illustrates a block diagram of a contact center according to an embodiment of the present disclosure.

Fig. 2 shows a schematic representation of a queue according to an embodiment of the present disclosure.

Fig. 3 shows a schematic representation of a queue according to an embodiment of the present disclosure.

Fig. 4 shows a schematic representation of a queue according to an embodiment of the present disclosure.

Fig. 5 shows a schematic representation of a queue according to an embodiment of the present disclosure.

Fig. 6 shows a schematic representation of a queue according to an embodiment of the present disclosure.

Fig. 7 shows a schematic representation of a queue according to an embodiment of the present disclosure.

Fig. 8 illustrates a flow diagram of a hybrid behavior pairing method according to an embodiment of the disclosure.

Detailed Description

A typical contact center algorithmically distributes contacts arriving at the contact center to agents that may be used to process the contacts. Sometimes, a contact center may be in the "L1 state" and have agents available and waiting to be assigned to inbound or outbound contacts (e.g., phone calls, internet chat sessions, email). At other times, the contact center may be in the "L2 state" and have contacts waiting in one or more queues for agents to become available for distribution. Such an L2 queue may be an inbound, outbound, or virtual queue. The contact center system implements various policies for distributing contacts to agents in both the L1 and L2 states.

The present disclosure relates generally to a contact center system, traditionally referred to as an "automated call distribution" ("ACD") system. Typically, such ACD processes follow an initial "skill-based routing" ("SBR") process that serves to distribute contacts and agents among skill queues within a contact center. Such skill queues may differentiate between contacts and agents based on language capabilities at a particular task group, customer needs, or agent proficiency.

The most common conventional allocation method within the queue is "first-in-first-out" or "FIFO" allocation, where the longest waiting contact is allocated to the longest waiting agent. Some contact centers implement "performance-based routing" (PBR), in which the longest-waiting contact is assigned to the highest-performance available agent. There are generally two such variations of the dispensing method. For example, the FIFO may select the least utilized available agent rather than the longest waiting agent. More generally, the FIFO may select the agent that lags most in a particular metric(s). The FIFO may also order the queue of contacts, where higher priority contact types may precede lower priority contact types in the queue. Similarly, the PBR may be modified so that the proxy performance ranking may be altered according to the type of contact pending assignment (e.g., U.S. patent No. 7,798,876 to Bala et al). The PBR can also be modified to avoid an extreme imbalance of agent utilization by setting a maximum or minimum limit on agent utilization with respect to a colleague.

The FIFO variant is usually targeted for "fairness" as it is designed to balance the distribution of contacts to agents over time. PBR takes a different approach in which the distribution of contacts to agents is purposefully skewed to increase the utilization of higher performing agents and decrease the utilization of lower performing agents. This may be done by the PBR despite the potentially negative impact on ethics and time productivity due to fatigue of over-utilized agents and insufficient training opportunities and rewards of under-utilized agents.

The present disclosure is directed to an optimization strategy for distributing contacts to agents, such as a "behavioral pairing" or "BP" strategy, that improves upon conventional distribution methods. Behavioral pairing targets balanced utilization of agents within a queue (e.g., skill queue), while potentially improving overall contact center performance beyond what FIFO or PBR methods would actually be able to achieve. This is a remarkable achievement, as BP works with the same contacts and the same agents as either FIFO or PBR methods, approximately balancing the utilization of agents, as provided by FIFO, while improving the overall contact center performance in practice over that provided by FIFO or PBR.

BP improves performance by distributing agent and contact pairs in the following manner: the allocation of potential subsequent agent and contact pairs is considered so that when the benefits of all allocations are aggregated, it can outweigh the benefits of the FIFO and PBR policies. In some cases, the BP results in an immediate contact and proxy pair, which may be the opposite of what the FIFO or PBR would indicate. For example, in the instant case, the BP may select the shortest waiting contact or the lowest performance available agent. BP honors "offspring (position)" because the system distributes contacts to agents in the following manner: if such a decision increases the probability of better contact center performance over time, it is inherently forgotten what the highest performance choice might be at that moment.

As explained in detail below, embodiments of the present disclosure relate to a technique for "hybrid behavioral pairing" ("HBP") that combines the policies of BP with those of PBR in such a way that a contact center administrator can adjust the balance between the two. For example, a contact center administrator may choose to make BP the dominant mechanism for distributing agents from within a group, with a bias towards performance-based routing. Instead of targeting balanced proxy utilization, HBP may target skewed proxy utilization. In some configurations, the bias or skew may be slight; for example, the HBP policy may be calibrated to reduce or limit the number of occasions in which any one agent in a queue (e.g., skill queue) receives more than one contact pair before other agents in the queue.

Fig. 1 shows a block diagram of a contact center system 100 according to an embodiment of the present disclosure. The description herein describes network elements, computers, and/or components of systems and methods for simulating a contact center system that may include one or more modules. As used herein, the term "module" may be understood to refer to computing software, firmware, hardware, and/or various combinations thereof. However, a module is not to be construed as software that is not implemented on hardware, firmware, or recorded on a processor-readable recordable storage medium (i.e., the module is not software itself). It should be noted that the modules are exemplary. Modules may be combined, integrated, separated, and/or duplicated to support various applications. Also, functions described herein as being performed at a particular module may be performed at one or more other modules and/or by one or more other devices in place of, or in addition to, functions performed at the particular module. Additionally, modules may be implemented across multiple devices and/or other components, which may be local or remote to one another. Further, modules may be moved from one device to another and added to the device, and/or modules may be included in both devices.

As shown in fig. 1, the contact center system may include a central switch 110. Central switch 110 may receive incoming contacts (e.g., callers) or support outbound connections to contacts via a dialer, telecommunications network, or other module (not shown). Central switch 110 may include contact routing hardware and software to facilitate routing contacts between or to one or more PBX/ACD or other queuing or switching components within the contact center.

In the contact center system 100, the central switch 110 may not be necessary if there is only one contact center or if there is only one PBX/ACD routing component. If more than one contact center is part of the contact center system 100, each contact center can include at least one contact center switch (e.g., contact center switches 120A and 120B). Contact center switches 120A and 120B may be communicatively coupled to central switch 110.

Each contact center switch for each contact center may be communicatively coupled to a plurality of agents (or "pools" of agents). Each contact center switch may support a certain number of agents (or "agents") being logged in at one time. At any given time, the logged-on agent may be available and waiting to be connected to the contact, or the logged-on agent may be unavailable for any of a number of reasons, such as being connected to another contact, performing some post-call function (such as logging information about the call), or taking a break.

In the example of fig. 1, central switch 110 routes the contact to one of two contact centers via contact center switch 120A and contact center switch 120B, respectively. Each of the contact center switches 120A and 120B is shown with two agents each. Agents 130A and 130B may be logged into contact center switch 120A, and agents 130C and 130D may be logged into contact center switch 120B.

The contact center system 100 may also be communicatively coupled to an integration service from, for example, a third party vendor. In the example of fig. 1, the hybrid behavioral pairing module 140 can be communicatively coupled to one or more switches in a switch system of the contact center system 100, such as the center switch 110, the contact center switch 120A, or the contact center switch 120B. In some embodiments, the switches of the contact center system 100 may be communicatively coupled to a plurality of hybrid behavioral pairing modules. In some embodiments, the hybrid behavioral pairing module 140 may be embedded within a component of the contact center system (e.g., embedded in a switch or otherwise integrated therewith).

The hybrid behavioral pairing module 140 can receive information about agents (e.g., agents 130A and 130B) logged into the switch and about incoming contacts via another switch (e.g., central switch 110) from the switch (e.g., contact center switch 120A), or in some embodiments, from a network (e.g., the internet or a communications network) (not shown).

The hybrid behavioral pairing module 140 can process this information and determine which contacts should be paired (e.g., matched, distributed, routed) with which agents. For example, multiple agents are available and wait for a connection to the contact (L1 state), and the contact arrives at the contact center via a network or central switch. As explained below, without the hybrid behavior pairing module 140 or similar behavior pairing module, the contact center switch would typically automatically distribute the new contact to any available agent that has been waiting the longest amount of time for the agent, under a "fair" FIFO policy, or to any available agent that has been determined to be the highest performing agent, under a PBR policy.

With the hybrid behavioral pairing module 140 or similar behavioral pairing module, the contacts and agents may be given scores (e.g., percentages or percentage ranges/bandwidths) according to a pairing model or other artificial intelligence data model so that the contacts may be matched, paired, or otherwise connected to preferred agents. In some embodiments, the hybrid behavioral pairing module 140 may be configured with an HBP policy that mixes BP and PBR policies, targeting biased rather than balanced proxy utilization.

In the L2 state, multiple contacts are available and waiting to connect to the agent, and the agent becomes available. These contacts may be queued in a contact center switch, such as a PBX or ACD device ("PBX/ACD"). Without the hybrid behavioral pairing module 140 or similar behavioral pairing module, when an agent chooses not to be available, the contact center switch would typically connect the newly available agent to any contacts that have been waiting the longest amount of time in the queue, as in a "fair" FIFO policy or PBR policy. In some contact centers, priority queuing may also be incorporated.

With the hybrid behavioral pairing module 140 or similar behavioral pairing module in the L2 scenario, as in the L1 state described above, contacts and agents may be given percentages (or percentage ranges/widths, etc.) according to, for example, a model such as other artificial intelligence models, so that agents that will be available may be matched, paired, or otherwise connected to preferred contacts.

In the case of an HBP policy, a hybridization factor or function may be applied to one or more orderings of agents to achieve a desired balance between a BP policy that targets balanced utilization and a PBR policy that targets highly skewed utilization during periods of time when the contact center is in the L1 state (i.e., multiple agents are available for allocation).

In some embodiments, the hybridization function may combine two (or more) ranking or other types of ranking systems together. For example, a contact center may have four agents: agents A, B, C, and D ("A", "B", "C", and "D") are available for pairing with a contact. Contacts may be sorted according to a plurality of sorting systems. For example, in the case of a typical FIFO policy, the agents may be ordered according to how long each agent has been waiting to allocate relative to the other agents. In the case of a typical PBR policy, agents may be ranked according to how well each agent performs with respect to other agents for a certain metric. In the case of a BP policy, agents may be ranked according to the quality of each agent's "behavioral fit" relative to other agents.

One technique for combining the two orderings is to determine a sum. For example, if the PBR policy ordered four agents as a-1, B-2, C-3, and D-4, the PBR policy would preferably pair the highest performing agent a with the next contact. And if the BP policy ranks the agents as a-4, B-2, C-3, D-1, the BP policy will preferably pair the best-fitting agent D with the next contact. In this example of an HBP policy, the sum of the two orderings would be a-5, B-4, C-6, D-5. The HBP policy will preferably pair agent B with the next contact, which is the second highest performing and second best fitting agent according to the original ranking.

Other embodiments may use other techniques for combining multiple orderings of agents. For example, the HBP ordering may be a product obtained by multiplying two or more orderings. For another example, the HBP ordering may be a weighted sum or product obtained by scaling one or more of the orderings by a scaling factor. In this way, the HBP may be configured to weight the relative performance of the proxy more or less than the relative behavior fit of the proxy.

Fig. 2 illustrates a queue 200 according to an embodiment of the disclosure operating under a BP policy 210. Queue 200 represents a simplified hypothetical situation in which four types of contacts may be distributed to any of four agents in an environment where a contact center seeks to maximize a desired metric (e.g., sales). Four evenly distributed types of contacts are allocated a percentage range (or "bandwidth"): 0.00 to 0.25 ("0-25% contact"), 0.25 to 0.50 ("25-50%"), 0.50 to 0.75 ("50-75% contact"), and 0.75 to 1.00 (75-100% contact). Four agents occupy equally spaced percentage bandwidths and are allocated percentages at the midpoints of their respective ranges: 0.00 to 0.25 ("0.125 proxy"), 0.25 to 0.50 ("0.375 proxy"), 0.50 to 0.75 ("0.625 proxy"), and 0.75 to 1.00 ("0.875 proxy"). The four agents may also be ranked by performance according to a desired metric (e.g., sales) such that the lowest performing agent is assigned the lowest percentage (0.125 agent) and the highest performing agent is assigned the highest percentage (0.875 agent).

By applying a diagonal policy, 0-25% contacts may preferably be assigned to the 0.125 agent, 25-50% contacts may preferably be assigned to the 0.375 agent, 50-75% contacts may preferably be assigned to the 0.625 agent, and 75-100% contacts may preferably be assigned to the 0.875 agent. BP policy 210 targets balanced utilization, where each agent receives approximately the same proportion of contacts over time. Thus, there is no bias towards the PBR strategy under which proxy utilization would be skewed towards the 0.875 proxy which utilizes the highest performance more heavily.

One such technique for generating performance biased agent percentages in accordance with embodiments of the present disclosure is to adjust the "initial" midpoint percentage "AP of each agent by a hybridization function or factor_initial"such that a relatively higher ranked (e.g., higher performance) agent takes up relatively more bandwidth (i.e., disproportionate bandwidth) and, therefore, receives relatively more contacts than a lower ranked (e.g., lower performance) agent. For example, the hybridization function may raise the percentage of each contact by a power, as in equation 1 below:

AP_adjusted＝AP_initial ^κ(equation 1)

The power parameter (e.g., "κ" or "Kappa parameter" as in equation 1 may determine the amount of bias towards PBR, and higher values of Kappa generate larger amounts of bias). A Kappa parameter of 1.0 will not generate a bias (AP)_adjusted＝AP_initial). Thus, this "neutral" value of Kappa results in the goal of balancing proxy utilization. In fact, the BP policy 210 is equal to the Kappa-based HBP policy, where Kappa is equal to 1.0. As Kappa increases, the degree of skew utilized by the proxy is a function of the bias towards the PBRIncreasing towards an increase.

Fig. 3 shows a queue 300 that applies this technique using a Kappa value of 2.0. Queue 300 represents the same four types of contacts and the same four agents as in queue 200. However, in queue 300, the percentage midpoint of the agent has been squared (AP)_adjusted＝AP_initial ^2.0). Applying diagonal policy, lowest-ranked proxy (AP) in the case of HBP policy 310_adjusted0.016) will take the least bandwidth and receive the least contacts, and so on, up to the highest ranked Agent (AP)_adjusted0.766) which will occupy the maximum bandwidth and receive the most contacts.

In some embodiments, the bandwidth of each agent may be determined such that the percentage midpoint of the adjustment for each agent is the midpoint of the new adjusted bandwidth for each agent. For example, the bandwidth of the lowest ranked 0.016 contact type may be approximately 0.000 to 0.031. In other embodiments, the bandwidth of each agent may be determined by equally distributing the "distance" between the midpoints of adjacent adjusted percentages. For example, the bandwidth of the lowest ranked 0.016 agent may be approximately 0.000 to 0.079.

Another version of the HBP technique applied to the queue 300 in fig. 3 is to adjust the initial percentage range of each agent instead of the initial midpoint percentage of each agent as in equation 2 below:

AP_{adjusted_range}＝AP_{initial_range} ^κ(equation 2)

The effect will be the same: a relatively higher ranked (e.g., higher performing) agent takes up relatively more bandwidth and, therefore, receives relatively more contacts than a lower ranked (e.g., lower performing) agent.

Fig. 4 shows a queue 400 that applies this technique using a Kappa value of 2.0. Queue 400 represents the same four contact types and agents as in queue 300. However, in queue 400, the initial percentage range of agents has been squared (AP)_{adjusted_range}＝AP_{initial_range} ^2.0) Rather than its initial midpoint percentage. Should be under HBP policy 410With the diagonal policy, the lowest ranked agents (occupying the percentage range of accommodation from 0.00 to approximately 0.06, with an approximate midpoint of 0.03) will receive the fewest contacts, and so on, until the highest ranked agents (occupying the percentage range of accommodation from approximately 0.56 to 1.00, with an approximate midpoint of 0.82) will receive the most contacts.

Conceptually, the use of target skew will result in the next half of the agents receiving approximately one-quarter of the contacts and the first half of the agents receiving the other three-quarters. Other techniques for visualizing or implementing these hybridization functions or factors include adjusting the "fit function" of the diagonal strategy.

FIG. 5 shows queue 500 having the same percentage and extent of agents as in queue 100 (FIG. 1), and which has not yet been adjusted. Unlike queue 110, in which BP policy 110 may be visualized by a 45 degree diagonal (CP ═ AP), HBP policy 510 may be visualized by a different, hybrid fitting function (e.g., to "bend" or "buckle" the diagonal). In the example of fig. 5, the fitting function is an exponential function, as shown in equation 3 below:

CP＝AP^κ(equation 3)

Conceptually, instead of determining a preferred pairing by selecting the pair closest to the diagonal CP-AP as in BP policy 110, a preferred pairing in HBP policy 510 may be determined by selecting the pair closest to the exponential CP-AP^2.0As in queue 500, where Kappa equals 2.0. Notably, fitting to a curve (e.g., CP ═ AP)^2.0) The effect of (a) is a continuous mathematical simulation of a discontinuous process that widens or shrinks the percentage range (e.g., squares the percentage range and then fits to CP ═ AP), as in queue 400 and HBP policy 410 (fig. 4).

Many variations of hybridization functions may be used to vary the target utilization of agents according to their performance or other ranking or metric. For example, the hybridization function may be a piecewise function.

Fig. 6 shows queue 600 and HBP policy 610 that affect the utilization of the lower half of the proxy differently than the utilization of the upper half of the proxy. For exampleThe contact center may determine that half of the contact should be distributed to agents below average in a balanced manner (e.g., Kappa ═ 1.0), but that the other half of the contact should be distributed to agents above average (e.g., Kappa ═ 1.0) based on the relative performance of the agents above average>1.0). Thus, contacts ranging from 0% to 50% can be evenly distributed to lower performing agents (0.125 agents and 0.375 agents) visualized as a fit along the 45 degree line CP ≦ AP, where 0.00 ≦ AP<0.50 (or, for example, 0.00)<AP is less than or equal to 0.50, and the like). Contacts ranging from 50% to 100% can be distributed to higher performing agents (0.625 agents and 0.875 agents) as a function of their performance, such as scaling to an exponential function of this portion of the contacts and agents. HBP policy 610 may be visualized as following the exponential curve CP-2 (AP-0.5)^2.0+0.5, where Kappa is 2.0 and 0.50 ≦ AP<1.00。

Occasionally, such a strategy would result in some higher performing agents (here, the 0.625 agent) receiving fewer contacts over time than their lower performing colleagues. Fig. 7 shows queue 700 and HBP policy 710 that also affect the utilization of the next half of the proxy differently than the utilization of the first half of the proxy using a segmented hybridization function. For example, the contact center may determine that a larger portion of the contact should be distributed to agents above average (e.g., Kappa) based on their relative rankings>1.0) and the remainder of the contact should be distributed in a balanced manner to the agents below average (e.g., Kappa-1.0). Thus, for Kappa 2.0 and AP ≧ 0.50 (or AP)>0.50), the pairing may be along the exponential curve CP ═ AP^2.0And (6) fitting. For Kappa ═ 1.0 and AP<0.50, the pairing may be a fit along a linear function scaled to this portion of the contact and agent: CP is 0.5 · AP.

In a real-world contact center, there may be more or fewer agents, and more or fewer contact types in the queue. In these examples, each contact type is evenly distributed over the full range of percentage rankings; however, in some contact centers, the distribution of the ranges may be based on the frequency with which certain types of contacts arrive at the contact center relative to the frequency with which other types of contacts arrive. The simplified example described above with four proxies and four contact types is used to illustrate the effects of the implicit form of HBP, such as those based on Kappa parameters and exponential scaling or other hybridization functions. However, HBP, including Kappa-based technologies, may also be applied to larger, more complex real-world contact centers.

In some embodiments, Kappa may be selected or adjusted to vary the bias towards the PBR (or skew in proxy utilization). For example, a Kappa of less than 2.0 (e.g., 1.0, 1.01, 1.1, 1.2, 1.5, etc.) will result in a relatively lower bias towards the PBR than the above example where Kappa is equal to 2.0. For example, if a contact center administrator wants to avoid the occurrence of a higher performing agent receiving multiple calls while a lower performing agent remains idle, a significantly lower value for Kappa may be more appropriate than 2.0. Conversely, a Kappa greater than 2.0 (e.g., 2.01, 2.1, 2.5, 200.0, etc.) will result in a relatively more bias toward the PBR.

Importantly, the impact on agent utilization is subtle under the Kappa-based HBP policy, as they controllably affect the extent to which agents wait between contacts. By increasing the power of the boosted agent percentage, the present invention controllably reduces the average time between contacts for higher ranked agents and increases the average time between contacts for relatively lower ranked agents. Similarly, lowering the power of the boosted proxy percentage has the opposite effect. For a neutral BP policy (e.g., Kappa ═ 1.0), each agent has approximately the same expected average latency between contacts. As Kappa increases, the relative expected average latency gradually (e.g., exponentially) decreases as the relative proxy performance increases.

In some embodiments, HBP policies may use potentially more progressive techniques to target relative proxy utilization. For example, agents may be assigned relative "utilization adjustments" based on agent rankings. In one example, the highest ranked agent may be assigned a relative utilization adjustment of 100%, the second highest agent may be assigned a relative utilization of 99%, the third highest agent may be assigned a relative utilization of 98%, and so on. In this example, the target utilization of the second highest ranked agent would be 99% of the target utilization of the highest performing agent. The relative utilization adjustment may be more aggressive in other configurations. For example, the highest ranked agent may be assigned a relative utilization of 100%, the second highest ranked agent may be assigned a relative utilization of 90%, the third highest ranked agent may be assigned a relative utilization of 80%, and so on. In this example, the target utilization of the second highest ranked agent would be 90% of the target utilization of the highest performing agent.

Fig. 8 illustrates a hybrid behavior pairing method 800 according to an embodiment of the disclosure. At block 810, the hybrid behavior pairing method 800 may be initiated.

At block 810, a percentage (or n-piece, quantile, range of percentages, bandwidth, or other type of "score" or range of scores, etc.) may be determined for each available contact. For the case where contacts are waiting in the queue, a percentage may be determined for each of the contacts waiting in the queue. For the case where a contact is not waiting in the queue, a percentage may be assigned to the next contact to reach the contact center. Based on the information about the contacts, the percentage may be bounded by a range of percentages defined for a particular type of contact or a particular group of contacts. The percentage limit or range may be based on a frequency distribution or other metric for the type of contact. Within the percentage range of the type, the percentages may be randomly assigned.

In some embodiments, the percentages may be ordered according to a particular metric or combination of metrics to be optimized in the contact center, and contacts determined to have a relatively high percentage may be considered "higher value" contacts for the contact center because these contacts are more likely to contribute to a higher overall performance in the contact center. For example, a relatively high percentage of contacts may have a relatively high likelihood of making a purchase.

In some embodiments, a percentage may be determined for a contact at the time the contact arrives at the contact center. In other embodiments, the percentage may be determined for the contact at a later point in time, such as when the contact arrives at a particular skills queue or ACD system, or when a pairing request is made.

After the percentage has been determined for each contact available for pairing, the behavioral pairing method 800 may proceed to block 820. In some embodiments, block 820 may be performed before or simultaneously with block 810.

At block 820, a percentage may be determined for each available agent. For the case where the agents are idle waiting for contacts to arrive, a percentage may be determined for each of the idle agents. For the case where all agents in the queue are busy, the percentage may be determined as the next agent that becomes available. The percentage may be bounded by a range of percentages (e.g., "bandwidth") defined based on all agents assigned to a queue (e.g., skill queue) or only available agents assigned to a particular queue. In some embodiments, the bounds or ranges of percentages may be based on desired agent utilization (e.g., for fairness, efficiency, or performance).

In some embodiments, the agent percentages may be ranked according to a particular metric or combination of metrics to be optimized at the contact center, and agents determined to have a relatively high percentage may be considered to be higher performing agents of the contact center. For example, a relatively high percentage of agents may have a relatively high likelihood of making a sale.

In some embodiments, the percentage of agents may be determined at the time the agents become available within the contact center. In other embodiments, the percentage may be determined at a later point in time, such as when a pairing request is made.

After the percentages have been determined for each available agent and contact, the behavioral pairing method 800 may continue to block 830.

At block 830, a hybridization function may be applied to the proxy percentage (or proxy percentage range or bandwidth). For example, Kappa values may be determined for exponential hybridization functions or fitting curves or lines. In some embodiments, the hybridization function may act on a single ordering that implicitly incorporates both behavioral fitting and performance information. In other embodiments, the hybridization function may combine (e.g., add, multiply, weight) multiple orderings of agents. After the hybridization functions have been applied or otherwise determined or configured, the mixing behavior pairing method 800 may continue to block 840.

At block 840, based on the hybridization function, a pair of available contacts and available agents may be determined based on the percentage (or percentage range) determined for each available contact at block 810 and for each available agent at block 820. In some embodiments, the selection may be determined based on the percentage or percentage range for each available agent adjusted at block 830. In some embodiments, the pair may be determined according to a diagonal policy, where contacts and agents having a more similar percentage (or the most similar percentage) may be selected for pairing. For example, the hybrid behavioral pairing module may select the contact-agent pairing having the smallest absolute difference between the score of the contact and the score of the agent. In some embodiments, the diagonal policy may be visualized as a 45 degree diagonal. In other embodiments, the diagonal policy may be visualized as a hybridization function (e.g., such as an exponential function, a linear function, a logarithmic function, a sigmoidal function, a logical function, a continuous differentiable function to a metafunction, or a discontinuous differentiable function such as a piecewise function).

In some cases, multiple agents may idle when a contact arrives (L1 state). In the case of HBP, a newly available contact may be paired with a selected one of the available agents that has a more similar adjusted percentage or range of percentages as the percentage of contacts than the other available agents. In other cases, multiple contacts may wait in the queue when an agent becomes available (L2 state). In the case of HBP, a newly available agent may be paired with a selected one of the contacts waiting in the queue, which has a percentage more similar to the agent's adjusted percentage or percentage range than the other contacts waiting in the queue.

In some cases, selecting a pairing based on the similarity of scores may result in selecting an instant pairing that may not be the highest performing instant pairing, but instead increases the likelihood of a better future pairing.

After pairing has been determined at block 840, the hybrid behavioral pairing method 800 may continue to block 850. At block 850, modules within the contact center system may connect the contacts and agents of the contact-agent pair to each other. For example, the behavioral pairing module may instruct an ACD system or other routing device that a particular contact may be distributed to a particular agent.

After connecting the contacts and agents at block 850, the behavioral pairing method 800 may end. In some embodiments, the behavioral pairing method 800 may return to block 840 for determination of one or more additional pairings (not shown). In other embodiments, the behavioral pairing method 800 may return to block 810 or block 820 to determine (or re-determine) a percentage or range of percentages for available contacts or agents (not shown), and then apply (or re-apply) the hybridization function at block 840.

At this point, it should be noted that the hybrid behavioral pairing in a contact center system according to the present disclosure as described above may involve, to some extent, the processing of input data and the generation of output data. The input data processing and output data generation may be implemented in hardware or software. For example, certain electronic components may be employed in a behavioral pairing module or similar or related circuitry as described above in accordance with the present disclosure for implementing functionality associated with behavioral pairing in a contact center system. Alternatively, one or more processors operating in accordance with instructions may implement the functionality associated with behavioral pairing in a contact center system in accordance with the present disclosure as described above. If so, it is within the scope of the disclosure that such instructions may be stored on one or more non-transitory processor-readable storage media (e.g., a magnetic disk or other storage medium) or transmitted to one or more processors via one or more signals embodied in one or more carrier waves.

The present disclosure is not to be limited in scope by the specific embodiments described herein. Indeed, other various embodiments and modifications of the disclosure, in addition to those described herein, will be apparent to those of ordinary skill in the art from the foregoing description and accompanying drawings. Accordingly, such other embodiments and modifications are intended to fall within the scope of the present disclosure. Further, while the present disclosure has been described herein in the context of at least one particular implementation in at least one particular environment for at least one particular purpose, those of ordinary skill in the art will recognize that its usefulness is not limited thereto and that the present disclosure may be beneficially implemented in any number of environments for any number of purposes. Accordingly, the claims set forth below should be construed in view of the full breadth and spirit of the present disclosure as described herein.

Claims (27)

1. A method for hybrid behavioral pairing in a contact center system, comprising:

determining, by at least one computer processor communicatively coupled to the contact center system and configured to perform a behavioral pairing operation in the contact center system, a first ordering of a plurality of agents in accordance with a behavioral pairing policy having balanced agent utilization;

determining, by the at least one computer processor, a second ordering of the plurality of agents according to a performance-based routing policy with unbalanced agent utilization;

determining, by at least one computer processor, a third ordering of the plurality of agents according to a hybrid behavior pairing policy, wherein the third ordering is determined according to the hybrid behavior pairing policy as having a combination of the first ordering and the second ordering of skewed agent utilization between the balanced agent utilization and the unbalanced agent utilization; and

establishing, by the at least one computer processor, a connection between a contact and an agent in a switch of the contact center system based on the third ordering of the plurality of agents.

2. The method of claim 1, wherein the behavioral pairing policy comprises a diagonal pairing policy.

3. The method of claim 1, further comprising: determining, by the at least one processor, a target skew amount for the skew agent utilization.

4. The method of claim 3, wherein a combination of the first ordering and the second ordering is a weighted sum.

5. The method of claim 1, wherein the first ordering can be expressed as a percentage or a range of percentages.

6. The method of claim 1, wherein the third ordering can be expressed as a percentage or a range of percentages adjusted according to a combination of the first ordering and the second ordering.

7. The method of claim 1, wherein under the hybrid behavior pairing strategy, higher performing agents are paired more frequently than lower performing agents.

8. The method of claim 1, wherein under the hybrid behavioral pairing strategy, higher performing agents are paired with a greater number of contact types than lower performing agents.

9. The method of claim 1, wherein under the hybrid behavioral pairing strategy, higher performing agents are paired with higher frequency contact types than lower performing agents.

10. A system for hybrid behavioral pairing in a contact center system, comprising:

at least one computer processor communicatively coupled to the contact center system and configured to perform behavioral pairing operations in the contact center system, wherein the at least one computer processor is configured to:

determining a first ranking of a plurality of agents according to a behavioral pairing policy having balanced agent utilization;

determining a second ordering of the plurality of agents according to a performance-based routing policy with unbalanced agent utilization;

determining a third ordering of the plurality of agents according to a hybrid behavior pairing policy, wherein the third ordering is determined according to the hybrid behavior pairing policy as having a combination of the first ordering and the second ordering of skewed agent utilization between the balanced agent utilization and the unbalanced agent utilization;

and

establishing, in a switch of the contact center system, a connection between a contact and an agent based on the third ordering of the plurality of agents.

11. The system of claim 10, wherein the behavioral pairing policy comprises a diagonal pairing policy.

12. The system of claim 10, wherein the at least one computer processor is further configured to: determining a target skew amount for the skewed agent utilization.

13. The system of claim 12, wherein a combination of the first ordering and the second ordering is a weighted sum.

14. The system of claim 10, wherein the first ordering can be expressed as a percentage or a range of percentages.

15. The system of claim 10, wherein the third ordering can be expressed as a percentage or a range of percentages adjusted according to a combination of the first ordering and the second ordering.

16. The system of claim 10, wherein under the hybrid behavior pairing strategy, higher performing agents are paired more frequently than lower performing agents.

17. The system of claim 10, wherein under the hybrid behavioral pairing strategy, higher performing agents are paired with a greater number of contact types than lower performing agents.

18. The system of claim 10, wherein under the hybrid behavioral pairing strategy, higher performing agents are paired with higher frequency contact types than lower performing agents.

19. An article of manufacture for hybrid behavioral pairing in a contact center system, comprising:

a non-transitory processor-readable medium; and

instructions stored on the medium;

wherein the instructions are configured to be readable from the medium by at least one computer processor communicatively coupled to the contact center system and configured to perform behavioral pairing operations in the contact center system, and thereby cause the at least one computer processor to operate to:

determining a first ranking of a plurality of agents according to a behavioral pairing policy having balanced agent utilization;

determining a second ordering of the plurality of agents according to a performance-based routing policy with unbalanced agent utilization;

determining a third ordering of the plurality of agents according to a hybrid behavior pairing policy, wherein the third ordering is determined according to the hybrid behavior pairing policy as having a combination of the first ordering and the second ordering of skewed agent utilization between the balanced agent utilization and the unbalanced agent utilization;

and

establishing, in a switch of the contact center system, a connection between a contact and an agent based on the third ordering of the plurality of agents.

20. The article of manufacture of claim 19, wherein the behavioral pairing policy comprises a diagonal pairing policy.

21. The article of manufacture of claim 19, wherein the at least one computer processor is further caused to operate to: determining a target skew amount for the skewed agent utilization.

22. The article of manufacture of claim 21, wherein a combination of the first ordering and the second ordering is a weighted sum.

23. The article of claim 19, wherein the first ordering may be expressed as a percentage or a range of percentages.

24. The article of manufacture of claim 19, wherein the third ordering can be expressed as a percentage or a range of percentages adjusted according to a combination of the first ordering and the second ordering.

25. The article of manufacture of claim 19, wherein higher performing agents are paired more frequently than lower performing agents under the mixed behavior pairing policy.

26. The article of claim 19, wherein under the hybrid behavioral pairing strategy, higher performing agents are paired with a greater number of contact types than lower performing agents.

27. The article of claim 19, wherein under the mixed behavior pairing strategy, higher performing agents are paired with higher frequency contact types than lower performing agents.