Patent Issued for System and method for automatically assigning a customer call to an agent (USPTO 11936818): Massachusetts Mutual Life Insurance Company

2024 APR 04 (NewsRx) -- By a News Reporter-Staff News Editor at Insurance Daily News -- A patent by the inventors Merritt, Sears (Groton, MA, US), filed on March 15, 2023, was published online on March 19, 2024, according to news reporting originating from Alexandria, Virginia, by NewsRx correspondents.

Patent number 11936818 is assigned to Massachusetts Mutual Life Insurance Company (Springfield, Massachusetts, United States).

The following quote was obtained by the news editors from the background information supplied by the inventors: “Customer contact centers, also herein called call centers, provide an important interface for customers/partners of an organization to contact the organization. The contact can be for a request for a product or service, for trouble reporting, service request, etc. The contact mechanism in a typical call center is via a telephone, but it could be via a number of other electronic channels, including e-mail, etc.

“The typical contact center consists of a number of human agents, with each assigned to a telecommunication device, such as a phone or a computer for conducting email or Internet chat sessions, that is connected to a central switch. Using these devices, the agents are generally used to provide sales, customer service, or technical support to the customers or prospective customers of a contact center or a contact center’s clients.

“Typically, a contact center or client will advertise to its customers, prospective customers, or other third parties a number of different contact numbers or addresses for a particular service, such as for sales, for billing questions, or for technical support. The customers, prospective customers, or third parties seeking a particular service will then use this contact information, and the incoming caller will be routed at one or more routing points to a human agent at a contact center who can provide the appropriate service. Contact centers that respond to such incoming contacts are typically referred to as “inbound contact centers.”

“Conventionally, a contact center operation includes a switch system that connects callers to agents. In an inbound contact center, these switches route incoming callers to a particular agent in a contact center, or, if multiple contact centers are deployed, to a particular contact center for further routing. When a call lands at a contact center (which can be physically distributed, i.e., the agents may or may not be on a single physical location), it important to route the call to an appropriate call agent, and also to provide call agents with information that will enable the agent to handle the call more effectively.

“In an example of conventional methods for routing an inbound call, if there are eight agents at a contact center, and seven are occupied with contacts, the switch will generally route the inbound caller to the one agent that is available. If all eight agents are occupied with contacts, the switch will typically put the contact on hold and then route it to the next agent that becomes available. More generally, the contact center will set up a queue of incoming callers and preferentially route the longest-waiting callers to the agents that become available over time. Such a pattern of routing contacts to either the first available agent or the longest-waiting agent is sometimes referred to as “round-robin” contact routing. In round-robin contact routing, eventual matches and connections between a caller and an agent are essentially random.

“Some attempts have been made to improve upon these standard yet essentially random processes for connecting a caller to an agent. In some customer contact systems, calls are distributed based on static attributes of the caller (e.g., customer identifier), type of problem (typically obtained by prompting the caller to interact with the IVR system) and capability profile of the agents. Conventional contact center systems make only limited use of dynamic, time-dependent attributes used in agent selection, such as traffic parameters (load on an agent, time of day etc.), information on agent performance, or dynamic information on customers. The call assignment process typically is not informed by prior experiences the same or similar customers had with the same or similar agents.”

In addition to the background information obtained for this patent, NewsRx journalists also obtained the inventors’ summary information for this patent: “In an embodiment, a method of the present invention automatically assigns or routes an inbound call from a customer to one of a plurality of agents, the agent being selected on the basis of likelihood of a favorable outcome. The call routing procedure is effected promptly upon receiving the customer call, with no need for preliminary interaction with the customer before routing the call. The method calculates outcome predictions by applying one or more predictive models determined for the identified customer to model variables. In an embodiment, the method routes the identified customer to one of the plurality of agents only after identifying an agent for which the calculated outcome prediction satisfies a favorable outcome criterion. In an embodiment, the call routing procedure uses time-series analysis of time-dependent data for more efficient modeling outcome predictions for each a set of call routing options in order to select one of a plurality of agents with the most favorable outcome prediction.

“The method of the invention calculates outcome predictions by applying the predictive model for the identified customer to dynamic model variables, i.e., model variables with values that change over time and/or model variables that include a plurality of instances at different points in time. In an embodiment, a model variable may represent call history data for the identified customer, the call history data including a plurality of customer interaction records recorded at different points in time. In this embodiment, each customer interaction record may include timestamp data, e.g., representative of the time of recording that record. In addition to calculating outcome predictions by applying the predictive model to dynamic model variables, the method may apply the predictive model to static model variables, i.e., model variables with a value or set of values that does not change over time.

“In an embodiment, together with selecting an agent from the plurality of agents, the method selects a product from a plurality of products, e.g., for potential sale of the selected product to a prospective or existing customer. In an embodiment, the method provides the selected agent with information based upon the analysis used for selecting the agent to receive the call, to assist the agent in effectively handling the call.

“In an embodiment, the method includes initially receiving a customer call for assistance, e.g., via telephone, the customer call including one or more customer identifier such as telephone number. Based upon the customer identifier, without requesting additional information from the customer, the method retrieves customer profile information for an identified customer from one or more internal databases of the call center, and/or one or more third-party databases.

“In an embodiment, the retrieved customer profile information includes risk data associated with the identified customer. In an embodiment, the customer profile data includes call history data associated with the identified customer. In one aspect of this embodiment, the call history data includes one or more customer interaction records including timestamp data associated with each of the customer interaction records.”

The claims supplied by the inventors are:

“1. A method, comprising: upon receiving a customer call at a call receiving device, the customer call including a customer identifier, retrieving, by a processor, customer data for the customer identifier comprising one or more customer records including timestamp data associated with each customer record; executing, by the processor, a predictive model to generate an outcome prediction for each of a set of call routing options by applying the predictive model to values of model variables over a time interval, wherein the time interval is a historical time period ending at a current time, determining, by the processor, whether any outcome prediction generated for the set of call routing options satisfies a favorable outcome criterion; in the event none of the outcome predictions generated for the set of call routing options satisfies the favorable outcome criterion, dynamically adjusting, by the processor, the time interval and repeating the executing step and the determining step by applying the predictive model to values of the model variables over the dynamically adjusted time interval; and in the event the outcome prediction generated for one of the set of call routing options satisfies the favorable outcome criterion, directing, by the processor, the call receiving device to route the customer call to the one of the set of call routing options.

“2. The method of claim 1, wherein the model variables comprise the customer data including the timestamp data associated with each customer record.

“3. The method as defined in claim 1, wherein the customer records for the customer identifier comprise one or more customer interaction records including timestamp data associated with each customer interaction record.

“4. The method of claim 3, wherein the model variables comprise the customer data including the timestamp data associated with each customer interaction record.

“5. The method as defined in claim 1, wherein the customer data comprise customer profile data associated with the customer identifier and call history data comprising one or more customer interaction records including timestamp data associated with each customer interaction record, wherein the model variables comprise the customer profile data and the call history data.

“6. The method as defined in claim 1, wherein the dynamically adjusting the time interval increases the time interval by increasing the historical time period ending at the current time.

“7. The method as defined in claim 6, wherein the dynamically adjusting the time interval effects a time-series analysis in which the time interval is increased by a standard rate.

“8. The method as defined in claim 1, wherein each of the set of call routing options comprises the customer identifier and one of a plurality of agents, wherein routing the customer call to the one of the set of call routing options comprises routing the customer call to the one of the plurality of agents comprised in the one of the set of call routing options.

“9. The method as defined in claim 8, further comprising the step of retrieving, by the processor, agent profile data for each of the plurality of agents, wherein the model variables further comprise the agent profile data retrieved for each of the plurality of agents.

“10. The method as defined in claim 8, further comprising the step, following the step of directing the call receiving device to route the customer call, of reporting to the one of the plurality of agents information on one or more of the model variables having a largest effect on the outcome prediction generated for the one of the set of call routing options.

“11. The method as defined in claim 8, further comprising the step of determining targeted marketing data associated with one or more products selected from a products database.

“12. The method as defined in claim 11, wherein in the executing step, each call routing option comprises the customer identifier, the one of the plurality of agents, and one of the one or more products selected from the products database.

“13. The method as defined in claim 12, further comprising the step, following the step of directing the call receiving device to route the customer call, of reporting to the one of the plurality of agents the targeted marketing data associated with the one of the one or more products selected from the products database.

“14. The method as defined in claim 1, wherein the predictive model is a Markov model.

“15. The method as defined in claim 1, wherein the model variables in the predictive model are selected using a random-forests learning method.

“16. A system, comprising: a call-receiving device for receiving a customer call; non-transitory machine-readable memory that stores customer data associated with a plurality of identified customers; a processor, configured to execute a call routing module, wherein the processor in communication with the non-transitory machine-readable memory executes a set of instructions instructing the processor to: upon receiving a customer call at a call receiving device, the customer call including a customer identifier, retrieve customer data for the customer identifier comprising one or more customer records including timestamp data associated with each customer record; execute a predictive model to generate an outcome prediction for each of a set of call routing options by applying the predictive model to values of model variables over a time interval, wherein the time interval is a historical time period ending at a current time, determine whether any outcome prediction generated for the set of call routing options satisfies a favorable outcome criterion; in the event none of the outcome predictions generated for the set of call routing options satisfies the favorable outcome criterion, dynamically adjust the time interval and repeating the executing step and the determining step by applying the predictive model to values of the model variables over the dynamically adjusted time interval; and in the event the outcome prediction generated for one of the set of call routing options satisfies the favorable outcome criterion, direct the call receiving device to route the customer call to the one of the set of call routing options.

“17. The system as defined in claim 16, wherein the dynamically adjusting the time interval increases the recent time interval by increasing the historical time period ending at the current time.

“18. The system as defined in claim 17, wherein the dynamically adjusting the time interval effects a time-series analysis in which the recent time interval is increased by a standard rate.

“19. The system as defined in claim 16, wherein each of the set of call routing options comprises the customer identifier and one of a plurality of agents, wherein route the customer call to the one of the set of call routing options comprises route the customer call to the one of the plurality of agents comprised in the one of the set of call routing options.

“20. The system as defined in claim 19, wherein the non-transitory machine-readable memory further stores the agent profile data for each of the plurality of agents, wherein the set of instructions further instructs the processor to retrieve the agent profile data for each of the plurality of agents from the non-transitory machine-readable memory, wherein the model variables further comprise the agent profile data retrieved for each of the plurality of agents.”

URL and more information on this patent, see: Merritt, Sears. System and method for automatically assigning a customer call to an agent. U.S. Patent Number 11936818, filed March 15, 2023, and published online on March 19, 2024. Patent URL (for desktop use only): https://ppubs.uspto.gov/pubwebapp/external.html?q=(11936818)&db=USPAT&type=ids

(Our reports deliver fact-based news of research and discoveries from around the world.)