Call Routing from First to Second Dialog of Interactive Voice Response System

Abstract

A customer service call system that provides both manual (assisted by a service representative) and automated problem solving dialogs. A routing process determines which dialog is appropriate for incoming calls. The routing process determines a probability that the automated system will resolve the problem and also determines an expected hold time for the customer to reach the manual dialog. The routing process uses this information to determine how the call should be routed.

Description

TECHNICAL FIELD OF THE INVENTION

This invention relates to interactive voice response systems for providing customer service, and more particularly to a method for routing calls from manual to automated dialogs.

BACKGROUND OF THE INVENTION

A business or company that provides services and/or products to clients or customers may provide their customers with customer service in the form of a customer service center that handles customer requests. Customer requests may comprise requesting new products or services, getting support for a product or service, asking questions about a product or service, etc. In non-automated systems, when a customer calls the service center with a request, the service center manually routes the call to an agent that services the customer's request.

Interactive Voice Response (IVR) systems are systems that provide information in the form of recorded messages over telephone lines in response to customer input in the form of spoken words or touch tone signaling. Examples of IVR systems are those implemented by banks, which allow customers to check their balances from any telephone, and systems for providing automated stock quotes.

IVR may be used to both to acquire information from, or enter data into, a database. For example, banks and credit card companies use IVR systems so that their customers can receive up-to-date account information instantly and easily without having to speak directly to a person. An IVR system may also be used to gather information. For example, an IVR telephone survey might prompt the customer to answer questions by pushing the numbers on a touch-tone telephone.

When fully implemented, an IVR system does not require human interaction over the telephone. The customer's interaction with the database is predetermined by the extent to which the IVR system will allow the customer access to the data.

IVR systems have been successful because they generate significant cost savings due to the ability to route calls to the appropriate call center without having to incur the cost of a service representative to accomplish the routing. Once the call is routed, subsequent dialog with the customer could be additional IVR dialog or it could be assisted by a service representative. Thus, a given customer service task could be partly implemented with IVR and partly with human assistance.

The absence of a service representative can sometimes be frustrating to the customer, such as when an IVR system simply provides recorded answers that assume the customer has already correctly diagnosed a problem. On the other hand, although non-automated systems do provide a live service representative, customers can become frustrated if they are required to be put on hold to wait for service.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 illustrates a partially automated customer service dialog.

FIG. 2 illustrates a fully automated customer service dialog.

FIG. 3 illustrates the aggregated responses of customers when queried how long they would wait for service representative, given an alternative to attempt an automated problem solving system.

FIG. 4 illustrates a method of routing calls from a partially to a fully automated problem solving system in accordance with the invention.

FIG. 5 illustrates a customer service call system having a routing unit in accordance with the invention.

DETAILED DESCRIPTION OF THE INVENTION

The following description is directed to a method of overcoming the limitations of existing IVR systems. For purposes of the following description, “automated” dialogs are those that are implemented using IVR technology, typically in the form of recorded menu type selections but may be in the form of any type of prompt for interactive customer responses. The customer responses may be by touch tone, by voice recognition, TTY, some other type of telephone key pad, or some combination of these.

The method provides a customer service call system with two alternative dialogs; one partially automated and partially manual and the other fully automated. A customer who places a call to a call center begins at the partially automated dialog. When appropriate, the customer's call is routed from the partially automated dialog to the fully automated dialog. In this manner, the system balances the advantages and disadvantages IVR-implemented customer service call-centers.

Specifically, the method described herein provides customers with an IVR-type problem detection dialog with the opportunity to speak with a service representative for help in diagnosing and resolving the problem. However, the method also addresses the objections of customers to remaining in queues for long hold times by determining when they may be better served by an automated system.

For purposes of this description, customer service tasks are presented to call centers, with the customer initiating the call by telephone. Each task is deemed a “problem” to be solved, with the problem solving having three phases: detection, diagnosis, and resolution.

FIG. 1 illustrates a problem solving system 10, which is partially IVR-implemented. Specifically, problem detection is presented as an IVR-implemented dialog. That is, the customer identifies the problem to be solved from a menu of items. However, the diagnosis and problem resolution dialogs of system 10 are “manual”, in the sense that once the specific problem is detected, a service representative (SR) is brought on-line to assist the customer with diagnosis and resolution of the problem.

In other words, system 10 requires the customer to detect a problem, then report the problem to a service representative who is trained in diagnosing the problem. The problem detection is IVR implemented in the sense that the customer responds to a menu type list of choices, by using either touch tone or voice responses. During the subsequent dialog, which is manual, the service representative asks the customer appropriate questions to aid in the diagnosis and to determine the solution.

FIG. 2 illustrates a problem solving system 20 that is fully IVR implemented. The entire dialog of system 20 aids the customer in problem detection and diagnosis without any assistance from a service representative. System 20 incorporates a set of diagnostic questions, which assist the customer in mapping a detected problem to a solution. Thus, system 20 requires the customer to detect and diagnose the problem by answering queries. The queries can be in a “tree” type format, with the response to any particular question determining which query is to follow.

The diagnostic questions of system 20 are developed using the following principles. First, the questions are designed to permit the customer to successfully complete the task. Second, each question has an appropriate level of readability and comprehension. Third, the questions must be two sentences or less, and contain three instructions or less. Fourth, the questions must provide pause, repeat, backup, opt out, and get summary functionality.

FIG. 3 illustrates a hold time function, derived from customer responses to a survey type query. As indicated above, one problem with manual systems, such as the manual dialog of system 10, is that customers may experience long hold times. FIG. 3 illustrates the hold time that a customer will willingly endure before speaking to a customer service representative, given the alternative of using an automated system, such as system 20. As illustrated, the more likely the automated system is to provide a solution, the less time the customer is willing to wait for a customer service representative. As explained below, FIG. 3 represents a “wait or switch” function, which may be used to determine when actual calls should be routed from a manual dialog to an automated dialog.

As an example of how the data of FIG. 3 might be obtained, a customer survey could be conducted. A sample group of customers might be presented with the following scenario: “You've called your Internet Service Provider to get help because you can't connect to the Internet. You've been put on hold and an announcement says you can either opt out to an automated problem solving system or you can wait to speak with a customer service representative. There is an [n]% chance that the APS could solve your problem. How long would you be willing to wait for a service representative before using the automated system?”

FIG. 4 illustrates how the customer survey data of FIG. 3 may be used to develop a call-routing method, which permits a customer's call to be routed from a manual dialog, such as that of system 10, to an automated dialog, such as that of system 20. As explained below, the routing decision is based on two factors: an expected hold time for the manual dialog, and the probability that the automated dialog will resolve the customer's problem.

For a given call, the expected hold time and the probability of success are quantified and integrated into an algorithm, which provides an index value for the call. The index value is compared to a threshold value. Depending on the result of the comparison, the call is then left with the manual system 10 or routed to an automated system 20. Alternatively, instead of the customer's call being automatically routed, the customer could be asked to choose whether he or she desires to wait for a service representative or to use the automated system.

Referring to the specific steps of FIG. 4, Step 41 is determining the task to be performed for the customer. In the system 10 of FIG. 1, task identification is accomplished when the customer selects a menu option at the outset of the IVR problem detection dialog. In other embodiments, Step 41 could be performed using a live service representative who interacts with the customer to determine the nature of the problem.

Step 42 determines the probability that the detected task can be successfully resolved by the automated system 20. For this step, it is assumed that some basis exists for determining this probability, such as statistical data representing past calls and success rates. The task identification information is the subject of a database lookup to determine the probability that the customer will receive a satisfactory resolution from the automated system. Step 43 is determining the expected hold time for the customer. This may be accomplished by querying the call center queue size.

Step 44 is using the probability of success value (determined in Step 42) and the expected hold time value (determined in Step 43) to calculate a customer wait index. The following equation is an example of how these two values may be used in this manner:

Wait Index=HT+74.6 PS−44 PS²−33.25,

where HT is the expected hold time in minutes and PS is the probability of solution.

The above function is derived from customer preference data, such the data illustrated in FIG. 3. For example, the curve of FIG. 3 can be fitted to an equation that best matches the curve, resulting in a function such as that set out above.

Step 45 is comparing the Wait Index to a threshold value. A threshold of 0 indicates that the current hold time (HT) and the current probability of solution (PS) values both match predetermined customer expectancies.

Step 46 is routing the call based on the result of the comparison. Positive values of the Wait Index indicate that a combination of HT and PS are greater than the threshold. In this case, the call is automatically routed to the automated system. Negative values of the Wait Index indicate that the combination of HT and PS are less than the threshold. In this case, the customer remains in the manual dialog queue to speak to a service representative. In either case, the customer may alternatively be queried to determine if he or she would like to be routed to the automated system.

The following example illustrates how the above-described method works for a call with HT=10 and PS=0.50.

$\begin{array}{c}\mathrm{Wait}\mathrm{Index}=10+74.6\left(\mathrm{.5}\right)-44{\left(\mathrm{.5}\right)}^2-33.25\\ =10+37.3-11-33.25\\ =3.05\end{array}$

This positive value results in the call being routed to the automated system, or the customer being queried about remaining in the queue.

The following example illustrates how the method works for a call with HT=1 and PS=0.50.

$\begin{array}{c}\mathrm{Wait}\mathrm{Index}=1+74.6\left(\mathrm{.5}\right)-44{\left(\mathrm{.5}\right)}^2-33.25\\ =1+37.3-11-33.25\\ =-5.95\end{array}$

This negative value results in the call remaining in the queue or the customer being queried about opting for the automated system.

In accordance with the above-described method, when hold times do not meet customer expectancies, customers may be automatically routed to an automated problem system 20, which can assist in problem diagnosis. The use of the wait index, with thresholds based on customer data may also be used to give the customer control over the decision to opt for the automated system.

FIG. 5 represents a customer call center system in accordance with the invention. The customer initially interacts with an IVR problem detection dialog 51, such as the detection dialog of FIG. 1. In other embodiments, dialog 51 could be a manual dialog, in which the customer interacts with a service representative. Once the task to be solved is determined, a routing unit 52 is programmed to calculate a wait index, compare the index to a threshold, and obtain a routing decision, as described above in connection with FIG. 4. Depending on the routing decision, the call is routed to the manual system 53 where the customer interacts with a live service representative or to the automated system 54 where the customer interacts with an IVR dialog. It is assumed that each of the elements of FIG. 5 is implemented with appropriate hardware and software. The routing unit is implemented with computer programming and databases appropriate for performing the above described functions. Suitable implementations of the various dialogs are known in the art or capable of being developed.

OTHER EMBODIMENTS

Although the present invention has been described in detail, it should be understood that various changes, substitutions, and alterations can be made hereto without departing from the spirit and scope of the invention as defined by the appended claims.

Claims

1. A method of routing a customer's call to either a manual dialog or an automated dialog of a customer service call system, comprising the steps of: detecting a task to be solved by the call;determining a probability value, representing the probability that the task will be solved by the automated dialog;predicting the hold time to reach the manual dialog;calculating a wait index as a function of the hold time and the probability value;comparing the results of the calculating step to a threshold wait index; androuting the call to the manual dialog or the automated dialog, based on the results of the comparing step.

2. The method of claim 1, wherein the detecting step is performed by the customer's selection from an interactive voice response menu.

3. The method of claim 1, wherein the determining step is performed by accessing data that is representative of customer calls using the automated dialog.

4. The method of claim 1, wherein the predicting step is performed by accessing data representing a current queue of the call system.

5. The method of claim 1, wherein the calculating step is performed by basing the index on the function, which represents the results of a customer survey.

6. The method of claim 1, wherein the routing step is performed by automatically routing the call.

7. The method of claim 1, wherein the routing step is performed by querying the customer to make a routing decision.

8. A routing system for customer service system, which has alternative manual and automated dialogs for solving customer problems, comprising: a problem detection process that elicits the task to be performed from the customer; anda routing unit that performs the following process: determines a probability value, representing the probability that the task will be solved by the automated dialog; predicts the hold time to reach the manual dialog; calculates a wait index as a function of the hold time and the probability value; compares the results of the calculating step to a threshold wait index; and routes the call to the manual dialog or the automated dialog, based on the results of the comparing step.

9. The system of claim 8, wherein the problem detection process is an interactive voice response process.

10. The system of claim 8, wherein the problem detection process is a manual process.

11. The system of claim 8, further comprising a database representing past customer calls using the automated dialog, accessible by the routing unit.

12. The system of claim 8, wherein the routing unit predicts the hold time by accessing data representing a current queue of the call system.

13. The system of claim 8, wherein the routing unit automatically routes the call.

14. The system of claim 8, wherein the routing unit queries the customer to make a routing decision.

15. A routing system for customer service system, which has alternative manual and automated dialogs for solving customer problems, comprising: a problem detection process that elicits the task to be performed from the customer;a routing unit that performs the following process: determines a probability value, representing the probability that the task will be solved by the automated dialog; predicts the hold time to reach the manual dialog; calculates a wait index as a function of the hold time and the probability value; compares the results of the calculating step to a threshold wait index; and routes the call to the manual dialog or the automated dialog, based on the results of the comparing step;an automated dialog process which presents queries to the customer, receives responses to the queries, and generates new queries based on the responses.

16. The system of claim 15, wherein the problem detection process is an interactive voice response process.

17. The system of claim 15, wherein the problem detection process is a manual process.

18. The system of claim 15, further comprising a database representing past customer calls using the automated dialog, accessible by the routing unit.

19. The system of claim 15, wherein the routing unit predicts the hold time by accessing data representing a current queue of the call system.

20. The system of claim 15, wherein the routing unit automatically routes the call.

21. (canceled)