WO2012107112A1

WO2012107112A1 - Interactive voice response system

Info

Publication number: WO2012107112A1
Application number: PCT/EP2011/060950
Authority: WO
Inventors: Munish Agarwal; Vikram AGARWAL; Akshay MEHRA
Original assignee: Telefonaktiebolaget L M Ericsson (Publ)
Priority date: 2011-02-08
Filing date: 2011-06-29
Publication date: 2012-08-16

Abstract

There is provided a method for dynamically configuring an optimized call flow in an Interactive Voice Response (IVR) system. The IVR system being configured for incoming calls from a mobile user subscribed to a network operator in a mobile communications system. In an embodiment, the method includes collecting usage data associated with a plurality of mobile users subscribed to the network operator and analyzing the usage data to identify a call flow pattern associated with each of the plurality of mobile users. The method further includes grouping the plurality of mobile users into a plurality of groups based at least in part on the call flow pattern and the usage data. Further, the method includes determining a plurality of optimized call flows corresponding to each of the groups and associating an optimized call flow with each of the plurality of groups.

Description

INTERACTIVE VOICE RESPONSE SYSTEM TECHNICAL FIELD

Implementations described herein relate generally to Interactive Voice Response (IVR) systems and particularly to an auto-configuring interactive voice response system that configures an optimized call path for an incoming call.

BACKGROUND

Many modern day service providers employ an Interactive Voice Response System (IVR) to facilitate client query resolutions and other call based service options. Interactive Voice Response (IVR) corresponds to a technology that allows a computing system to interact with humans via voice and Dual Tone Multiple Frequency (DTMF) keypad inputs. For example, in the telecommunications industry, the IVR system allows customers or subscribers to interact with a company's database via a telephone keypad (e.g. using DTMF) or by speech recognition. The subscribers can then service their own inquiries by following the IVR dialogue. The IVR systems can respond with pre-recorded or dynamically generated audio to further direct users on how to proceed. IVR applications can be used to control almost any function where the interface can be broken down into a series of simple interactions. IVR systems deployed in typical mobile communication networks are sized to handle large call volumes. IVR systems are used by network operator's to reduce the need for an expensive human workforce for simple day-to-day customer interactions. Nevertheless, while an IVR system is less expensive than a human based response network, the costs of an IVR system can be significant. One of the major components contributing to the cost of an IVR system is the platform cost which is typically licensed by number of voice channels or ports. Often the number of calls to a service provider's IVR system exceeds the number of available ports or channels. A typical incoming call to the IVR system would be guided via a default call flow where at one or more stages in the call flow, the caller is invoked to make a selection from a voice menu by either speaking or pressing a button on some the calling device. In some cases, it so happens that the subscriber calling the IVR system requires a particular IVR menu option but has to pass through a certain call flow path to reach the desired IVR menu option. This leads to unwanted delay for the customer and also the engagement of one of the IVR systems ports throughout the duration of the incoming call. Hence, it is desirable to free the voice channels or ports from an ongoing incoming call at the earliest time possible. This can be done by directing the incoming call to the desired IVR menu destination quickly, avoiding redundant options in the call flow path.

Existing methods to minimize the length of time an incoming call is in an IVR system include storing the optimal call flow details for each individual subscriber. However, for a large customer database such a solution requires a large data store which in turn requires a significant and costly processing power and storage resources. Indeed, if implemented such an approach limits the scalability of the IVR system by making it costly and complex. Accordingly, there is a need for an improved IVR system.

SUMMARY

It is an object of the IVR system disclosed to optimize call flow associated with an incoming call to the IVR system. Another object of the disclosed IVR system is to minimize holding time during an incoming call. A further object of the disclosed IVR system is to minimize the port and platform requirements in an IVR system.

Embodiments of a method for dynamically configuring an optimized call flow in an Interactive Voice Response (IVR) system are disclosed. The IVR system being configured for incoming calls from a mobile user subscribed to a network operator in a mobile communications system. In an embodiment, the method includes collecting usage data associated with a plurality of mobile users subscribed to the network operator and analyzing the usage data to identify a call flow pattern associated with each of the plurality of mobile users. The method further includes grouping the plurality of mobile users into a plurality of groups based at least in part on the call flow pattern and the usage data. Further, the method includes determining a plurality of optimized call flows corresponding to each of the groups and associating an optimized call flow with each of the plurality of groups.

Embodiments of call flow management node for dynamically configuring call flows in an IVR system are disclosed. The IVR system being configured for mobile users subscribed to an operator network. According to an embodiment, the call flow management node includes a statistics collection module configured to collect user data associated with a plurality of mobile users from the IVR system. The call flow management bode further includes a pattern recognition module configured to identify a usage pattern of each of the plurality of mobile users and a grouping module configured to group the plurality of mobile users into a plurality of groups based at least in part on the usage pattern and the usage data. In addition, call flow management node includes a call flow optimization module configured to determine optimal call flows corresponding to each group of the plurality of mobile users and a call flow mapping module configured to associate an optimized call flow with each of the plurality of groups.

Embodiments of a method for dynamically selecting an optimized call flow in an Interactive Voice Response (IVR) system are disclosed. The IVR system being configured for incoming calls from mobile users subscribed to a network operator in a mobile communications system. According to an embodiment, the method includes receiving an incoming call from a mobile user subscribed to a network operator and identifying a group associated with the mobile user. The method further includes determining the optimized call flow corresponding to the identified group and applying the optimized call flow to the incoming call.

Additional features of the embodiments described herein will be set forth in the description that follows, and in part will be obvious from the description, or may be learned by the practice of the described methods and apparatus. The features and advantages of the embodiments may be realized and obtained by means of the system and combinations particularly pointed out in the appended claims. These and other features of the presently disclosed system will become more fully apparent from the following description and appended claims, or may be learned by the practice of the described methods and apparatus as set forth hereinafter.

BRIEF DESCRIPTION OF THE DRAWINGS

To further clarify the above and other advantages and features of the methods and apparatus as set forth herein, a more particular description will be rendered by references to specific embodiments thereof, which are illustrated in the appended drawings. It is appreciated that these drawings depict only typical embodiments and are therefore not to be considered limiting of its scope. The methods and apparatus as set forth herein will be described and explained with additional specificity and detail with the accompanying drawings in which:

FIG. 1 illustrates an exemplary auto-configuring IVR system for dynamically configuring call flows in an IVR system;

FIG. 2 illustrates an exemplary IVR menu in a call flow according to an embodiment; FIG. 3 illustrates call flow between user, IVR system, and call flow management system according to an exemplary implementation;

FIG. 4 illustrates exemplary call flows for three different calls to IVR system according to an implementation;

FIG. 5 illustrates an exemplary comparison of usage patterns during training phase according to an embodiment;

FIG. 6 illustrates call flows, corresponding percentage access, and parameter details according to an embodiment;

FIG. 7 illustrates an exemplary graphical distribution of number of callers and parameter value (account balance) of callers;

FIG. 8 illustrates an exemplary graphical distribution of number of callers and parameter value (age) of callers;

FIG. 9 illustrates an exemplary method for dynamically configuring optimized call flow in Interactive Voice Response (IVR) system; and

FIG. 10 illustrates an exemplary method for dynamically selecting optimized call flow in

Interactive Voice Response (IVR) system.

DETAILED DESCRIPTION

Embodiments are disclosed for dynamically configuring an optimized call flow in an Interactive Voice Response (IVR) system. As described above, one of the major components contributing to the cost of an IVR system is the platform cost which is typically licensed on voice channels or ports. The cost for platforms and the ports can be significant. It may be desirable in such scenarios to direct the incoming call to the desired IVR menu option as soon as possible without having to go through redundant call flow path and free the voice channels or ports from an ongoing incoming call.

Disclosed methods and systems reduce or eliminate unwanted holding time for an incoming call by automatically configuring the IVR system to implement an optimized call flow for an incoming call. The incoming call thus would be directed to an IVR menu option that the user is most likely to choose thereby reducing the holding time for the incoming call. Conventional methods to minimize holding time of an incoming call mandate large storage requirements and high processing power. The disclosed systems and methods herein deploy probabilistic models to determine an optimized call flow thereby reducing such storage and processing requirements. The basic principle of the disclosed systems and methods is "Making the most used modules accessible in the least amount of time" . In addition, the requirement for number of ports in an IVR system depends directly on the 'mean call time' which is essentially the time for which a customer engages with the IVR system. The number of ports required can be reduced if the IVR call flow is designed in such a way that the mean call time is minimized.

Accordingly, the method for dynamically configuring an optimized call flow in an Interactive Voice Response (IVR) system includes collecting usage data associated with a plurality of mobile users subscribed to the network operator and analyzing the usage data to identify a call flow pattern associated with each of the plurality of mobile users. The method further includes grouping the plurality of mobile users into a plurality of groups based at least in part on the call flow pattern and the usage data. Further, the method includes determining a plurality of optimized call flows corresponding to each of the groups and associating an optimized call flow with each of the plurality of groups. The IVR system described herein can be configured for incoming calls from a mobile user subscribed to a network operator in a mobile communications system.

In addition, embodiments of methods for dynamically selecting an optimized call flow in an Interactive Voice Response (IVR) system are disclosed. According to an embodiment, the method includes receiving an incoming call from a mobile user subscribed to a network operator and identifying a group associated with the mobile user. The method further includes determining the optimized call flow corresponding to the identified group and applying the optimized call flow to the incoming call.

A concept relevant to the disclosed systems and methods is the fact that IVR usage pattern of a subscriber or a mobile user depends on several parameters (network, system, and external). Hence, the disclosed systems determine a relationship between the usage pattern and one or more such parameters and predict how a given subscriber or mobile user would use an IVR system over a period of time and/or a number of occasions. The usage of the IVR system corresponds to the series of choice of IVR menu options made by the mobile user. The object of this is to predict an optimized call flow for each of the subscribers and associating the optimized call flow with an incoming call from the subscribers. The exemplary methods involve training a call flow management node while associating a call flow with an incoming call to the IVR system from the subscriber. Over a period, the call flow management node attains higher accuracies and an optimized call flow is associated with each of the subscribers to the network operator.

By basing the usage pattern on a group of parameters rather than solely on the subscriber calling in, a different call flow can be selected for the same subscriber under different condition (parameters). Such use of parameters provides a more realistic call flow selection and does not require storing of the optimized call flow for each subscriber.

For the purposes of the ongoing description, in most of the instances, the following terms would have the meaning as defined below:

• Call flow: A "call flow" corresponds to a path traced by the subscriber while traversing through an IVR menu by selecting one or more options provided therein.

• Call flow Usage Pattern (or simply usage pattern or call flow pattern): A usage pattern is a representation of usage pattern of a call describing what and when the call flow services were used.

• Pattern Variance: Pattern variance corresponds to a quantitative difference between two call flow usage patterns.

• Tolerance: In the context of call flow usage patterns, tolerance corresponds to acceptable value of Pattern Variance between two call flow usage patterns to be considered as similar

• Usage group: A logical grouping based on parameters which have a similar Call flow Usage pattern. Typical parameters include Network parameter like Location information, Hosting System Parameters like Gender, Age, Current Salary, Account Balance, outstanding credit, or external Parameters like Stock Market Value or change, Time of the day. The parameters will vary depending upon the system with which the IVR is connected to.

• Optimized call flow: For each call flow usage pattern there is one and only one call flow that provides the least call holding time and best customer satisfaction which is called the Optimized call flow.

Fig. 1 illustrates an exemplary call flow management node 102 for dynamically configuring call flows in an IVR system 103. The IVR system 103 can be configured to receive incoming calls from a plurality of mobile users subscribed to a network operator. The IVR system 103 can correspond to all modern day IVR systems, modules, subsystems either stand alone or integrated with other telecommunication modules deployed by service providers (network operators) to facilitate query resolutions, and other call based service options in mobile communications network. Although, the call flow management node 102 is shown as a module or block separate from the IVR system 103, it may be appreciated that in various other embodiments, the call flow management node 102 can reside in the IVR system 103 without departing from the scope of this description.

The call flow management node 102 includes a statistics collection module 104 configured to collect usage data associated with the plurality of mobile users from the IVR system 103. Every network operator or mobile service provider maintains/stores the usage details of subscribers or mobile users for various purposes, for example, but not limited to, charging, profiling, etc. In an implementation, the collected usage data includes network parameters, system parameters, and external parameters. The network parameters may include one or more of user location and on net/off net status associated with the plurality of mobile users. The system parameters include one or more of gender, age, current salary, account balance, and outstanding credit. The external parameters include one or more of stock market value and time of day. The statistics collection module 104 further obtains the call flow (if available) that was triggered for a call, the options that were selected by the user in the IVR menu, start time and end time for one or more modules in the IVR menu.

For example, the usage data for a call flow 200 shown in fig. 2 is as follows:

STARTTIME:24102010100001,MSISDN:9810830178,ANSWERCALL:5sec,MAINMENU: 10 sec,B ALENQ : 20 sec,ENDTIME: 24102010100046 In the example, above, the incoming call was initiated by the cell number "9810830178" on 24^th September 2010 at 10:00:01. The caller had spent 5 seconds in the welcome module, 10 seconds in main menu module and 20 seconds in balance enquiry module respectively. Figure 2 illustrates. The call flow management node 102 further includes a pattern recognition module 106 configured to identify a usage pattern (also referred to as call flow pattern) of each of the plurality of mobile users. In an embodiment, the identification of usage pattern is based at least in part on the usage data from the statistics collection module 104. The usage pattern includes information relating to selected IVR menu options, a start time, and an end time for each module of an IVR menu. It may be appreciated that in any IVR system 103, there will be an IVR menu that allows mobile users or subscribers to interact with the network operator's database via a telephone keypad or by speech recognition. The subscribers can then service their own inquiries by following the IVR dialogue and selecting appropriate options from the IVR menu. The IVR systems 103 typically responds with pre-recorded or dynamically generated audio to further direct users on how to proceed.

For instance, when a mobile user or a subscriber has finished calling the IVR system 103, during a training phase of the call flow management node 102, the statistics collection module 104 obtains the usage data and forwards it to the pattern recognition module 106. The pattern recognition module 106 first calculates a usage pattern for the call which includes the call flow details such as the IVR menu along with the time taken in each IVR module. Fig. 4 illustrates example call flows 400 for Call 1, Call 2, and Call 3. For instance, call flow 402 for Call 1 has the usage pattern as shown below in Table 1 :

Table 1:

Similarly, the usage pattern for call flow 404 corresponding to Call 2 has the call details as shown in Table 2 below:

Table 2:

Yet another example usage pattern for Call 3 as per the call flow 406 is shown below in Table 3: Table 3:

The call flow management node 102 also includes a grouping module 108 configured to group the plurality of mobile users or subscribers into a plurality of groups based at least in part on the usage pattern and the usage data. The grouping module 108 makes it possible for the call flow management node to associate a call flow to a group rather than an individual according to an embodiment. In such an embodiment, the storage space required would be less as compared to the case where an individually optimized call flow is to be associated with the individual subscribers.

The call flow management node 102 can also includes a call flow optimization module 110 configured to determine optimal call flows corresponding to each group of the plurality of mobile users. The call flow management node 102 further includes a call flow mapping module 112 configured to associate an optimized call flow with each of the plurality of groups. In one of the implementations, the call flow mapping module 112 is further configured to select and map one of the optimized call flows (e.g. 114) to an incoming call from a mobile device 116 based on the group corresponding to the mobile user initiating the incoming call. For example, call flow for group 1 and group 2 have been illustrated under the call flows 114 in fig. 1.

An IVR call flow comprises of a set of menu items and modules used for user interaction. When a call is initiated from the user to the IVR these modules are executed depending on the user profile and the inputs provided by the user to menu items. An example of a typical IVR menu that is provisioned in existing IVR systems is illustrated in fig. 2. As shown, the IVR menu 200 includes one or more IVR modules. For example, the IVR menu 200 includes welcome module 202 and main menu module 204. The main menu 204 can include further options such as, balance enquiry module 206, refill module 214, and additional services menu module 216. The balance enquiry module 208 can further include detailed balance menu module 208. Similarly, the advanced services menu module 216 includes language change module 218, pin change module 220, family and friend module 222, and value voucher refill module 224. The detailed balance menu 208 includes an exit module 210 and a detailed enquiry module 212. An incoming call from a mobile user can be directed to one or more of these modules as per the options selected by the mobile user. A series of such selection made from the beginning to the end of call entailing access to one or more of the modules constitute a call flow and associated usage pattern.

In operation, the method of arriving at a suitable or optimized call flow for a particular incoming call involves two phases: the analysis phase or the training phase and the selection phase.

Fig. 3 illustrates interactions between the IVR system 103 and the call flow management node 102. As shown in the figure, a user 302 initiates a call to the IVR system 103. The IVR system 103 extracts usage data (e.g. parameters) and also captures usage pattern, and call flow details for the incoming call. The IVR system 103 forwards such details and usage data to the call flow management node 102. The call flow management node 102 selectively processes the usage data and the usage pattern details during the training phase to draw inferences and store results for future training cycles. Fig. 4, for instance, illustrates an exemplary training sample set 400. The call flow management node 102 identifies a usage pattern (such as table 1, table 2, table 3) corresponding to a given incoming call. For example, fig. 4 shows three call flows 402, 404, 406, for three calls, i.e. call 1, call 2, and call 3 respectively. The purpose of usage pattern recognition is to consider as many patterns as possible for arriving finally at an optimized call flow. To this end, the call flow management node 102 defines metrics to distinguish one usage pattern from the other. A user of the system 100 can set pre-determined threshold values (referred to as "tolerance" in this description) to classify two usage patterns as similar or dissimilar.

In a successive progression, the pattern recognition module 106 compares the identified usage pattern with an existing usage pattern if any. It may be appreciated that the training phase may be an ongoing stage wherein the call flow management node 102 may be subjected to repeated training phases to improve the accuracies with regard to optimization of call flows. In such as case, a second or subsequent training phase will have an existing pattern identified in any of the previous training phases. Furthermore, every training phase may result in additional usage patterns different from those already recorded. Such usage patterns are also included in the "existing usage patterns" for next or subsequent training. Accordingly, the call flow optimization module 110 calculates a variance between the identified usage pattern and the existing usage pattern. For purposes of the ongoing description, "variance" can be defined as a quantitative difference between two usage patterns or call flow patterns.

Pattern variance between two calls can be calculated as the deviation between two calls after a common part of the cal (time taken by caller A after common part of the call + time taken by caller B after common part of the call). For example, if caller A calls the IVR system 103, hears the Welcome prompt (welcome module 202), goes to the main menu (main menu module 204), and selects balance enquiry (balance enquiry module 206) as his choice, the call holding time = 5 (Answer Call) + 10 (Welcome) + 20 (Balance Enquiry) + 10 (Dedicated Balances) + 5 (End Call) = 50 seconds. If caller B calls the IVR, hears the Welcome prompt, goes to the Main menu, and selects

Refill as his choice, the call holding time = 5 (Answer Call) + 10 (Welcome) + 20 (Refill) = 35 seconds. The deviation between these two calls can be calculated as 35 (Balance Enquiry + Dedicated Balances +End Call) + 20 (Refill) = 50. If caller B had hung up his call after just entering his voucher number, the call holding time would have been = 5 + 10 + 15 = 30 seconds. The deviation in this case would be calculated as 30 + 15 = 45 seconds.

Further, the predetermined threshold value for tolerance can be determined and set by a call flow designer depending on one or more of the lengths of the voice announcements and the mean holding time which will be based on a call flow design

Fig. 5 illustrates such a comparison according to an embodiment. For instance, block 502 depicts the comparison for call 1, block 504 depicts the comparison for call 2, and block 506 depicts the comparison for call 3. As shown in case of call 1, there being no existing usage pattern, the identified pattern as identified in fig. 4 for call 1 is designated as Pattern 1 for the incoming call or the mobile user. If the variance between the identified usage pattern and all the existing patterns (for the same mobile user) is beyond a predetermined threshold or "tolerance", then the identified usage pattern is stored as the new usage pattern or call flow pattern for the corresponding incoming call or the mobile user. For instance, in the comparison shown in 504 for call 2, the identified usage pattern in fig. 4 for call 2 is compared with the existing pattern (i.e. Pattern 1). The variance (=20) computed during the comparison in 504 is greater than the predetermined threshold (=2 in this embodiment). Thus, the identified usage pattern is set as Pattern 2 for the incoming call or the mobile user. If there are more than one existing usage patterns having variance below the tolerance then the identified usage pattern is considered similar to that existing usage pattern which has the least variance. For instance, in the comparison shown in 506 for call 3, the identified usage pattern for call 3 in fig. 4 is compared with existing patterns (Pattern 1, Pattern 2). The variance computed during the comparison in 506 with Pattern 1 is "1" and with pattern 2 is "18". Since variance in case of Pattern 2 is greater than the predetermined threshold (=2 in the current embodiment), and the variance in case of Pattern 1 is less than the predetermined threshold, the identified pattern is considered similar to Pattern 1.

Therefore, after repeated training phases and multiple incoming calls, the call flow management module 102 would have stored a plurality of usage patterns, corresponding to an incoming call or a mobile user. The usage data (for e.g. network parameters, system parameters, and external parameters) are also stored along with the call flow usage pattern for subsequent analysis (training phase) in future. The call flow management node 102 also groups the mobile user or subscriber based on the usage data and usage pattern details obtained from the IVR system 103. Subsequently, the call flow management node 102 determines an optimal call flow corresponding to each of the groups. In an embodiment, an optimal call flow may correspond to that usage pattern which has a maximum percentage access (of the usage pattern) by a mobile user over time. For instance, fig. 6 illustrates the three sample call flows shown in fig. 4, the corresponding percentage access, and the parameter details 602 for the corresponding call flow.

The IVR system 103 logs the usage pattern and usage data for all the incoming calls. The usage data would include parameters like location, gender, age, current salary, account balance, outstanding dues, stock market value, time of the day, etc. The call flow management node 102 performs an offline analysis process of this data to create groups of mobile users. Once the grouping is over, an appropriate call flow (optimized call flow) is assigned to each of the groups so that when a caller calls in to the IVR system 103, the most appropriate call flow is selected based on the group during the selection phase.

The following table 4 depicts a relationship between the usage pattern and the groups according to an example embodiment.

Table 4:

In an embodiment, the grouping module 108, for each usage pattern, plots all the parameter values against the number of callers or mobile users. Subsequently, the grouping module 108, for each parameter, compares graphs of various usage patterns and finds out an overlap of the graph amongst the usage patterns. The parameter that has the least overlapping area amongst the usage pattern plays the biggest influence on the usage pattern. Such a parameter is selected by the grouping module 108 as the parameter for grouping with the maximum weight. Next, the grouping module 108 selects other parameters in increasing order of the overlapping area and reducing the weight of the parameters accordingly. The grouping module 108 halts the parameter data processing when the overlap is midway between the maximum and minimum overlap. The grouping module 108 then creates groups from a single usage patterns based on the above weighted parameters.

For example, in an embodiment, 'account balance' and 'age' may correspond to caller parameters. When each caller or mobile user calls the IVR system 103, the call details, or usage data logged by the IVR system 103 would include 'account Balance' and 'age' corresponding to the mobile user. Such data is made available to the call flow management module 102 (and the grouping module 108) that creates the graphs as shown in fig. 7 and fig. 8 respectively.

Fig. 7 illustrates a graph 700 between number of callers or mobile users (represented by Y-axis 702) and the parameter 'account balance' (represented by X-axis 704). As shown in the graph 700, callers or mobile users with account balance in the range of 12000 to 24000 use Pattern 1 marked as 706 in the graph. Furthermore, callers or mobile users with account balance lying in the range of 30000 to 37000 use Pattern 2 marked as 708 in figure 7.

Fig. 8 illustrates a graph 800 between number of callers or mobile users (represented by Y-axis 802) and the parameter 'age' (represented by X-axis 804). As shown in the graph 800, callers or mobile users with age between 30 to 55 years use Pattern 1 marked as 806 in the graph and callers with age between 34 to 46 use Pattern 2 marked as 808 in figure 8.

Now, it can be inferred from the graph 800 that using the 'age' parameter, there cannot be an established usage pattern since there is no unique criteria to separate the groups from each other. On the other hand, the 'account balance' parameter does have an effect on the usage pattern of the callers or mobile users. Thus, callers with 'account balance between 12000 to 24000' are more likely to follow Pattern 1 and callers with 'account balance between 30000 to 37000' are more likely to follow Pattern 2. Hence, the grouping module 108 identifies one or more grouping criteria for creating a plurality of groups of a plurality of mobile users. It may be appreciated that more than one grouping criteria can be used for grouping of mobile users. The following table 5 shows an example of groups created based on different criteria in an embodiment.

Table 5:

There is an optimal call flow for each usage pattern and the optimal call flow may be directly generated out of the usage pattern. The call flow optimization module 110 determines an optimized call flow for each usage pattern and stores the same. In an embodiment, an optimized call flow may correspond to that usage pattern which has a maximum percentage access (of the usage pattern) by a mobile user over time. This is based on an assumption that the percentage of access invariably shows the probability of the mobile user accessing the same usage pattern the next time. Over repeated cycles of training, a particular usage pattern can thus be designated as an optimized call flow for the incoming call or the mobile user and for the corresponding group. For instance, fig. 6 illustrates the three sample call flows shown in fig. 4, the corresponding percentage access, and the parameter details 602 for the corresponding call flow.

Table 6 below shows the optimized call flows suitable for the corresponding usage pattern in an embodiment.

Table 6:

The call flow management node 102 stores the usage data, identified usage pattern, a list of usage patterns (including the existing patterns), calculated variances, group IDs (based on grouping of the mobile users), predetermined thresholds for tolerance, percentage access, and optimized call flows in suitable formats. The suitability of data formats and data structures for ease of access and speed of processing may be an important consideration for storing the exemplary data mentioned above. It may be appreciated that any of known data structures may be employed for the purposes of the ongoing description without departing from the scope of the apparatus and methods disclosed herein. For example, the data may be stored as a look up table (LUT) that may be accessed during any stage of operation for multiple purposes. The stored data can be used during subsequent training phases or the selection phase. The call flow management node 102 may be subjected to repeated training cycles after the expiry of a predetermined time to take into consideration, the changing dynamics, and parameter values of the incoming calls or mobile users. A network operator may schedule such training phases regularly to ensure that optimized call flows does not lead to dissatisfaction of the mobile users or the consumers instead of providing then a lesser holding time in the IVR menu. The selection phase corresponds to the live run of the call flow management node 102. In the selection phase, the IVR system 103 receives an incoming call from the mobile device 116. The call flow management node 102 receives the incoming call details from the IVR system 103 and identifies a group associated with the incoming call or the mobile user initiating the incoming call. Subsequently, the call flow management node 102 selects the optimal call flow for the identified group and maps the selected optimal call flow to the incoming call. The call flow management module 102 provides the optimized call flow (e.g. 114) to the IVR system 103 which executes the call flow for the incoming call. Fig. 9 illustrates a method 900 for dynamically configuring an optimized call flow in an

Interactive Voice Response (IVR) system 103. The IVR system 103 being configured for incoming calls from a mobile user subscribed to a network operator in a mobile communications system. It may be noted that the method 900 embodies the training phase and hence the method will be described with specific references to the training phase but is not limiting to any phase as such.

At block 902, usage data associated with a plurality of mobile users subscribed to the network operator is collected. The IVR system 103 extracts usage data (e.g. parameters) and also captures usage pattern, and call flow details for the incoming call. The IVR system 103 forwards such details and usage data to the call flow management node 102. The statistics module 104 collects such usage data from the IVR system 103 for further processing. In an implementation, the usage data includes network parameters, system parameters, and external parameters. The network parameters include location information; the system parameters include one or more of gender, age, current salary, account balance, and outstanding credit. The external parameters may include one or more of stock market value and time of the day.

At block 904, the usage data is analyzed to identify a call flow pattern (or usage pattern) associated with each of the plurality of mobile users. The call flow pattern includes selected menu options, start time, and end time for each module (as shown in fig. 2) of the Interactive Voice Response (IVR) menu 200. In a successive progression, the pattern recognition module 106 identifies a usage pattern based on an analysis of the usage data. The analysis may include one or more comparison of the identified usage pattern with one or more existing patterns. The analysis may further include defining tolerance for the comparison. During the analysis, two or more of call flow patterns may be considered as similar if a pattern variance between the two or more call flow patterns is below a predetermined threshold tolerance. In an embodiment, the analysis includes determining percentage access for each of the determined call flow patterns.

At block 906, the plurality of mobile users is grouped into a plurality of groups based at least in part on the call flow pattern and the usage data. The grouping module 108 groups the mobile user or subscriber based on the usage data and call flow pattern details.

At block 908, a plurality of optimized call flows corresponding to each of the groups are determined. The call flow optimization module 110 determines an optimized call flow for each usage pattern and stores the same. In an embodiment, an optimized call flow may correspond to that usage pattern which has a maximum percentage access (of the usage pattern) by a mobile user over time.

At block 910, an optimized call flow is associated with each of the plurality of groups. The call flow mapping module 112 maps the optimized call flows to each of the plurality of groups as shown in table 6.

Fig. 10 illustrates a method 1000 for dynamically selecting an optimized call flow in an Interactive Voice Response (IVR) system 103 in an embodiment. The IVR system 103 being configured for incoming calls from mobile users subscribed to a network operator in a mobile communications system. In one of the implementations, the method 1000 at least in part describes the selection phase.

At block 1002, an incoming call is received from a mobile user subscribed to a network operator. The mobile user calls the IVR system 103 using the mobile device 114. The call details are logged in by the IVR system 103 and forwarded to the call flow management node 102.

At block 1004, a group associated with the mobile user is identified. The grouping module 102 based on the usage data, identifies the group corresponding to the mobile user. This may involve accessing a data structure such as table 4 and table 5.

At block 1006, the optimized call flow corresponding to the identified group is determined. In an embodiment, the call flow mapping module 112 during such a determination, may access a data structure (e.g. a look-up table) storing group IDs and corresponding optimal call flows (such as table 6).

At block 1008, the optimized call flow is applied to the incoming call. The call flow management node 102 sends the optimized call flow to the IVR system 103 which then applies the optimized call flow to the incoming call.

In yet another implementation, the method 1000 may further include the steps of the training phase. For instance, the method 1000 may include collecting usage data associated with a plurality of mobile users subscribed to the network operator and analyzing the usage data to identify a call flow pattern associated with each of the plurality of mobile users. The method 1000 may also include grouping the plurality of mobile users into a plurality of groups based at least in part on the call flow pattern and the usage data and determining a plurality of optimized call flows corresponding to each of the groups. The method 100 may further include associating an optimized call flow with each of the plurality of groups. The grouping may include mapping each of the one or more parameters with a usage pattern associated with the plurality of mobile users.

Disclosed systems and methods can provide for automatic monitoring of the usage of the IVR system 103 and analysis of the usage data based on customer profile, and reconfiguration of the call flow of the IVR system 103. Such features reduce the call time but still maintain the same functionality of the IVR system 103 as before. The mobile user would not feel any difference due to introduction of the call flow management node 102, with regard to the functionality of the IVR 103, except the reduction of mean call time or holding time. Doing so not only reduces the cost of the IVR platform but also leads to customer satisfaction as customers can get the required service in less time.

Disclosed systems and methods reduces huge processing power required in the current systems by not storing the call flows for individual subscribers but storing optimized call flows for groups rather than individuals. The disclosed systems further help the network service provider to perform logical grouping of the mobile users. The structure of these groups can be made dynamic by multiple and recurring training cycles as compared to the static structure of groups in the current systems. This feature enables automatic reshaping/remodeling of the call flow management node 102 as per the change of the users behavior in time. The disclosed systems and methods also provide inputs for a more focused launch for the new services introduced by network service providers.

It will be appreciated that the teachings of the present document, the disclosed systems, and methods can be implemented as a combination of hardware and software. The software is preferably implemented as an application program comprising a set of program instructions tangibly embodied in a computer readable medium. The application program is capable of being read and executed by hardware such as a computer or processor of suitable architecture. Similarly, it will be appreciated by those skilled in the art that any examples, flowcharts, functional block diagrams and the like represent various exemplary functions, which may be substantially embodied in a computer readable medium executable by a computer or processor, whether or not such computer or processor is explicitly shown. The processor can be a Digital Signal Processor (DSP) or any other processor used conventionally that is capable of executing the application program or data stored on the computer-readable medium.

The example computer-readable medium can be, but is not limited to, (Random Access Memory) RAM, (Read Only Memory) ROM, (Compact Disk) CD or any magnetic or optical storage disk capable of carrying application program executable by a machine of suitable architecture. It is to be appreciated that computer readable media also includes any form of wired or wireless transmission. Further, in another implementation, the method disclosed herein can be incorporated on a hardware medium using ASIC or FPGA technologies.

Aspects of the system disclosed herein may also be implemented in methods and/or computer program products. Accordingly, the disclosed method may be embodied in hardware and/or in hardware/software (including firmware, resident software, microcode, etc.). Furthermore, the method may take the form of a computer program product on a computer- usable or computer-readable storage medium having computer-usable or computer-readable program code embodied in the medium for use by or in connection with an instruction execution system. The actual software code or specialized control hardware used to implement embodiments described herein is not limiting of the described system. Thus, the operation and behavior of the aspects were described without reference to the specific software code - it being understood that one would be able to design software and control hardware to implement the aspects based on the description herein. Furthermore, certain portions of the disclosed methods may be implemented as "logic" that performs one or more functions. This logic may include hardware, such as an application specific integrated circuit or field programmable gate array or a combination of hardware and software.

It is to be appreciated that the subject matter of the claims are not limited to the various examples an language used to recite the principle of the disclosure, and variants can be contemplated for implementing the claims without deviating from the scope. Rather, the embodiments described herein encompass both structural and functional equivalents thereof.

While certain present embodiments and methods of practicing the same have been illustrated and described herein, it is to be distinctly understood that the system disclosed herein is not limited thereto but may be otherwise variously embodied and practiced within the scope of the following claims.

Claims

CLAIMS:

1. A method (900) for dynamically configuring an optimized call flow in an Interactive Voice Response (IVR) system (103), the IVR system (103) being configured for incoming calls from a mobile user subscribed to a network operator in a mobile communications system, the method comprising:

collecting (902) usage data associated with a plurality of mobile users subscribed to the network operator;

analyzing (904) the usage data to identify a call flow pattern associated with each of the plurality of mobile users;

grouping (906) the plurality of mobile users into a plurality of groups based at least in part on the call flow pattern and the usage data;

determining (908) a plurality of optimized call flows corresponding to each of the groups; and

associating (910) an optimized call flow with each of the plurality of groups.

2. The method according to claim 1, wherein the call flow pattern comprises selected menu options, start time, and end time for each module of the Interactive Voice Response (IVR) menu (200).

3. The method according to claim 1, wherein the usage data comprises network parameters, system parameters, and external parameters.

4. The method according to claim 3, wherein network parameters include location information.

5. The method according to claim 3, wherein system parameters include one or more of gender, age, current salary, account balance, and outstanding credit.

6. The method according to claim 3, wherein external parameters include one or more of stock market value and time of the day.

7. The method according to claim 1, wherein the analyzing comprises determining call flow pattern subsequent to a call to the IVR system (103).

8. The method according to claim 7, wherein the analyzing comprises considering two or more of call flow patterns as similar if a pattern variance between the two or more call flow patterns is below a predetermined threshold tolerance.

9. The method according to claim 7, wherein the analyzing comprises determining percentage access for each of the determined call flow patterns.

10. A call flow management node (102) for dynamically configuring call flows in an Interactive Voice Response (IVR) system (103), the IVR system (103) being configured for mobile users subscribed to an operator network, the system comprising:

a statistics collection module (104) configured to collect usage data associated with a plurality of mobile users from the IVR system (103);

a pattern recognition module (106) configured to identify a usage pattern of each of the plurality of mobile users;

a grouping module (108) configured to group the plurality of mobile users into a plurality of groups based at least in part on the usage pattern and the usage data;

a call flow optimization module (110) configured to determine optimal call flows corresponding to each group of the plurality of mobile users; and

a call flow mapping module (112) configured to associate an optimized call flow with each of the plurality of groups.

11. The call flow management node (102) according to claim 10, wherein the usage pattern comprises selected menu options, start time, and end time for each module of Interactive Voice Response (IVR) menu (200).

12. The call flow management node (102) according to claim 10, wherein the usage data comprises network parameters, system parameters, and external parameters.

13. The call flow management node (102) according to claim 12 wherein network parameters include location information.

14. The call flow management node (102) according to claim 12, wherein system parameters include one or more of gender, age, current salary, account balance, and outstanding credit.

15. The call flow management node (102) according to claim 12, wherein external parameters include one or more of stock market value and time of day.

16. The call flow management node (102) according to claim 10, wherein the call flow mapping module (112) being further configured to select and map one of the optimized call flows (114) to an incoming call from a mobile user to the IVR system (103) based on a determination of the group to which the mobile user belongs to.

17. A method for dynamically selecting an optimized call flow in an Interactive Voice Response (IVR) system (103), the IVR system (103) being configured for incoming calls from mobile users subscribed to a network operator in a mobile communications system, the method comprising:

receiving an incoming call from a mobile user subscribed to a network operator; identifying a group associated with the mobile user;

determining the optimized call flow corresponding to the identified group; and applying the optimized call flow to the incoming call.

18. The method according to claim 17 further comprising:

associating (910) an optimized call flow with each of the plurality of groups.

19. The method according to claim 18, wherein the one or more parameters comprise network parameters, system parameters, and external parameters.

20. The method according to claim 18, wherein the grouping comprises mapping each of the one or more parameters with a usage pattern associated with the plurality of mobile users.

21. The method according to claim 17, wherein the determining comprises accessing a lookup table storing group IDs, and corresponding optimal call flows.