Process optimization method and apparatus, and device

Abstract

The present application provides a process optimization method and apparatus, and a device. The apparatus where a process optimization module is located determines a target sub-IVR process to be optimized from among different sub-IVR processes on the basis of values of service parameters of the sub-IVR processes and reference values of the service parameters, determines, during optimization of the target sub-IVR process, a target node associated with the corresponding service parameter from among a plurality of nodes comprised in the target sub-IVR process, selects a target optimization scheme for optimizing the target node from an optimization scheme set, and optimizes the target sub-IVR process by optimizing the target node in the target sub-IVR process on the basis of the target optimization scheme. Because the target node is a node associated with the corresponding service parameter, by optimizing the target node, the value of the service parameter of the optimized target sub-IVR process is improved to the maximum extent, thereby achieving quick and convenient IVR process optimization.

Description

Process optimization method, device and equipment

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application claims priority to the Chinese patent application filed with the State Intellectual Property Office of the People's Republic of China on December 28, 2023, with application number 202311855810.0 and application name "A process optimization method, device and equipment", all contents of which are incorporated by reference into this application. Technical Field

[0003] The present application relates to the field of data processing technology, and in particular to a process optimization method, device and equipment. Background Art

[0004] Self-service processes, also known as interactive voice response (IVR), are a type of self-service offered by corporate call centers to customers and play a crucial role in their operations. Optimizing the service flow within the IVR (hereinafter referred to as the IVR process for ease of description) is a key task for call center operations managers and a crucial component in improving the operational efficiency of enterprise call center systems.

[0005] In related technologies, optimizing IVR processes typically requires operations managers to manually analyze business metrics collected by reporting and statistical tools and develop optimization solutions. However, this manual approach is inefficient and labor-intensive. Therefore, a technical solution that can quickly and easily optimize IVR processes is urgently needed.

[0006] Summary of the Invention

[0007] The embodiments of the present application provide a process optimization method, apparatus, and device for quickly and easily optimizing an IVR process.

[0008] In a first aspect, an embodiment of the present application provides a process optimization method, which can be executed by a device where a process optimization module is located. In this method:

[0009] The device where the process optimization module is located can obtain the values ​​of the business parameters of each sub-IVR process, and based on the difference between the values ​​of the business parameters of each sub-IVR process and the reference values ​​corresponding to the business parameters, determine the target sub-IVR process to be optimized from each sub-IVR process; when optimizing the target sub-IVR process, the target node associated with the business parameter can be determined from the multiple nodes included in the target sub-IVR process, and the target optimization scheme that can be used to optimize the target node can be selected from the optimization scheme set, thereby achieving optimization of the target sub-IVR process.

[0010] Through the above-mentioned method, the device where the process optimization module is located can intelligently determine the target sub-IVR process to be optimized from each sub-IVR process based on the values ​​of the business parameters of each sub-IVR process and the reference values ​​of the business parameters. When optimizing the target sub-IVR process to be optimized, the device where the process optimization module is located can first determine the target node associated with the business parameters from the multiple nodes included in the target sub-IVR process, and select a target optimization scheme that can optimize the target node from the optimization scheme set. The target sub-IVR process can be optimized by optimizing the target node in the target sub-IVR process based on the target optimization scheme. Since the determined target node is a node associated with the business parameter, by optimizing these target nodes, the value of the business parameter of the optimized target sub-IVR process can be improved to the greatest extent, thereby achieving the purpose of quickly and easily optimizing the IVR process.

[0011] In one possible implementation, when the device where the process optimization module is located determines the target node associated with the business parameter from the multiple nodes included in the target sub-IVR process, it can be based on the association relationship between the business parameters of the sub-IVR process and the indicator parameters of the node obtained in advance to determine the target node associated with the business parameter from the multiple nodes included in the target sub-IVR process.

[0012] In this way, based on the pre-acquired association between the service parameters of the sub-IVR process and the index parameters of the node, the target node associated with the service parameters can be quickly and accurately determined, thereby improving the speed and accuracy of IVR process optimization.

[0013] In one possible implementation, when determining the target node, the device where the process optimization module is located may first obtain a correlation coefficient that can characterize the correlation between the indicator parameters and the business parameters of each node included in the target sub-IVR process based on the pre-obtained correlation between the business parameters of the sub-IVR process and the indicator parameters of the node, and then determine the target node based on the correlation coefficient.

[0014] Through the above-mentioned manner, the speed and accuracy of determining the target node can be improved based on the correlation coefficient that can characterize the correlation between the indicator parameter of the node and the business parameter.

[0015] In one possible implementation, when determining the target node based on the above-mentioned correlation coefficient, the device where the process optimization module is located can first obtain a ranking position that can characterize the degree of excellence between each node based on the correlation coefficient and the value of the indicator parameter of each node included in the target sub-IVR process; and then, based on the ranking position between each node, a set number of target nodes to be optimized can be determined from each node.

[0016] Through the above method, since the target node can be selected based on a comprehensive evaluation of multiple dimensions such as the correlation coefficient and the node's index parameters, the accuracy of determining the target node can be improved to the greatest extent possible. By adjusting the target node selected based on this method, the values ​​of the business parameters of the optimized target sub-IVR process can be improved to the greatest extent, thereby achieving the purpose of optimizing the IVR process quickly and accurately.

[0017] In one possible implementation, when determining (obtaining) the ranking order that characterizes the superiority or inferiority between each node, the device where the process optimization module is located may first obtain the sub-ranking position of each node in terms of each indicator parameter based on the value of the indicator parameter taken by each node included in the target sub-IVR process; and then obtain the ranking position that can characterize the superiority or inferiority between each node based on the weighted value between each correlation coefficient and the sub-ranking position of each node.

[0018] Through the above method, based on each correlation coefficient and the sub-ranking position of each node in each indicator parameter, the ranking position that can characterize the degree of superiority and inferiority between each node can be accurately obtained, thereby improving the accuracy of determining the target node to be optimized. By adjusting the target node selected based on this method, the value of the business parameter of the optimized target sub-IVR process can be improved to the greatest extent, thereby realizing quick and accurate optimization of the IVR process.

[0019] In one possible implementation, the association between the business parameters of the sub-IVR process and the index parameters of the node obtained by the device where the process optimization module is located can be obtained based on a trained association model. The business parameters of the sub-IVR process can be input into the trained association model. The association model can determine the index parameters of the node associated with the business parameters of the input sub-IVR process based on the business parameters of the input sub-IVR process.

[0020] In one possible implementation, the training process of the association relationship model includes:

[0021] The device where the first model training module is located obtains the sample business parameters of any sub-IVR process in the sample set, and the sample business parameters correspond to the sample labels of the correlation coefficients between the sample business parameters of the sub-IVR process and the index parameters of each node in the sub-IVR process. The device where the first model training module is located inputs the obtained sample business parameters of the sub-IVR process into the correlation model to be trained, and the correlation model operates on the input sample business parameters of the sub-IVR process, determines the index parameters of the node associated with the sample business parameters of the sub-IVR process, and outputs the correlation coefficient identification labels between the sample business parameters of the sub-IVR process and the index parameters of each node in the sub-IVR process. The device where the first model training module is located trains the correlation model to be trained based on the correlation coefficient identification labels and the correlation coefficient sample labels.

[0022] Through the above method, based on the trained association model, the association between the business parameters of the sub-IVR process and the indicator parameters of the node can be quickly and accurately obtained. Based on this association, the accuracy of determining the target node can be improved, thereby improving the speed and accuracy of IVR process optimization.

[0023] In one possible implementation, when the device where the process optimization module is located selects a target optimization scheme from a set of optimization schemes, it can be based on the correspondence between the index parameters of the nodes of the sub-IVR process obtained in advance and the optimization schemes, and determine the optimization scheme in the optimization scheme set that corresponds to the value of the index parameter of the target node of the target sub-IVR process as the target optimization scheme.

[0024] Through the above method, the determined target optimization scheme can be the optimization scheme corresponding to the value of the indicator parameter of the target node of the target sub-IVR process in the optimization scheme set. The target optimization scheme can be better applied to the target node. When the target node in the target sub-IVR process is optimized based on the target optimization scheme, the purpose of optimizing the target sub-IVR process can be achieved to the greatest extent possible.

[0025] In one possible implementation, the device where the process optimization module is located can determine one or more optimization schemes corresponding to the values ​​of the indicator parameters of the target node of the target sub-IVR process obtained based on the optimization scheme library model as candidate schemes, and can display each candidate scheme. The operation management personnel can select a scheme from the candidate schemes as the target optimization scheme, or can modify a candidate scheme. The device where the process optimization module is located can identify the candidate scheme selected by the operation management personnel (user) or the candidate scheme modified by the user, and can determine the candidate scheme selected by the user or the candidate scheme modified by the user as the target optimization scheme.

[0026] Through the above-mentioned manner, the candidate solution selected by the user or the candidate solution modified by the user can be determined as the target optimization solution, which can improve the flexibility and accuracy of determining the target optimization solution.

[0027] In one possible implementation, the device where the process optimization module is located can obtain one or more optimization schemes corresponding to the values ​​of the indicator parameters of the target node of the target sub-IVR process based on the optimization scheme library model, and use these optimization schemes as candidate schemes. Each candidate scheme carries priority information, and the target optimization scheme can be selected from the candidate schemes based on the priority information.

[0028] Through the above approach, the flexibility and accuracy of determining the target optimization solution can be improved based on the priority information of the candidate solutions.

[0029] In one possible implementation, the correspondence between the index parameters of the nodes of the sub-IVR process and the optimization scheme pre-obtained by the device where the process optimization module is located is obtained based on a trained optimization scheme library model. The values ​​of the index parameters of the nodes of the sub-IVR process can be input into the optimization scheme library model, and the optimization scheme library model can determine the optimization scheme corresponding to the values ​​of the index parameters of the nodes of the input sub-IVR process.

[0030] In one possible implementation, the training process of the optimization solution library model includes:

[0031] The device where the second model training module is located obtains the values ​​of the sample index parameters of several nodes of any sub-IVR process in the sample set, and the sample index parameters correspond to the optimization solution sample labels applicable to the values ​​of the sample index parameters of these several nodes of the sub-IVR process. The device where the second model training module is located can input the obtained values ​​of the sample index parameters of several nodes of the sub-IVR process into the optimization solution library model to be trained. The optimization solution library model calculates the values ​​of the sample index parameters of several nodes of the input sub-IVR process, determines the optimization solution applicable to the values ​​of the sample index parameters of several nodes of the sub-IVR process, and outputs the optimization solution identification label. The device where the second model training module is located uses the optimization solution identification label and the optimization solution sample label to train the optimization solution library model to be trained.

[0032] Through the above method, based on the trained optimization solution library model, the correspondence between the indicator parameters of the nodes of the sub-IVR process and the optimization solution can be quickly and accurately obtained. Based on this correspondence, the accuracy of determining the target optimization solution can be improved, thereby improving the speed and accuracy of IVR process optimization.

[0033] In one possible implementation, when determining a target sub-IVR process from each sub-IVR process, the device where the process optimization module is located may, for each sub-IVR process, count the difference between the value of each service parameter of the sub-IVR process and the corresponding reference value; and count the weighted value between the preset weight of each service parameter of the sub-IVR process and the corresponding difference; in addition, the reference weighted value between the reference value of each service parameter and the preset weight may be counted, and the target sub-IVR process may be determined from each sub-IVR process based on the difference between the above-mentioned weighted value of each sub-IVR process and the reference weighted value.

[0034] Through the above method, the preset weights of the business parameters and the difference between the values ​​of the business parameters and the reference values ​​can be combined to comprehensively measure the target sub-IVR process that needs to be optimized from multiple perspectives, so that the target sub-IVR process to be optimized can be flexibly and accurately determined.

[0035] In one possible implementation, the device where the process optimization module is located can also perform a comparative test on the optimized target sub-IVR process and the target sub-IVR process before optimization. If the test results show that the values ​​of the business parameters of the optimized target sub-IVR process are inferior to the values ​​of the business parameters of the target sub-IVR process before optimization, the target sub-IVR process before optimization can still be adopted; if the test results show that the values ​​of the business parameters of the optimized target sub-IVR process are superior to the values ​​of the business parameters of the target sub-IVR process before optimization, the optimized target sub-IVR process can be adopted.

[0036] Through the above-mentioned method, it can be ensured to the greatest extent that the currently adopted target sub-IVR process is a sub-IVR process with better service parameter values, thereby ensuring the execution effect of the sub-IVR process to the greatest extent.

[0037] In one possible implementation, the device where the process optimization module is located can also obtain a reference correlation coefficient between the indicator parameters of each node and each business parameter in the target sub-IVR process based on the values ​​of the business parameters corresponding to the optimized target sub-IVR process and the target sub-IVR process before optimization and the values ​​of the indicator parameters of the nodes contained in the test results; based on the reference correlation coefficient, the correlation model is optimized and trained, etc., to optimize and update the correlation between the business parameters of the sub-IVR process and the indicator parameters of the nodes previously obtained based on the correlation model.

[0038] Through the above methods, we can continuously accumulate experience in optimizing the IVR process, continuously optimize the association model, continuously optimize the association between the business parameters of the sub-IVR process and the indicator parameters of the node, and continuously improve the optimization effect of the IVR process.

[0039] In a possible implementation, the device where the process optimization module is located may also optimize and update the correspondence between the obtained indicator parameters of the nodes of the sub-IVR process and the optimization solution based on the test results.

[0040] Through the above method, we can continuously accumulate optimization experience for the IVR process, continuously optimize the above optimization solution library model, continuously optimize the correspondence between the indicator parameters of the nodes of the sub-IVR process and the optimization solution, and continuously improve the optimization effect of the IVR process.

[0041] In a second aspect, an embodiment of the present application further provides a process optimization device, which has the function of implementing the behavior of the device where the process optimization module is located in the method example of the first aspect above. The beneficial effects can be found in the description of the first aspect and will not be repeated here. The functions can be implemented by hardware or by hardware executing corresponding software. The hardware or software includes one or more modules corresponding to the above functions. The structure of the process optimization device may include a process optimization module.

[0042] The process optimization module is used to: obtain the value of at least one business parameter of each sub-IVR process in the IVR process within the target period, and determine a target sub-IVR process from each sub-IVR process based on the value of each business parameter and the reference value corresponding to each business parameter; determine a target node associated with the business parameter in the target sub-IVR process from multiple nodes included in the target sub-IVR process; select a target optimization solution from the optimization solution set, the target optimization solution being used to optimize the target node; and optimize the target sub-IVR process based on the target optimization solution.

[0043] As a possible implementation, the process optimization module may determine the target node based on the association between the pre-obtained service parameters of the sub-IVR process and the index parameters of the node.

[0044] As a possible implementation method, the process optimization module can obtain each correlation coefficient based on the correlation between the business parameters of the sub-IVR process and the indicator parameters of the node obtained in advance. Each correlation coefficient represents the correlation between any indicator parameter and the business parameter of each node included in the target sub-IVR process; the target node can be determined based on each correlation coefficient.

[0045] As a possible implementation method, the process optimization module can obtain a ranking position that represents the degree of excellence between each node based on each correlation coefficient and the value of the indicator parameter of each node included in the target sub-IVR system; based on the ranking position between each node, determine a set number of target nodes to be optimized from each node.

[0046] As a possible implementation method, the process optimization module can obtain the sub-ranking position of each node in each indicator parameter based on the value of each node in the indicator parameter; and can obtain the ranking position that represents the degree of excellence between each node based on each correlation coefficient and the sub-ranking position of each node.

[0047] As a possible implementation method, the association relationship between the business parameters of the sub-IVR process and the index parameters of the node is obtained based on a trained association model. The association model can determine the index parameters of the node associated with the business parameters of the input sub-IVR process based on the business parameters of the input sub-IVR process.

[0048] As a possible implementation method, the process optimization module can determine the optimization scheme corresponding to the value of the indicator parameter of the target node of the target sub-IVR process in the optimization scheme set as the target optimization scheme based on the correspondence between the indicator parameters of the nodes of the sub-IVR process obtained in advance and the optimization scheme.

[0049] As a possible implementation method, the process optimization module can determine at least one optimization scheme in the optimization scheme set that corresponds to the value of the indicator parameter of the target node of the target sub-IVR process as a candidate scheme, and display each of the candidate schemes; determine the target optimization scheme based on the candidate scheme selected by the user or the candidate scheme modified by the user.

[0050] As a possible implementation method, the correspondence between the index parameters of the nodes of the sub-IVR process and the optimization scheme is obtained based on the trained optimization scheme library model. The optimization scheme library model can determine the optimization scheme corresponding to the value of the index parameter of the node of the input sub-IVR process based on the value of the index parameter of the node of the input sub-IVR process.

[0051] As a possible implementation method, the process optimization module can count the difference between the value of each business parameter of each sub-IVR process and the reference value of the corresponding business parameter for each sub-IVR process; and count the weighted value between the preset weight of each business parameter of the sub-IVR process and the corresponding difference; based on the weighted value of each sub-IVR process and the reference weighted value between the reference value of each business parameter and the preset weight, determine the target sub-IVR process from each sub-IVR process.

[0052] As a possible implementation method, the process optimization module can also perform a comparative test on the optimized target sub-IVR process and the target sub-IVR process before optimization. If the values ​​of the business parameters of the optimized target sub-IVR process in the test results are worse than the values ​​of the business parameters of the target sub-IVR process before optimization, the currently used target sub-IVR process is restored to the target sub-IVR process before optimization; otherwise, if the values ​​of the business parameters of the optimized target sub-IVR process in the test results are better than the values ​​of the business parameters of the target sub-IVR process before optimization, the optimized target sub-IVR process is adopted.

[0053] As a possible implementation method, the process optimization module can also obtain a reference correlation coefficient between the index parameters of each node and each business parameter in the target sub-IVR process based on the values ​​of the business parameters corresponding to the optimized target sub-IVR process and the target sub-IVR process before optimization and the values ​​of the index parameters of the nodes contained in the test results; based on the reference correlation coefficient, the correlation between the obtained business parameters of the sub-IVR process and the index parameters of the nodes is optimized and updated.

[0054] As a possible implementation method, the process optimization module may also optimize and update the correspondence between the obtained indicator parameters of the nodes of the sub-IVR process and the optimization solution based on the test results.

[0055] In a third aspect, the present application further provides a computing device comprising a processor and a memory, and may further comprise a communication interface, wherein the processor executes program instructions in the memory to perform the method provided in the first aspect or any possible implementation of the first aspect. The memory is coupled to the processor and stores program instructions and data necessary for determining the process optimization process. The communication interface is used to communicate with other devices, such as obtaining the values ​​of the business parameters of each sub-IVR process, obtaining the association between the business parameters of the sub-IVR process and the indicator parameters of the node, obtaining the correspondence between the indicator parameters of the node of the sub-IVR process and the optimization solution, etc.

[0056] In a fourth aspect, the present application provides a computing device system comprising at least one computing device. Each computing device comprises a memory and a processor. The processor of at least one computing device is configured to access code in the memory to execute the method provided in the first aspect or any possible implementation of the first aspect.

[0057] In a fifth aspect, the present application provides a computer-readable storage medium. When the computer-readable storage medium is executed by a computing device, the computing device performs the method provided in the first aspect or any possible implementation of the first aspect. The storage medium stores a program. The storage medium includes, but is not limited to, volatile memory, such as random access memory, and non-volatile memory, such as flash memory, a hard disk drive (HDD), and a solid state drive (SSD).

[0058] In a sixth aspect, the present application provides a computing device program product, comprising computer instructions that, when executed by a computing device, cause the computing device to perform the method provided in the aforementioned first aspect or any possible implementation of the first aspect. The computer program product may be a software installation package, and when the method provided in the aforementioned first aspect or any possible implementation of the first aspect is required, the computer program product may be downloaded and executed on the computing device.

[0059] In a seventh aspect, the present application also provides a computer chip, which is connected to a memory, and is used to read and execute a software program stored in the memory, and to execute the method described in the above-mentioned first aspect and various possible implementation methods of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] FIG1 is a schematic diagram of the architecture of a process optimization system provided in an embodiment of the present application;

[0061] FIG2 is a schematic diagram of the architecture of another process optimization system provided in an embodiment of the present application;

[0062] FIG3 is a schematic diagram of a process optimization method provided in an embodiment of the present application;

[0063] FIG4 is a schematic diagram of another process optimization method provided in an embodiment of the present application;

[0064] FIG5 is a schematic structural diagram of a process optimization device provided in an embodiment of the present application;

[0065] 6 and 7 are schematic diagrams of the structure of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0066] Before introducing a process optimization method, device and equipment according to the embodiments of the present application, some concepts involved in the embodiments of the present application are first explained.

[0067] (1) Interactive Voice Response (IVR)

[0068] IVRs provide automated voice services and are the primary tool for enterprise call centers to provide self-service to customers. As the gateway to a call center, IVRs were initially designed to serve two purposes: first, to identify customer needs and direct them through various keystrokes; and second, to reduce increasing labor costs by providing automated voice services. IVRs utilize a customer-oriented voice directory, enabling information queries and command execution based on customer selections (either through the telephone keypad or voice input). Customers can conduct self-service transactions through direct voice input.

[0069] The IVR process can include multiple sub-IVR processes, where different sub-IVR processes can handle different services. For example, after a customer enters the IVR process by making a phone call, he or she can select the sub-IVR process for handling mobile phone services or the sub-IVR process for handling laptop services according to the voice prompt.

[0070] (2) Business parameters of the IVR process (sub-IVR process).

[0071] The business parameters of the IVR process, also known as the KPIs for the IVR process, can be used to reflect the rationality of the IVR process and whether the IVR process meets customer needs. These business parameters can include the IVR diversion rate, the IVR first-call resolution rate, and the IVR repeat call rate.

[0072] Among them, the IVR diversion rate usually refers to: within a period of time, the proportion (ratio) of the number of calls completed through self-service in the IVR process in the total number of calls. The total number of calls is the sum of the number of calls completed through self-service in the IVR process (for the convenience of description, referred to as the first call number) and the number of calls that were not resolved by the self-service in the IVR process and were transferred to manual customer service for processing (for the convenience of description, referred to as the second call number). That is, the total number of calls can be the sum of the first call number and the second call number.

[0073] IVR diversion rate is a key indicator for evaluating IVR effectiveness. Generally speaking, a higher IVR diversion rate indicates that the corresponding IVR process can better provide self-service to customers.

[0074] The IVR first-time resolution rate generally refers to the proportion (ratio) of the number of calls within a period of time in which a customer's contact needs are satisfactorily resolved in the first self-service contact through the IVR process, without the need for follow-up work, callbacks, or secondary confirmation, and which can solve the customer's problem in one go, to the above-mentioned number of first calls. The number of first calls is the number of calls completed through the self-service in the IVR process within a period of time.

[0075] As you can see, the IVR first-time resolution rate represents the rate at which customer issues are resolved immediately. It's a key indicator of IVR system efficiency and customer satisfaction. Generally speaking, a higher IVR first-time resolution rate indicates that the IVR process is better at providing self-service and meeting customer needs.

[0076] The IVR repeat call rate is also known as the IVR repeat call rate. The call center has a data backend that collects IVR transaction logs and IVR contact history, presenting this data in detailed reports. The detailed reports can be filtered to identify unique customer identifiers, thereby counting the number of customers and the total number of calls from these customers within a certain time period. The formula for calculating the repeat call rate is as follows:

[0077] IVR repeat call rate = (total number of calls with recorded customer numbers - total number of customers with recorded customer numbers) / total number of calls with recorded customer numbers.

[0078] Generally speaking, the lower the IVR repeat call rate, the better the corresponding IVR process can provide customers with self-service and the more it can meet customer needs.

[0079] Sometimes, a single business parameter may not objectively reflect (measure) the rationality (quality) of the IVR process. For example, a certain IVR process may have a high IVR diversion rate. However, this may simply be because the IVR process lacks an option to transfer to manual customer service, or the option is buried too deep for users to find, resulting in the false impression of a high IVR diversion rate. In this case, multiple business parameters such as the IVR diversion rate and the IVR first-time resolution rate can be combined to measure the quality of the IVR process.

[0080] (3) Node index parameters.

[0081] A node can be understood as each link required for handling business in an IVR process (sub-IVR process). For example, a node may include inputting an ID number, inputting a password, etc. The node's indicator parameters can be used to reflect whether the settings of each node in the IVR process are reasonable. The node's indicator parameters may include: node surfing volume, node surfing rate, node repeat playback volume (also known as node repeat visit volume), node direct rate, node visit rate, node failure volume, node key press sequence, etc. Among them, this application does not limit the specific content of the nodes and indicator parameters. The nodes and indicator parameters listed here are only examples. The nodes and indicator parameters are related to the specific IVR process.

[0082] Node surfing volume generally refers to the number of visits from the current node's successor nodes (lower-level nodes) back to the current node. This is often caused by customers visiting the subsequent node but finding that they cannot solve the problem and need to return to the predecessor node (upper-level node) to search for the required service. Node surfing volume can reflect the rationality of the current node configuration and the usability of the IVR process. A higher node surfing volume indicates a more unreasonable node configuration in the IVR process.

[0083] The node surfing rate is typically calculated as the percentage of the current node's node surfing volume relative to the current node's total node visits (total node visits). Similar to the node surfing volume, a higher node surfing rate indicates a less rational node setup within the IVR process. The total node visits volume will be explained later, so I won't elaborate on it here.

[0084] Node repeat playback typically refers to the number of times the same node is accessed repeatedly (by selecting the replay button). Node repeat playback reflects the clarity and accuracy of the node's voice prompts. A high node repeat playback rate typically indicates that the node prompts are unclear, inaccurate, or contain content that the user struggles to understand.

[0085] The node direct rate usually refers to the proportion of visits to the current node (node ​​direct volume) in the direction indicated by the access hierarchy relationship between nodes (forward access) to all visits to the current node (total node visits). The node direct rate can reflect the frequency of node visits and the effectiveness of the node. Generally speaking, the higher the node direct rate, the more reasonable the node settings in the IVR process are. The node direct rate can be expressed by the formula: node direct rate = node direct volume / total node visits, where total node visits = node direct volume + node surfing volume + node repeated playback volume.

[0086] Node failures generally refer to the number of visits that exit the IVR process without resolving the customer's issue (unsuccessful processing). Generally speaking, a higher node failure rate indicates a more inappropriate node configuration within the IVR process.

[0087] The node key sequence generally indicates the order in which nodes are accessed within the IVR process. Generally, nodes that are earlier in the key sequence are more easily accessible to customers.

[0088] The following is a brief introduction to the design concept of the embodiment of this application:

[0089] In related technologies, report statistics tools can collect statistics and analyze access logs of customers accessing IVR processes, presenting business indicators related to the IVR process. The business indicators can include the values ​​of the business parameters of each sub-IVR process contained in the IVR process, as well as the values ​​of the indicator parameters of the nodes contained in each sub-IVR process. When optimizing the IVR process, operations management personnel usually need to manually analyze the business indicators counted by the report statistics tool based on the operations management personnel, and manually summarize and conclude the optimization plan for the IVR process. After the operations management personnel manually summarize and conclude the optimization plan for the IVR process, IVR process customization developers or business management personnel who master IVR process orchestration technology can manually orchestrate the optimized IVR process online or offline based on the business processing logic requirements in the optimization plan based on the IVR orchestration tool. The optimized IVR process can be loaded and run based on the tools provided by the IVR process running environment, so that the optimized IVR process can be used by customers, such as processing customer call requests, recording customer access logs to the IVR process, etc.

[0090] In related technologies, the process of determining an optimization plan and optimizing the IVR process according to the optimization plan must be done manually, which is inefficient and has high labor costs.

[0091] In view of this, the present application proposes a process optimization method, device and apparatus for quickly and easily optimizing the IVR process. Referring to FIG1 , which is a schematic diagram of the architecture of a process optimization system provided in an embodiment of the present application, the process optimization system includes a client 200 and an apparatus 100 in which a process optimization module 101 is located.

[0092] The client 200 has data input and data acquisition functions. The client 200 can be deployed on the user side. The operation manager (user) of the IVR process can input the IVR process optimization goals through the client 200. The IVR process optimization goals input by the operation manager can include the values ​​that the IVR diversion rate, IVR first solution rate and other business parameters want to achieve (for the convenience of description, called reference values) and the weights corresponding to each business parameter (for the convenience of description, called preset weights), etc. The client 200 can obtain the IVR process optimization goals input by the operation manager.

[0093] There is a connection between the client 200 and the device 100 where the process optimization module 101 is located. The client 200 can transmit the data of the acquired IVR process optimization target to the device 100 where the process optimization module 101 is located, so that the device 100 where the process optimization module 101 is located can determine the target sub-IVR process to be optimized from each sub-IVR process based on the IVR process optimization target.

[0094] The client 200 can also have a transmission function. Since the client 200 is closer to the operation management personnel (users), the client 200 can also transmit the optimization records of the IVR process to the users, and can also display candidate solutions when the user's assistance in determining the target optimization solution is needed. The user can manually select or modify the candidate solution through the client 200. The device 100 where the process optimization module 101 is located can determine the candidate solution selected or modified by the user through the client 200 as the target optimization solution.

[0095] The device 100 where the process optimization module 101 is located can optimize the IVR process. Still referring to FIG1 , the device 100 where the process optimization module 101 is located can include: the process optimization module 101 , the report statistics tool module 102 , the IVR arrangement tool module 103 , and the IVR operating environment providing tool module 104 .

[0096] Among them, the report statistics tool module 102 can count and analyze the access logs of customers accessing the IVR process, and present business indicators related to the IVR process. The business indicators can include the values ​​of the business parameters of each sub-IVR process included in the IVR process, as well as the values ​​of the indicator parameters of the nodes included in each sub-IVR process, etc.

[0097] The process optimization module 101 can determine the target sub-IVR process to be optimized from each sub-IVR process based on the values ​​of the business parameters of each sub-IVR process and the reference values ​​of the business parameters configured by the operation management personnel, determine the target node associated with the business parameters from the nodes included in the target sub-IVR process, and select a target optimization scheme that can optimize the target node from the optimization scheme set.

[0098] The IVR orchestration tool module 103 may provide an IVR orchestration tool for orchestrating and optimizing the IVR process. In an embodiment of the present application, after the process optimization module 101 determines the target optimization solution, it may generate optimization operation instructions for the target nodes based on the adjustment information for the nodes in the target sub-IVR process included in the target optimization solution, thereby driving the IVR orchestration tool module 103. Based on the IVR orchestration tool and the optimization operation instructions, the IVR orchestration tool module 103 adjusts (optimizes) the key sequence, playback content, etc. of the nodes included in the target sub-IVR process, thereby automatically and intelligently orchestrating and optimizing the target sub-IVR process.

[0099] Among them, the process optimization module 101 can also be called the module where the IVR analysis and optimization tool is located. The process optimization module 101 and the IVR orchestration tool module 103 can be two independent modules or integrated into one module. This application does not make any specific restrictions on this.

[0100] The IVR operating environment providing tool module 104 may be used to load and run the optimized IVR process so that the optimized IVR process can be used by the customer, for example, to process the customer's call request, record the customer's access log to the IVR process, etc.

[0101] The embodiments of the present application do not limit the specific form of the device 100 where the process optimization module is located. The device 100 where the process optimization module is located may be a hardware device, such as a computing device, a computing device cluster, or a chip or processor in a computing device. The device 100 where the process optimization module is located may also be a software device, such as process optimization software, a container, or a virtual machine deployed on one or more computing devices.

[0102] In Figure 1, only the device 100 where the process optimization module is located interacts with the user through the client 200 as an example. In actual applications, the device 100 where the process optimization module is located can also be directly deployed on the user side to interact directly with the user to obtain the IVR process optimization goals provided by the user, as well as the candidate solutions selected or modified by the user.

[0103] In addition, the modules contained in the device 100 where the process optimization module is located can also be deployed in a plurality of independent devices in a distributed manner. Refer to Figure 2, which is a schematic diagram of the architecture of another process optimization system provided in an embodiment of the present application. The process optimization system includes a device 100 where the process optimization module 101 is located, a device 300 where the report statistics tool module 102 is located, a device 400 where the IVR orchestration tool module 103 is located, and a device 500 where the IVR operating environment providing tool module 104 is located. Whether the four modules, the process optimization module 101, the report statistics tool module 102, the IVR orchestration tool module 103, and the IVR operating environment providing tool module 104, are integrated into the same device (such as integrated into the device 100 where the process optimization module 101 is located), or are distributed and deployed in multiple devices, the roles played by these four modules can remain unchanged. For example: the device 300 where the report statistics tool module 102 is located can perform statistics and analysis on the access log of the customer's access to the IVR process based on the report statistics tool module 102, and present business indicators related to the IVR process. The business indicators may include the values ​​of the business parameters of each sub-IVR process included in the IVR process, and the values ​​of the indicator parameters of the nodes included in each sub-IVR process, etc.

[0104] The device 100 where the process optimization module 101 is located can, based on the process optimization module 101, determine the target sub-IVR process to be optimized from each sub-IVR process based on the values ​​of the business parameters of each sub-IVR process and the reference values ​​of the business parameters configured by the operation management personnel, determine the target node associated with the business parameters from the nodes included in the target sub-IVR process, and select a target optimization scheme that can optimize the target node from the optimization scheme set.

[0105] The device 400 where the IVR orchestration tool module 103 is located can provide an IVR orchestration tool for orchestrating and optimizing the IVR process. After the process optimization module 101 determines the target optimization plan, it can generate optimization operation instructions for the target node based on the adjustment information of the node in the target sub-IVR process contained in the target optimization plan. The device 100 where the process optimization module 101 is located drives the device 400 where the IVR orchestration tool module 103 is located, so that the IVR orchestration tool in the IVR orchestration tool module 103 can adjust (optimize) the key sequence, playback content, etc. of the nodes contained in the target sub-IVR process according to the optimization operation instructions, thereby realizing automatic and intelligent orchestration of the optimized target sub-IVR process. The device 500 where the IVR operating environment providing tool module 104 is located can load and run the optimized IVR process based on the IVR operating environment providing tool module 104, so that the optimized IVR process can be used by customers, which will not be repeated here.

[0106] The following describes a process optimization method provided by an embodiment of the present application in conjunction with Figure 3. For ease of understanding, the embodiment of the present application takes the integration of the four modules, namely the process optimization module 101, the report statistics tool module 102, the IVR arrangement tool module 103, and the IVR operating environment providing tool module 104, into the device 100 where the process optimization module is located as an example to explain the process optimization solution provided by the embodiment of the present application. The process optimization method provided by the embodiment of the present application is applied to the device 100 where the process optimization module is located. When executing the process optimization method provided by the embodiment of the present application, the device 100 where the process optimization module is located can perform the following steps:

[0107] Step 301: The device 100 where the process optimization module is located obtains the value of at least one service parameter of each sub-IVR process in the interactive voice response (IVR) process within the target period, and determines the target sub-IVR process from each sub-IVR process based on the value of each service parameter and the reference value corresponding to each service parameter.

[0108] The report statistics tool module 102 in the device 100 where the process optimization module is located can count and analyze the users' access to the IVR process within the target period, and can count the values ​​of several (at least one) business parameters such as the IVR diversion rate, IVR first-resolution rate, and IVR process repeat dialing rate of each sub-IVR process included in the IVR process within the target period based on the users' access to the IVR process. Among them, the present application does not make specific limitations on the sub-IVR processes included in the IVR process, and different sub-IVR processes can handle different services. The present application does not make specific limitations on the specific process of the report statistics tool module 102 determining the values ​​of the business parameters of each sub-IVR process, nor does it make specific limitations on the specific duration of the target period, and it can be flexibly set according to needs.

[0109] In order to enable the device 100 where the process optimization module is located to identify the target sub-IVR process that needs to be optimized, the operation and management personnel of the IVR process, etc. can pre-configure the IVR process optimization target. The IVR process optimization target can include the values ​​that each of the several business parameters wants to achieve (for the convenience of description, referred to as reference values). In addition, the IVR process optimization target can also include the weight corresponding to each business parameter (for the convenience of description, referred to as preset weight). Among them, this application does not specifically limit the number of business parameters included in the IVR process optimization target, the reference values ​​of the business parameters, the preset weights, etc., and the sum of the preset weights of the business parameters included in the IVR process optimization target can be a value such as 1.

[0110] The device 100 where the process optimization module resides can obtain reference values ​​and preset weights corresponding to each business parameter configured by the operations administrator. The operations administrator in the embodiments of the present application may be a person without experience in process optimization. They only need to configure the relevant IVR process optimization targets according to business needs. The device 100 where the process optimization module resides can then automatically and intelligently optimize the IVR process based on the process optimization method provided in the present application, thereby lowering the threshold for using the process optimization method, reducing process optimization costs, and improving optimization efficiency.

[0111] The device 100 where the process optimization module is located can select (determine) a target sub-IVR process that needs to be optimized from each sub-IVR process based on the values ​​of the business parameters of each sub-IVR process and the reference values ​​corresponding to the business parameters. When determining the target sub-IVR process, it can be that for each sub-IVR process, the difference between the value of the business parameter of the sub-IVR process and the reference value of the corresponding business parameter is counted (calculated), and then based on the difference, the target sub-IVR process is selected from each sub-IVR process. Exemplarily, each sub-IVR process whose business parameter value does not reach the reference value of the corresponding business parameter can be determined as a target sub-IVR process in turn. For example, for each sub-IVR process to be optimized whose service parameter value does not reach the corresponding reference value, the sub-IVR processes to be optimized can be sorted in descending order according to the difference between the service parameter value of each sub-IVR process to be optimized and the corresponding reference value. In order from front to back, each sub-IVR process to be optimized can be determined as a target sub-IVR process, and each sub-IVR process to be optimized can be optimized separately.

[0112] In addition, when selecting a target sub-IVR process from each sub-IVR process, the target sub-IVR process can also be selected based on the difference between the values ​​of the above-mentioned business parameters and the reference values ​​of the business parameters, as well as the preset weights of the business parameters. For example, for each sub-IVR process, the difference between the values ​​of the business parameters of the sub-IVR process and the corresponding reference values ​​can be counted (calculated), and the weighted value between the preset weights of the business parameters of the sub-IVR process and the difference between the above-mentioned corresponding business parameters can be counted (for convenience of description, referred to as multi-dimensional weighted values). Then, each sub-IVR process whose multi-dimensional weighted value does not reach the weighted value between the reference value of the business parameter and the preset weight (for convenience of description, referred to as the reference weighted value) is selected as the target sub-IVR process to be optimized. For example, for each sub-IVR process to be optimized whose multi-dimensional weighted value does not reach the reference weighted value, the sub-IVR processes to be optimized can be sorted in descending order according to the difference between the multi-dimensional weighted value and the reference weighted value, and each sub-IVR process to be optimized can be determined as a target sub-IVR process in order from front to back, and each sub-IVR process to be optimized can be optimized. In addition, after sorting the sub-IVR processes to be optimized in descending order according to the difference between the multi-dimensional weighted value and the reference weighted value, the first set number of sub-IVR processes in the sorting can be determined as target sub-IVR processes in turn. In addition, each sub-IVR process whose difference (deviation) between the multi-dimensional weighted value and the reference weighted value is greater than a set deviation threshold can be determined as a target sub-IVR process, and this application does not make specific restrictions on this.

[0113] Among them, for the convenience of understanding, the following takes the business parameters including business parameter a and business parameter b, where the preset weight of business parameter a is 50% and the preset weight of business parameter b is 50% as an example to explain the process of determining the multidimensional weighted value and the reference weighted value. The reference weighted value can be: the reference value of business parameter a × 50% + the reference value of business parameter b × 50%. The process of determining the multidimensional weighted value is similar to the process of the reference weighted value. The multidimensional weighted value can be: (the value of business parameter a - the reference value of business parameter a) × 50% + (the value of business parameter b - the reference value of business parameter b) × 50%.

[0114] Since the present application can combine the preset weights of the business parameters and the difference between the values ​​of the business parameters and the reference values, it can comprehensively measure the target sub-IVR process that needs to be optimized from multiple perspectives, and can flexibly and accurately determine the target sub-IVR process to be optimized.

[0115] Step 302: The apparatus 100 where the process optimization module is located determines a target node associated with the service parameter in the target sub-IVR process from a plurality of nodes included in the target sub-IVR process.

[0116] Considering that the sub-IVR process contains multiple nodes required for handling corresponding services, such as nodes for entering ID numbers and passwords, when optimizing the target sub-IVR process, the order of the nodes contained in the sub-IVR process (key press sequence) can be adjusted, the content played by the nodes can be adjusted, or the nodes can be deleted, etc., to achieve optimization of the sub-IVR process. In order to quickly and accurately optimize the target sub-IVR process, the nodes associated with the service parameters in the target sub-IVR process can be first determined, and these nodes can be used as target nodes to be optimized. By optimizing the target nodes, the target sub-IVR process can be optimized.

[0117] When determining the target node, the device 100 where the process optimization module is located can quickly and accurately determine the target node associated with the business parameters in the target sub-IVR process based on the association relationship between the business parameters of the sub-IVR process obtained in advance and the index parameters of the node. For example, when the "node repeat visits" of a node in the target sub-IVR process is too high, it can be considered that the playback content of the node (node ​​prompt voice content) and the like may be set unreasonably, resulting in the customer having to enter the same node multiple times and re-listen to the same node multiple times, which may affect the business parameter of the target sub-IVR process, the first solution rate. Therefore, the association relationship between the business parameters of the sub-IVR process obtained and the index parameters of the node can include information that there is an association between the business parameter of the first solution rate of the target sub-IVR process and the index parameter of the node repeat visits of the node. The process optimization model can determine the node as the target node based on the association relationship.

[0118] Since target nodes are nodes associated with service parameters, optimizing these target nodes can maximize the improvement of the service parameter values ​​of the optimized target sub-IVR process, thereby optimizing the target sub-IVR process. Furthermore, since the relationship between the pre-obtained service parameters of the sub-IVR process and the node's indicator parameters can be used to quickly and accurately determine the target nodes associated with the service parameters, the speed and accuracy of IVR process optimization can be improved.

[0119] The association relationship between the business parameters of the above-mentioned sub-IVR process and the index parameters of the node can be obtained based on the trained association relationship model. For example, the business parameters of any sub-IVR process can be input into the trained association relationship model, and the association relationship model can operate on the business parameters of the input sub-IVR process to determine the index parameters of the node associated with the business parameters of the input sub-IVR process. Among them, the training process of the association relationship model can be executed by the device where the model training module (for the convenience of description, referred to as the first model training module) is located. The device where the first model training module is located and the device 100 where the process optimization module is located can be the same device or different devices, and this application does not make specific restrictions on this. The training process of the association relationship model can be as follows:

[0120] The device where the first model training module is located obtains the sample business parameters of any sub-IVR process in the sample set, and the sample business parameters correspond to the sample labels of the correlation coefficients between the sample business parameters of the sub-IVR process and the indicator parameters of each node in the sub-IVR process. The correlation coefficient sample labels can be manually marked by the operation and management personnel, or they can be calculated based on the values ​​of the business parameters of the sub-IVR process in several statistical periods and the values ​​of the indicator parameters of the reception in the sub-IVR process. This application does not specifically limit the specific process of determining the correlation coefficient sample labels.

[0121] The device where the first model training module is located can input the sample business parameters of the acquired sub-IVR process into the association model to be trained. The association model operates on the input sample business parameters of the sub-IVR process, determines the index parameters of the nodes associated with the sample business parameters of the sub-IVR process, and outputs an identification label of the association coefficient between the sample business parameters of the sub-IVR process and the index parameters of each node in the sub-IVR process. The device where the first model training module is located can train the association model to be trained based on the association coefficient identification label output by the association model and the association coefficient sample label of the corresponding sample business parameter.

[0122] For example, the device where the first model training module is located can determine whether the recognition result of the association relationship model is accurate based on whether the association relationship coefficient identification label is consistent with the association relationship coefficient sample label. In a specific implementation, if the association relationship coefficient identification label is inconsistent with the association relationship coefficient sample label, it can be considered that the recognition result of the association relationship model is inaccurate, and the parameters of the association relationship model can be adjusted to train the association relationship model. When adjusting the parameters of the association relationship model, a gradient descent algorithm can be used to perform back propagation on the gradient of the parameters of the association relationship model, etc., to adjust the parameters of the association relationship model. The above operations can be performed on the sample business parameters of each sub-IVR process in the sample set. When the preset convergence conditions are met, it is determined that the training of the association relationship model is completed.

[0123] The preset convergence condition may be that the sample business parameters of the sub-IVR processes in the sample set are accurately identified by the association model, the number of sample business parameters is greater than a set number, or the number of iterations of training the association model reaches a set maximum number of iterations. These conditions can be flexibly set in specific implementations and are not specifically limited here.

[0124] In the embodiment of the present application, the association relationship between the service parameters of the sub-IVR process and the indicator parameters of the node can be quickly and accurately obtained based on the trained association relationship model.

[0125] When determining a target node based on the pre-obtained association relationship between the business parameters of the sub-IVR process and the index parameters of the node, the device 100 where the process optimization module is located can first obtain, based on the association relationship, an association coefficient that can characterize the degree of association between the index parameters of each node included in the target sub-IVR process and the business parameters, and then determine the target node based on the association coefficient.

[0126] The obtained association between the target sub-IVR process's service parameters and the node's index parameters may include an association coefficient that characterizes the degree of correlation between the index parameters of each node in the target sub-IVR process and the service parameters. For example, the association coefficient characterizing the degree of correlation between the index parameters of each node in the target sub-IVR process and the service parameters may be a positive correlation coefficient, a negative correlation coefficient, or an irrelevant correlation coefficient. The irrelevant correlation coefficient may have a value of 0. When the correlation coefficient between a node's index parameter and a service parameter is 0, it is assumed that the node does not affect the value of the target sub-IVR process's service parameters. The positive correlation coefficient may have a value greater than 0 and less than 1. When the correlation coefficient between a node's index parameter and a service parameter is a positive correlation coefficient, it is assumed that as the value of the node's index parameter increases, the value of the service parameter of the target sub-IVR process also increases. The larger the positive correlation coefficient, the greater the magnitude of the increase in the value of the corresponding service parameter. On the contrary, the value of the negative correlation coefficient can be any negative number that is not less than -1 and less than 0, wherein, when the correlation coefficient between the index parameter of a certain node and a certain business parameter is a negative correlation coefficient, it can be considered that when the absolute value of the value of the index parameter of the node increases, the value of the business parameter of the target sub-IVR process decreases accordingly, and the larger the absolute value of the negative correlation coefficient, the greater the amplitude of the decrease in the value of the corresponding business parameter. In other words, it can generally be that the larger the absolute value of the correlation coefficient, the greater the correlation between the index parameter of the corresponding node and the business parameter of the sub-IVR process. Of course, it can also be that the smaller the absolute value of the correlation coefficient, the greater the correlation between the index parameter of the corresponding node and the business parameter of the sub-IVR process, and this application does not make specific restrictions on this. For ease of understanding, the embodiment of the present application takes the example that when the absolute value of the correlation coefficient is larger, the correlation between the index parameter of the corresponding node and the business parameter of the sub-IVR process is greater as an example.

[0127] The above correlation coefficient can be used to quickly and accurately determine the target node to be optimized that is associated with the business parameter. For example, a node whose absolute value of the correlation coefficient between its indicator parameter and the business parameter is greater than a set correlation coefficient threshold can be determined as the target node to be optimized that is associated with the business parameter.

[0128] Since the present application can determine the target node based on the correlation coefficient that can characterize the correlation between the indicator parameter of the node and the business parameter, the speed and accuracy of determining the target node can be improved.

[0129] When determining the target node based on the above-mentioned correlation coefficient, the target node can be comprehensively determined from multiple dimensions in combination with the values ​​of the indicator parameters. Specifically, the ranking that can characterize the degree of superiority or inferiority between the nodes can be obtained based on the correlation coefficient corresponding to each indicator parameter and the values ​​of the indicator parameters of each node. For the convenience of description, for example, the correlation coefficient that characterizes the correlation between a certain indicator parameter (such as indicator parameter A) of each node and the business parameter can be called the correlation coefficient corresponding to the indicator parameter A. Similarly, for the convenience of description, the indicator parameter A can be called the indicator parameter corresponding to the correlation coefficient.

[0130] For example, when the value of the index parameter of a certain node is less ideal and the correlation coefficient corresponding to the index parameter is larger, it can be considered that the configuration of the node is more unreasonable (worse) and the node needs to be optimized more, and the node can be determined as the target node to be optimized first. When the value of the index parameter of a certain node is more ideal, it can be considered that the configuration of the node is more reasonable (better) and the node does not need to be optimized. The ranking of the degree of excellence (to be optimized) between the nodes can be comprehensively determined by combining the ideal value of the index parameter of each node and the size of the correlation coefficient corresponding to the index parameter with an unideal value. Among them, the present application does not make a specific limitation on the degree of excellence represented by the ranking. For example, the higher the ranking is, the worse the configuration of the corresponding node is and the more it needs to be optimized. The set number of nodes with a higher ranking can be determined as the target node. In addition, the lower the ranking is, the worse the configuration of the corresponding node is and the more it needs to be optimized. The set number of nodes with a lower ranking can be determined as the target node.

[0131] Since the target node can be selected based on a comprehensive evaluation of multiple dimensions including the correlation coefficient and the node's index parameters, the accuracy of determining the target node can be improved to the greatest extent possible. By adjusting the target node selected based on this method, the values ​​of the business parameters of the optimized target sub-IVR process can be improved to the greatest extent, thereby achieving the purpose of optimizing the IVR process quickly and accurately.

[0132] In order to quickly and accurately obtain a ranking that can characterize the degree of superiority or inferiority between nodes, each node can be sorted for each indicator parameter based on the value of the indicator parameter taken by each node included in the target sub-IVR process, thereby obtaining a sorting sequence for each node in the indicator parameter dimension (for the convenience of description, the sorting sequence can be referred to as a sub-sorting sequence), and then the ranking position of each node in the sub-sorting sequence can be obtained (for the convenience of description, the ranking position can be referred to as a sub-ranking position). Alternatively, each indicator parameter can be sorted based on the product (weighted value) of the value of each node in the indicator parameter and the correlation coefficient corresponding to the indicator parameter, thereby obtaining a sub-ranking position of each node in the indicator parameter dimension.

[0133] Wherein, the embodiment of the present application does not limit the way in which each node is sorted in the sub-sorting sequence corresponding to each index parameter. For example, when the value of the index parameter of a node is higher, it represents that the value of the index parameter is less ideal, the current configuration of the node is more unreasonable, and the node needs to be optimized more, then the order of the value of the index parameter from large to small can be used to sort each node in descending order, and a sub-sorting sequence can be obtained, and the node with the highest ranking in the sub-sorting sequence can be preferentially determined as the target node to be optimized. For another example, when the value of the index parameter of a node is lower, it represents that the value of the index parameter is less ideal, the current configuration of the node is more unreasonable, and the node needs to be optimized more, then the order of the value of the index parameter from small to large can be used to sort each node in ascending order, and a sub-sorting sequence can be obtained, and the node with the highest ranking in the sub-sorting sequence can be preferentially determined as the target node to be optimized. It is understandable that no matter which method is used, it can be preferred to determine the node with the less ideal value of the index parameter as the target node.

[0134] After obtaining the sub-ranking rank of each node in each indicator parameter dimension, the target node can be comprehensively determined by combining the sub-ranking rank of each node in each indicator parameter dimension and the correlation coefficient corresponding to each indicator parameter. For example, for each node, the weighted value between the sub-ranking rank of the node in each indicator parameter dimension and the corresponding correlation coefficient can be obtained. According to the weighted value, each node can be comprehensively sorted to obtain a comprehensive sorting sequence. The ranking rank of each node in the comprehensive sorting sequence can be considered as a ranking rank that can characterize (reflect) the degree of superiority or inferiority between each node. According to the order of each node in the comprehensive sorting sequence, a set number of target nodes to be optimized can be determined, so as to prioritize the nodes with a high correlation with the business parameters of the target sub-IVR process and a less ideal (inferior) value of the indicator parameters as target nodes. This application does not specifically limit the number of target nodes to be determined (set number), for example, it can be 10 or 8, etc., and can be flexibly set according to demand.

[0135] For ease of understanding, the process of determining the target node provided in the embodiment of the present application is explained below through a specific embodiment:

[0136] Assume that the target sub-IVR process includes 11 nodes, namely: node1, node2, node3, node4, node5, node6, node7, node8, node9, node10, node11. The target nodes are determined from the dimensions of three indicator parameters respectively. The three indicator parameters are respectively represented as: indicator parameter 1, indicator parameter 2, and indicator parameter 3, wherein the correlation coefficient corresponding to indicator parameter 1 is called correlation coefficient 1 (also called weight 1), the correlation coefficient corresponding to indicator parameter 2 is called correlation coefficient 2 (also called weight 2), and the correlation coefficient corresponding to indicator parameter 3 is called correlation coefficient 3 (also called weight 3). For the values ​​of the indicator parameters of each node and the values ​​of the correlation coefficients corresponding to the indicator parameters, please refer to Table 1. The values ​​of the correlation coefficients corresponding to the indicator parameters in Table 1 are only examples. The correlation coefficients corresponding to the indicator parameters can be the same or different. The sum of the correlation coefficients corresponding to the indicator parameters can be a value such as 1. This application does not make specific restrictions on this.

[0137] Table 1

[0138] Each node can be sorted according to its values ​​in indicator parameter 1, indicator parameter 2, and indicator parameter 3, and the sub-sorted sequence of each node in each indicator parameter dimension can be obtained. Assuming that the nodes are sorted in descending order based on the values ​​of indicator parameter 1 of each node from large to small, the sub-sorted position of each node in the sub-sorted sequence of indicator parameter 1 dimension can be obtained. Assuming that in the sub-sorted sequence of indicator parameter 1 dimension, the sub-sorted positions of node1, node2, and node4 are all 1 (node1, node2, and node4 are tied for first), the sub-sorted positions of node3, node5, node6, node7, and node8 are all 2, and the sub-sorted positions of node9, node10, and node11 are all 3.

[0139] Assume that the nodes are sorted in descending order based on the value of the indicator parameter 2 of each node, and the sub-sorting position of each node in the sub-sorting sequence of the indicator parameter 2 dimension is obtained. Assume that in the sub-sorting sequence of the indicator parameter 2 dimension, the sub-sorting positions of node1, node2 and node8 are all 1, the sub-sorting positions of node3, node4, node5, node6 and node7 are all 2, and the sub-sorting positions of node9, node10 and node11 are all 3.

[0140] Assume that the nodes are sorted in ascending order based on the value of the indicator parameter 3 of each node, and the sub-sorting position of each node in the sub-sorting sequence of the indicator parameter 3 dimension is obtained. Assume that in the sub-sorting sequence of the indicator parameter 3 dimension, the sub-sorting position of node9, node10 and node11 are all 1, the sub-sorting position of node4, node5, node7 and node8 are all 2, the sub-sorting position of node1, node2 and node3 are all 3, and the sub-sorting position of node6 is 4.

[0141] After obtaining the sub-ranking position of each node in each indicator parameter dimension, the weighted value between the sub-ranking position of the node in each indicator parameter dimension and the corresponding correlation coefficient of each indicator parameter can be obtained for each node. For example, for node node1, node1's sub-ranking position in indicator parameter 1 dimension is 1, and the corresponding correlation coefficient of indicator parameter 1 is 0.5; node1's sub-ranking position in indicator parameter 2 dimension is 1, and the corresponding correlation coefficient of indicator parameter 2 is 0.3; node1's sub-ranking position in indicator parameter 3 dimension is 3, and the corresponding correlation coefficient of indicator parameter 3 is 0.2. Then, the weighted value of node1's sub-ranking position in each indicator parameter dimension and the corresponding correlation coefficient = 1 × 0.5 + 1 × 0.3 + 3 × 0.2 = 1.4.

[0142] The above weighted value of each node can be calculated based on the same calculation method. For example, the weighted value of Node2's sub-ranking position in each indicator parameter dimension and the corresponding correlation coefficient = 1×0.5+1×0.3+3×0.2=1.4.

[0143] The weighted value of Node3's sub-ranking position in each indicator parameter dimension and the corresponding correlation coefficient = 2×0.5+2×0.3+3×0.2=2.2.

[0144] The weighted value of the sub-ranking position of Node4 in each indicator parameter dimension and the corresponding correlation coefficient = 1×0.5+2×0.3+2×0.2=1.5.

[0145] The weighted value of Node5's sub-ranking position in each indicator parameter dimension and the corresponding correlation coefficient = 2×0.5+2×0.3+2×0.2=2.

[0146] The weighted value of Node6's sub-ranking position in each indicator parameter dimension and the corresponding correlation coefficient = 2×0.5+2×0.3+4×0.2=2.4.

[0147] The weighted value of Node7's sub-ranking position in each indicator parameter dimension and the corresponding correlation coefficient = 2×0.5+2×0.3+4×0.2=2.4.

[0148] The weighted value of Node8's sub-ranking position in each indicator parameter dimension and the corresponding correlation coefficient = 2×0.5+1×0.3+2×0.2=1.7.

[0149] The weighted value of Node9's sub-ranking position in each indicator parameter dimension and the corresponding correlation coefficient = 3×0.5+3×0.3+1×0.2=2.6.

[0150] The weighted value of the sub-ranking position of Node10 in each indicator parameter dimension and the corresponding correlation coefficient = 3×0.5+3×0.3+1×0.2=2.6.

[0151] The weighted value of Node11's sub-ranking position in each indicator parameter dimension and the corresponding correlation coefficient = 3×0.5+3×0.3+1×0.2=2.6.

[0152] Assuming that the smaller the weighted value of the node obtained based on the above calculation method, the higher the correlation between the node and the business parameters of the target sub-IVR process, and the less ideal (worse) the value of the indicator parameter is. The nodes can be comprehensively sorted in order of weighted values ​​from small to large to obtain a comprehensive sorting sequence. Based on the ranking of each node in the comprehensive sorting sequence, a ranking that can characterize the degree of superiority and inferiority between each node is obtained. According to the order of the nodes in the comprehensive sorting sequence from front to back, a set number of nodes with the highest ranking can be determined as target nodes. For example, assuming that a total of 5 target nodes are selected, node1, node2, node4, node8 and node5 can be determined as target nodes respectively.

[0153] Of course, the nodes can also be comprehensively sorted in descending order of weighted values ​​to obtain a comprehensive sorting sequence, and a set number of nodes with a lower sorting order can be determined as target nodes according to the order of the nodes in the comprehensive sorting sequence from back to front. This application does not make any specific restrictions on this.

[0154] Since the present application can accurately obtain the ranking that can characterize the degree of superiority and inferiority between each node based on each correlation coefficient and the sub-ranking position of each node in each indicator parameter, the accuracy of determining the target node to be optimized can be improved. By adjusting the target node selected based on this method, the value of the business parameter of the optimized target sub-IVR process can be improved to the greatest extent, thereby realizing quick and accurate optimization of the IVR process.

[0155] Step 303: The device 100 where the process optimization module is located selects a target optimization solution from the optimization solution set, and the target optimization solution is used to optimize the target node.

[0156] Operation management personnel and the like can pre-configure an optimization solution set, which can include several optimization solutions that can optimize nodes of different sub-IVR processes. The device 100 where the process optimization module is located can obtain the optimization solution set, and can select a target optimization solution from the optimization solution set that can optimize the target node in the target sub-IVR process. For example, each optimization solution in the optimization solution set can correspond to identification information of the sub-IVR process and node to which the optimization solution is applicable. Based on this identification information, the device 100 where the process optimization module is located can select an optimization solution from the optimization solution set that is applicable to the target node in the target sub-IVR process, and determine the optimization solution as the target optimization solution.

[0157] In order to quickly and accurately determine the target optimization solution, the correspondence between the index parameters of the nodes of the sub-IVR process and the optimization solution can be obtained in advance. When the device 100 where the process optimization module is located selects the target optimization solution from the optimization solution set, the values ​​of the index parameters of the target node of the target sub-IVR process can be compared and matched with the values ​​of the index parameters in the above-mentioned correspondence (the correspondence between the index parameters of the nodes of the sub-IVR process and the optimization solution). The optimization solution contained in the successfully matched correspondence is used as the optimization solution corresponding to the value of the index parameter of the target node of the target sub-IVR process in the optimization solution set, and this optimization solution can be determined as the target optimization solution.

[0158] Since the target optimization scheme determined in this application is an optimization scheme in the optimization scheme set that corresponds to the value of the indicator parameter of the target node of the target sub-IVR process, the target optimization scheme can be better applied to the target node. When the target node in the target sub-IVR process is optimized based on the target optimization scheme, the purpose of optimizing the target sub-IVR process can be achieved to the greatest extent possible.

[0159] The correspondence between the index parameters of the nodes of the above-mentioned sub-IVR process and the optimization scheme can be obtained based on the trained optimization scheme library model. For example, the values ​​of the index parameters of the nodes of any sub-IVR process can be input into the trained optimization scheme library model, and the optimization scheme library model can calculate the index parameters of the nodes of the input sub-IVR process, thereby determining the optimization scheme corresponding to the values ​​of the index parameters of the nodes of the input sub-IVR process. Among them, the training process of the optimization scheme library model can be executed by the device where the model training module (for the convenience of description, referred to as the second model training module) is located. The device where the second model training module is located and the device 100 where the process optimization module is located can be the same device or different devices. In addition, the device where the second model training module is located and the device where the first model training module is located can be the same device or different devices. This application does not make specific restrictions on this. The training process of the optimization scheme library model can be as follows:

[0160] The device containing the second model training module obtains the values ​​of sample indicator parameters for several nodes of any sub-IVR process in the sample set. The sample indicator parameters correspond to optimization solution sample labels applicable to the values ​​of the sample indicator parameters of these several nodes of the sub-IVR process. The optimization solution sample labels can be obtained based on manual labeling by operations management personnel, etc., and this application does not specifically limit the specific process of optimizing solution sample labels.

[0161] The device where the second model training module resides can input the obtained sample index parameter values ​​of several nodes of the sub-IVR process into the optimization solution library model to be trained. The optimization solution library model operates on the input sample index parameter values ​​of several nodes of the sub-IVR process, determines an optimization solution applicable to the sample index parameter values ​​of several nodes of the sub-IVR process, and outputs an optimization solution identification label. The device where the second model training module resides can train the optimization solution library model to be trained based on the optimization solution identification label output by the optimization solution library model and the optimization solution sample label corresponding to the sample index parameter.

[0162] For example, the device where the second model training module is located can determine whether the recognition result of the optimization solution library model is accurate based on whether the optimization solution identification label is consistent with the optimization solution sample label. In a specific implementation, if the optimization solution identification label is inconsistent with the optimization solution sample label, it can be considered that the recognition result of the optimization solution library model is inaccurate, and the parameters of the optimization solution library model can be adjusted to train the optimization solution library model. Optionally, when adjusting the parameters of the optimization solution library model, a gradient descent algorithm can be used to backpropagate the gradient of the parameters of the optimization solution library model, etc., to adjust the parameters of the optimization solution library model. The above operation can be performed on the sample indicator parameters of each node of the sub-IVR process in the sample set, and when the preset convergence conditions are met, it is determined that the training of the optimization solution library model is completed.

[0163] The preset convergence condition can be satisfied by, for example, the number of applicable optimization solutions accurately identified by the optimization solution library model for the sample index parameters of the nodes of the sub-IVR process in the sample set is greater than a set number, or the number of iterations of training the optimization solution library model reaches a set maximum number of iterations. These settings can be flexibly made in specific implementations and are not specifically limited here.

[0164] The embodiment of the present application can quickly and accurately obtain the correspondence between the index parameters of the nodes of the sub-IVR process and the optimization scheme based on the trained optimization scheme library model. Based on this correspondence, the accuracy of determining the target optimization scheme can be improved, thereby improving the speed and accuracy of IVR process optimization.

[0165] The device 100 where the process optimization module is located can obtain one or more optimization schemes corresponding to the values ​​of the index parameters of the target node of the target sub-IVR process based on the optimization scheme library model. These optimization schemes can be used as candidate schemes respectively. Each candidate scheme can carry priority information. The target optimization scheme can be determined (selected) from the candidate schemes based on the priority information. For example, the priority information carried by any candidate scheme can include the priority coefficient or priority level of the candidate scheme applicable to the target sub-IVR process. The present application does not make specific restrictions on the specific priority coefficients and priority levels. The higher the priority coefficient or priority level, the more applicable the candidate scheme is to the target sub-IVR process. Of course, the lower the priority coefficient or priority level, the more applicable the candidate scheme is to the target sub-IVR process. The device 100 where the process optimization module is located can select a scheme that is most suitable for the target sub-IVR process from the candidate schemes based on the priority information carried by the candidate schemes as the target optimization scheme.

[0166] In order to increase the flexibility and accuracy of determining the target optimization plan, in the early stage of the operation of the optimization plan library model, the device 100 where the process optimization module is located can determine one or more optimization plans corresponding to the values ​​of the indicator parameters of the target node of the target sub-IVR process obtained based on the optimization plan library model as candidate plans, and can display each candidate plan. The operation manager (for the convenience of description, it can also be called the user) selects a plan from the candidate plans as the target optimization plan. The operation manager can also modify a candidate plan. The device 100 where the process optimization module is located can identify the candidate plan selected by the operation manager (user) or the candidate plan modified by the user, and can determine the candidate plan selected by the user or the candidate plan modified by the user as the target optimization plan.

[0167] Since the embodiment of the present application can determine the target optimization solution based on the candidate solution selected by the user or the candidate solution modified by the user, the flexibility and accuracy of determining the target optimization solution can be improved.

[0168] Step 304: The device 100 where the process optimization module is located optimizes the target sub-IVR process based on the target optimization solution.

[0169] After determining the target optimization scheme, the device 100 where the process optimization module is located can optimize the target sub-IVR process based on the target optimization scheme. For example, the process optimization module 101 in the device 100 where the process optimization module is located can generate optimization operation instructions for the target nodes based on the adjustment information for the nodes in the target sub-IVR process included in the target optimization scheme, and drive (call) the IVR orchestration tool module 103 in the device 100 where the process optimization module is located through interface calls, etc. The IVR orchestration tool module 103 adjusts (optimizes) the key sequence, playback content, etc. of the nodes included in the target sub-IVR process based on the IVR orchestration tool and the optimization operation instructions, thereby achieving automatic and intelligent optimization of the target sub-IVR process. Compared with the related technologies that still require IVR process customization developers or business managers who are proficient in IVR process orchestration technology, etc., manually choreographing and optimizing the IVR process according to the business processing logic requirements in the optimization plan requires business managers to have certain work experience. In the embodiments of the present application, the target sub-IVR process can be automatically and intelligently optimized, which does not rely on manual labor or manual work experience, can reduce labor costs and improve efficiency.

[0170] Since, in the embodiment of the present application, the device 100 where the process optimization module is located can intelligently determine the target sub-IVR process to be optimized from each sub-IVR process based on the values ​​of the business parameters of each sub-IVR process and the reference values ​​of the business parameters, when optimizing the target sub-IVR process to be optimized, the device 100 where the process optimization module is located can first determine the target node associated with the business parameters from the multiple nodes included in the target sub-IVR process, and select a target optimization scheme that can optimize the target node from the optimization scheme set, and optimize the target sub-IVR process based on the target optimization scheme, thereby achieving the purpose of quickly and easily optimizing the IVR process.

[0171] After optimizing the target sub-IVR process based on the target optimization solution, a comparative test can be performed on the optimized target sub-IVR process and the target sub-IVR process before optimization. For example, the process optimization module 101 can drive the IVR operating environment providing tool module 104. The IVR operating environment providing tool module 104 loads and runs the optimized target sub-IVR process based on the IVR process operating environment providing tool, and can record the customer's access log to the optimized target sub-IVR process. The report statistics tool module 102 can count and analyze the access log of the customer accessing the optimized target sub-IVR process, and present the values ​​of the business parameters of the optimized target sub-IVR process. The process optimization module 101 can obtain the values ​​of the business parameters of the target sub-IVR process, and can compare the values ​​of the business parameters of the optimized target sub-IVR process with the values ​​of the business parameters of the target sub-IVR process before optimization after running for the same length of time to obtain the test results of the comparative test. When the comparison test results show that the business parameter values ​​of the optimized target sub-IVR process are better than the business parameter values ​​of the target sub-IVR process before optimization, it can be considered that the optimized target sub-IVR process has achieved a certain optimization effect and the optimized target sub-IVR process can be adopted.

[0172] In addition, if the values ​​of the business parameters of the optimized target sub-IVR process are not only better than the values ​​of the business parameters of the target sub-IVR process before optimization, but also reach the reference values ​​corresponding to the business parameters, it can be considered that the target sub-IVR process has achieved the optimization goal.

[0173] If the service parameter values ​​of the optimized target sub-IVR process are merely better than those of the target sub-IVR process before optimization, but have not yet reached the corresponding reference values ​​for the service parameters, it can be considered that the target sub-IVR process has only achieved a certain optimization goal and requires further optimization to achieve the optimization goal. The optimized target sub-IVR process can be determined as the target sub-IVR process to be optimized in step 301. The optimized target sub-IVR process is further optimized based on the steps in steps 302-304 until the service parameter values ​​of the optimized target sub-IVR process reach the corresponding reference values, i.e., the optimization goal is achieved. This will not be further described here.

[0174] In order to continuously accumulate experience and continuously optimize the correspondence between the index parameters of the nodes of the sub-IVR process obtained based on the optimization solution library model and the optimization solution, the device 100 where the process optimization module is located can optimize and train the optimization solution library model based on the test results of the comparison test, so that the correspondence between the index parameters of the nodes of the sub-IVR process obtained based on the optimization solution library model and the optimization solution can be optimized and updated. Exemplarily, the priority information carried by the target optimization solution corresponding to the value of the index parameter of the target node of the target sub-IVR process obtained based on the optimization solution library model can be updated. For example, if the value of the business parameter of the target sub-IVR process after optimization is not only better than the value of the business parameter of the target sub-IVR process before optimization, but also reaches the reference value corresponding to the business parameter, the target optimization solution can be determined as the optimization solution most suitable for the target sub-IVR process, and the priority coefficient or priority level of the target optimization solution can be configured to be the highest priority coefficient or the highest priority level most suitable for the target sub-IVR process.

[0175] If the values ​​of the business parameters of the target sub-IVR process after optimization are only better than the values ​​of the business parameters of the target sub-IVR process before optimization, but have not yet reached the reference values ​​corresponding to the business parameters, the priority coefficient or priority level of the target optimization plan can be configured as a second-highest priority coefficient or second-highest priority level that is more suitable for but not the most suitable for the target sub-IVR process, etc., and can be flexibly set according to needs.

[0176] Among them, when the higher the value of a certain business parameter is, the more reasonable the corresponding IVR process setting is. When the value of the business parameter of the optimized target sub-IVR process is higher than the value of the business parameter of the target sub-IVR process before optimization, it is considered that the value of the business parameter of the optimized target sub-IVR process is better than the value of the business parameter of the target sub-IVR process before optimization. When the value of the business parameter of the optimized target sub-IVR process is lower than the value of the business parameter of the target sub-IVR process before optimization, it is considered that the value of the business parameter of the optimized target sub-IVR process is worse than the value of the business parameter of the target sub-IVR process before optimization.

[0177] When the lower the value of a certain business parameter is, the more reasonable the corresponding IVR process setting is, it can be that when the value of the business parameter of the target sub-IVR process after optimization is lower than the value of the business parameter of the target sub-IVR process before optimization, it is considered that the value of the business parameter of the target sub-IVR process after optimization is better than the value of the business parameter of the target sub-IVR process before optimization, and when the value of the business parameter of the target sub-IVR process after optimization is higher than the value of the business parameter of the target sub-IVR process before optimization, it is considered that the value of the business parameter of the target sub-IVR process after optimization is worse than the value of the business parameter of the target sub-IVR process before optimization.

[0178] In addition, if the comparison test results show that the service parameter values ​​of the optimized target sub-IVR process are inferior to the service parameter values ​​of the target sub-IVR process before optimization, it can be considered that the optimized target sub-IVR process has not optimized the target sub-IVR process, but may have a counter-effect. The device 100 where the process optimization module is located can restore the currently used target sub-IVR process (such as the optimized target sub-IVR process) to the target sub-IVR process before optimization. At the same time, in order to optimize the pre-acquired optimization solution library model and optimize the correspondence between the index parameters of the nodes of the sub-IVR process obtained based on the optimization solution library model and the optimization solution, the priority coefficient or priority level of the target optimization solution can be configured to the lowest priority coefficient or lowest priority level that is not applicable to the target sub-IVR process, or the target optimization solution can be directly removed from the above correspondence and excluded from the candidate solutions. At the same time, after restoring the currently used target sub-IVR process to the target sub-IVR process before optimization, the target sub-IVR process before optimization can be again determined as the target sub-IVR process to be optimized in step 301, and the target sub-IVR process can be further optimized based on the steps in steps 302 to 304 until the values ​​of the business parameters of the target sub-IVR process after optimization reach the reference values ​​corresponding to the business parameters, thereby achieving the optimization goal. This will not be repeated here.

[0179] Similar to the above-mentioned process of optimizing and training the optimization solution library model, in order to continuously optimize the association relationship model, accumulate experience, and improve the optimization effect of the IVR process, the device 100 where the process optimization module is located can obtain the correlation coefficient between the index parameters of each node in the target sub-IVR process and each business parameter based on the values ​​of the business parameters and the values ​​of the index parameters of the nodes corresponding to the optimized target sub-IVR process and the target sub-IVR process before optimization contained in the test results (for convenience of description, this correlation coefficient is referred to as the reference correlation coefficient); based on the reference correlation coefficient, the correlation relationship between the business parameters of the sub-IVR process and the index parameters of the nodes previously obtained can be optimized and updated. For example, assume that the above-mentioned association relationship obtained based on the association relationship model includes: there is an association relationship between business parameter A and nodes C1, C2, and C3 in the target sub-IVR process, and the correlation coefficient between business parameter A and node C1 is 0.5, the correlation coefficient between business parameter A and node C2 is 0.6, and the correlation coefficient between business parameter A and node C3 is -0.7. Based on the values ​​of the business parameters and the index parameters of the nodes corresponding to the optimized target sub-IVR process and the target sub-IVR process before optimization contained in the test results, the reference correlation coefficients obtained include: the correlation coefficient between business parameter A and node C1 is not 0.5 but 0.8, the correlation coefficient between business parameter A and node C2 is not 0.6 but 0.7, the correlation coefficient between business parameter A and node C3 is still -0.7, and the correlation coefficient between business parameter A and node C4 is 0.7. Based on the latest reference correlation coefficients, the correlation model can be optimized and trained so that the correlation model can learn that business parameter A is not only related to There is a correlation between nodes C1, C2, and C3 in the target sub-IVR process, and there is also a correlation with node C4 in the target sub-IVR process. These correlation information can also enable the correlation model to learn that the correlation coefficient between business parameter A and node C1 is 0.8, the correlation coefficient between business parameter A and node C2 is 0.7, the correlation coefficient between business parameter A and node C3 is -0.7, and the correlation coefficient between business parameter A and node C4 is 0.7. These specific correlation coefficients enable the optimized trained correlation model to provide more accurate information on the correlation between the business parameters of the sub-IVR process and the indicator parameters of the nodes.

[0180] For ease of understanding, the process optimization method provided by this application is explained below through a specific embodiment. Referring to Figure 4, it is a process diagram of another process optimization method provided by an embodiment of this application, which includes the following steps:

[0181] Step 401: The operations manager configures an IVR process optimization target in the device 100 where the process optimization module 101 is located.

[0182] For ease of understanding, the following example uses the IVR process optimization goals configured by the operations management personnel as follows: the IVR first-time solution rate is increased to 30% (i.e., the reference value of the IVR first-time solution rate is 30%), and the preset weight of the IVR first-time solution rate is 50%; the IVR diversion rate is increased to 80% (i.e., the reference value of the IVR diversion rate is 80%), and the preset weight of the IVR diversion rate is 50%.

[0183] Step 402: The process optimization module 101 obtains the values ​​of the IVR first-resolution rate and the IVR diversion rate of each sub-IVR process in the IVR process within the target period. For each sub-IVR process, the difference between the value of the IVR first-resolution rate of the sub-IVR process and the reference value of the IVR first-resolution rate of 30% configured by the operation and management personnel (for the convenience of description, referred to as the first difference) and the difference between the value of the IVR diversion rate of the sub-IVR process and the reference value of the IVR diversion rate of 80% configured by the operation and management personnel (for the convenience of description, referred to as the second difference) are calculated, and the multi-dimensional weighted value of the sub-IVR process is calculated: the first difference × 50% + the second difference × 50%.

[0184] The difference between the multidimensional weighted value of each sub-IVR process and the reference weighted value (30%×50%+80%×50%) is counted, and the sub-IVR process with the largest difference between the multidimensional weighted value and the reference weighted value is determined as the target sub-IVR process.

[0185] Step 403: The process optimization module 101 determines a target node associated with the service parameter in the target sub-IVR process from a plurality of nodes included in the target sub-IVR process based on the trained association relationship model.

[0186] Step 404: The process optimization module 101 generates (determines) a target optimization solution that can be used to optimize the target node based on the trained optimization solution library model.

[0187] Among them, in the early stage of running the optimization solution library model and the association relationship model, the operation management personnel can assist in determining the target optimization solution. The following describes a method provided by the embodiment of the present application for the operation management personnel to assist in determining the target optimization solution:

[0188] Step 1: The process optimization module 101 may use the optimization solutions obtained based on the optimization solution library model as candidate solutions and display each candidate solution.

[0189] Step 2: Operations management personnel can manually select or modify candidate solutions.

[0190] Step 3: The process optimization module 101 may determine the candidate solution selected or modified by the user (operation manager) as the target optimization solution.

[0191] The aforementioned step of having operations management personnel assist in determining the target optimization solution may be optional. For example, in the initial stages of the operation of the optimization solution library model and the association model, operations management personnel may assist in determining the target optimization solution, and may thereby continuously optimize the optimization solution library model and the association model. Subsequently, once the optimization solution library model and the association model have achieved a high degree of accuracy, operations management personnel may no longer be required to assist in determining the target optimization solution, and the target optimization solution may be determined directly based on the optimization solution library model.

[0192] Step 405 : The process optimization module 101 generates optimization operation instructions for the target nodes according to the adjustment information for the nodes in the target sub-IVR process included in the target optimization solution, and drives (calls) the IVR orchestration tool module 103 .

[0193] Step 406: The IVR orchestration tool module 103 adjusts (optimizes) the nodes included in the target sub-IVR process based on the configured IVR orchestration tool according to the optimization operation instruction, thereby automatically and intelligently optimizing the target sub-IVR process.

[0194] The nodes adjusted in the target sub-IVR process may include other nodes except the target node, which is not specifically limited in this application.

[0195] Step 407: The process optimization module 101 performs a comparative test on the optimized target sub-IVR process and the target sub-IVR process before optimization.

[0196] The following describes a method for performing a comparative test on the optimized target sub-IVR process and the unoptimized target sub-IVR process, provided in an embodiment of the present application (not shown):

[0197] If the values ​​of the business parameters of the optimized target sub-IVR process in the test results are worse than the values ​​of the business parameters of the target sub-IVR process before optimization, the currently used target sub-IVR process will be restored to the target sub-IVR process before optimization; otherwise, if the values ​​of the business parameters of the optimized target sub-IVR process in the test results are better than the values ​​of the business parameters of the target sub-IVR process before optimization, the optimized target sub-IVR process will be adopted.

[0198] In addition, based on the values ​​of the business parameters corresponding to the optimized target sub-IVR process and the target sub-IVR process before optimization and the values ​​of the index parameters of the nodes contained in the test results, a reference correlation coefficient between the index parameters of each node and each business parameter in the target sub-IVR process can be obtained; based on the reference correlation coefficient, the correlation model is optimized and updated, that is, the correlation between the business parameters of the sub-IVR process and the index parameters of the nodes obtained based on the correlation model can be optimized and updated.

[0199] The optimization solution library model may also be optimized and updated based on the test results, that is, the corresponding relationship between the index parameters of the nodes of the sub-IVR process obtained by the optimization solution library model and the optimization solution may be optimized and updated.

[0200] In addition, experimental verification shows that the process optimization method provided in the embodiment of the present application can increase the IVR first-time resolution rate of the call center's IVR process by about 5% to 10% and the IVR diversion rate by about 10% within 6 months of operation.

[0201] Based on the same inventive concept as the method embodiment, the embodiment of the present application also provides a process optimization device, which is used to execute the method executed by the device 100 where the process optimization module is located in the method embodiment. As shown in Figure 5, the process optimization device 500 includes a process optimization module 101. The process optimization module 101 can obtain the value of at least one business parameter of each sub-IVR process in the IVR process within the target period, and based on the value of each business parameter and the reference value corresponding to each business parameter, determine a target sub-IVR process from each sub-IVR process; determine a target node associated with the business parameter in the target sub-IVR process from multiple nodes included in the target sub-IVR process; select a target optimization solution from the optimization solution set, and the target optimization solution is used to optimize the target node; and optimize the target sub-IVR process based on the target optimization solution.

[0202] As a possible implementation, the process optimization module 101 may determine the target node based on the association between the pre-obtained service parameters of the sub-IVR process and the index parameters of the node.

[0203] As a possible implementation, the process optimization module 101 can obtain each correlation coefficient based on the correlation between the business parameters of the sub-IVR process and the index parameters of the node obtained in advance. Each correlation coefficient represents the correlation between any index parameter and the business parameter of each node included in the target sub-IVR process; the target node can be determined based on each correlation coefficient.

[0204] As a possible implementation, the process optimization module 101 can obtain a ranking position that characterizes the degree of excellence between each node based on each correlation coefficient and the value of the index parameter of each node included in the target sub-IVR system; based on the ranking position between each node, determine a set number of target nodes to be optimized from each node.

[0205] As a possible implementation, the process optimization module 101 can obtain the sub-ranking position of each node in each indicator parameter based on the value of each node in the indicator parameter; and can obtain the ranking position that represents the degree of excellence between each node based on each correlation coefficient and the sub-ranking position of each node.

[0206] As a possible implementation method, the association relationship between the business parameters of the sub-IVR process and the index parameters of the node is obtained based on a trained association model. The association model can determine the index parameters of the node associated with the business parameters of the input sub-IVR process based on the business parameters of the input sub-IVR process.

[0207] As a possible implementation method, the process optimization module 101 can determine the optimization scheme corresponding to the value of the indicator parameter of the target node of the target sub-IVR process in the optimization scheme set as the target optimization scheme based on the correspondence between the indicator parameters of the nodes of the sub-IVR process obtained in advance and the optimization scheme.

[0208] As a possible implementation method, the process optimization module 101 can determine at least one optimization solution in the optimization solution set that corresponds to the value of the indicator parameter of the target node of the target sub-IVR process as a candidate solution, and display each of the candidate solutions; determine the target optimization solution based on the candidate solution selected by the user or the candidate solution modified by the user.

[0209] As a possible implementation method, the correspondence between the index parameters of the nodes of the sub-IVR process and the optimization scheme is obtained based on the trained optimization scheme library model. The optimization scheme library model can determine the optimization scheme corresponding to the value of the index parameter of the node of the input sub-IVR process based on the value of the index parameter of the node of the input sub-IVR process.

[0210] As a possible implementation, the process optimization module 101 can count the difference between the value of each business parameter of each sub-IVR process and the reference value of the corresponding business parameter for each sub-IVR process; and count the weighted value between the preset weight of each business parameter of the sub-IVR process and the corresponding difference; based on the weighted value of each sub-IVR process and the reference weighted value between the reference value of each business parameter and the preset weight, determine the target sub-IVR process from each sub-IVR process.

[0211] As a possible implementation method, the process optimization module 101 can also perform a comparative test on the optimized target sub-IVR process and the target sub-IVR process before optimization. If the values ​​of the business parameters of the optimized target sub-IVR process in the test results are worse than the values ​​of the business parameters of the target sub-IVR process before optimization, the currently used target sub-IVR process is restored to the target sub-IVR process before optimization; otherwise, if the values ​​of the business parameters of the optimized target sub-IVR process in the test results are better than the values ​​of the business parameters of the target sub-IVR process before optimization, the optimized target sub-IVR process is adopted.

[0212] As a possible implementation method, the process optimization module 101 can also obtain a reference correlation coefficient between the index parameters of each node and each business parameter in the target sub-IVR process based on the values ​​of the business parameters corresponding to the optimized target sub-IVR process and the target sub-IVR process before optimization and the values ​​of the index parameters of the nodes contained in the test results; based on the reference correlation coefficient, the correlation between the obtained business parameters of the sub-IVR process and the index parameters of the nodes is optimized and updated.

[0213] As a possible implementation, the process optimization module 101 may also optimize and update the correspondence between the obtained indicator parameters of the nodes of the sub-IVR process and the optimization solution based on the test results.

[0214] The division of modules in the embodiments of the present application is illustrative and is merely a logical functional division. In actual implementation, other division methods may be used. Furthermore, the functional modules in the various embodiments of the present application may be integrated into a single processor, or may exist physically separately, or two or more modules may be integrated into a single module. The aforementioned integrated modules may be implemented in the form of hardware or software functional modules.

[0215] If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including a number of instructions for enabling a terminal device (which can be a personal computer, mobile phone, or network device, etc.) or a processor to execute all or part of the steps of the method of each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), magnetic disk or optical disk, etc., various media that can store program code.

[0216] The present application further provides a computing device 600 as shown in Figure 6. The computing device 600 includes a bus 601, a processor 602, a communication interface 603, and a memory 604. The processor 602, the memory 604, and the communication interface 603 communicate with each other via the bus 601.

[0217] Among them, the processor 602 can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc.

[0218] The memory 604 can be a dynamic random access memory (DRAM). In addition to DRAM, the memory 604 can also be other random access memories, such as static random access memory (SRAM). In addition, the memory 602 can also be a read-only memory (ROM). As for the read-only memory, for example, it can be a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), etc. The memory 604 can also be a flash memory medium (FLASH), a hard disk drive (HDD) or a solid state drive (SSD). The memory 604 stores computer program instructions, and the processor 602 executes the computer program instructions to execute the steps performed by the device 100 where the process optimization module is located in the method described in Figures 3 and 4. The memory 604 can also include software modules required for other running processes such as the operating system. The operating system can be LINUX TM ,UNIX TM ,WINDOWS TM wait.

[0219] The present application also provides a computing device system, comprising at least one computing device 700 as shown in FIG7 . The computing device 700 comprises a bus 701, a processor 702, a communication interface 703, and a memory 704. The processor 702, the memory 704, and the communication interface 703 communicate with each other via the bus 701. The at least one computing device 700 in the computing device system communicates with each other via a communication path.

[0220] The specific types of the processor 702 and the memory 704 can be found in the description of the processor 602 and the memory 604, which will not be described here. The processor 702 executes the computer program instructions stored in the memory 704 to execute part or all of the steps performed by the device 100 where the process optimization module is located in the method described in Figures 3 and 4. The memory may also include other software modules required for running processes, such as an operating system. The operating system may be LINUX TM ,UNIX TM ,WINDOWS TM wait.

[0221] At least one computing device 700 in the computing device system establishes communication with each other via a communication network, and each computing device 700 runs any one or multiple modules in the process optimization apparatus 500 .

[0222] The above embodiments can be implemented in whole or in part through software, hardware, firmware, or any other combination. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded or executed on a computer, the processes or functions described in accordance with the embodiments of the present invention are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server or data center that contains a collection of one or more available media. The available medium can be a magnetic medium (e.g., a floppy disk, hard disk, tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid state drive (SSD).

[0223] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0224] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the present application. It should be understood that each flow and / or box in the flow chart and / or block diagram, as well as the combination of the flow chart and / or box in the flow chart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device for implementing the functions specified in one or more flow charts and / or one or more boxes in the block diagram.

[0225] These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce a product including an instruction device that implements the functions specified in one or more processes in the flowchart and / or one or more boxes in the block diagram.

[0226] These computer program instructions may also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more flows of the flowchart and / or one or more blocks of the block diagram. Obviously, those skilled in the art may make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include such modifications and variations.

Claims

1. A process optimization method, characterized in that, Including: Obtain the values of at least one service parameter of each sub-IVR process in the interactive voice response (IVR) process within the target period. Based on the value of each service parameter and the corresponding reference value of each service parameter, determine the target sub-IVR process from the various sub-IVR processes; Determine, from the multiple nodes included in the target sub-IVR process, the target node in the target sub-IVR process that is associated with the service parameter; Select a target optimization plan from the set of optimization plans, where the target optimization plan is used to optimize the target node; Optimize the target sub-IVR process based on the target optimization plan.

2. The method according to claim 1, characterized in that, The determining, from the multiple nodes included in the target sub-IVR process, the target node in the target sub-IVR process that is associated with the service parameter includes: Determine the target node based on the pre-obtained association relationship between the service parameter of the sub-IVR process and the metric parameter of the node.

3. The method according to claim 2, characterized in that, The determining the target node based on the pre-obtained association relationship between the service parameter of the sub-IVR process and the metric parameter of the node includes: Based on the association relationship, obtain each association coefficient, where any association coefficient characterizes the degree of association between any metric parameter of each node included in the target sub-IVR process and the service parameter; Determine the target node based on each of the association coefficients.

4. The method according to claim 3, characterized in that, The determining the target node based on each of the association coefficients includes: Based on each of the association coefficients and the values of the metric parameters of the various nodes, obtain a ranking order representing the superiority and inferiority between the various nodes; Based on the ranking order between the various nodes, determine a set number of the target nodes from the various nodes.

5. The method according to claim 4, wherein The obtaining a ranking order representing the superiority and inferiority between the various nodes based on each of the association coefficients and the values of the metric parameters of the various nodes includes: For each metric parameter, based on the values of the various nodes in this metric parameter, obtain the sub-ranking order of the various nodes in this metric parameter; Based on each of the association coefficients and the sub-ranking orders of the various nodes, obtain a ranking order representing the superiority and inferiority between the various nodes.

6. The method according to claim 2, wherein The association relationship between the service parameter of the sub-IVR process and the metric parameter of the node is obtained based on a trained association relationship model, and the association relationship model is used to determine the metric parameter of the node associated with the input service parameter of the sub-IVR process according to the input service parameter of the sub-IVR process.

7. The method according to claim 1, wherein The selecting a target optimization plan from the set of optimization plans includes: Based on the pre-obtained correspondence between the metric parameter of the node of the sub-IVR process and the optimization plan, determine, from the set of optimization plans, the optimization plan corresponding to the value of the metric parameter of the target node of the target sub-IVR process as the target optimization plan.

8. The method according to claim 7, wherein The determining, from the set of optimization plans, the optimization plan corresponding to the value of the metric parameter of the target node of the target sub-IVR process as the target optimization plan includes: Determine at least one optimization solution corresponding to the value of the metric parameter of the target node of the target sub-IVR process in the set of optimization solutions, and display each of the candidate solutions; Determine the target optimization solution according to the candidate solution selected by the user or the candidate solution modified by the user.

9. The method according to claim 7, wherein The correspondence between the metric parameter of the node of the sub-IVR process and the optimization solution is obtained based on the trained optimization solution library model, and the optimization solution library model is used to determine the optimization solution corresponding to the value of the metric parameter of the node of the input sub-IVR process according to the value of the metric parameter of the node of the input sub-IVR process.

10. The method according to claim 1, characterized in that, The determining of the target sub-IVR process from the respective sub-IVR processes based on the value of each service parameter and the corresponding reference value of each service parameter includes: For each sub-IVR process, count the difference between the value of each service parameter of the sub-IVR process and the reference value of the corresponding service parameter; and count the weighted value between the preset weight of each service parameter of the sub-IVR process and the corresponding difference; Based on the weighted value of each sub-IVR process and the reference weighted value between the reference value of each service parameter and the preset weight, determine the target sub-IVR process from each sub-IVR process.

11. The method according to any one of claims 1 to 10, characterized in that, Further includes: Conduct a comparative test on the optimized target sub-IVR process and the pre-optimization target sub-IVR process. If the value of the service parameter of the optimized target sub-IVR process in the test result is inferior to the value of the service parameter of the pre-optimization target sub-IVR process, then restore the currently adopted target sub-IVR process to the pre-optimization target sub-IVR process; Otherwise, if the value of the service parameter of the optimized target sub-IVR process in the test result is superior to the value of the service parameter of the pre-optimization target sub-IVR process, then adopt the optimized target sub-IVR process.

12. The method according to claim 11, wherein Further includes: Based on the values of the service parameters and the values of the metric parameters of the nodes corresponding to the optimized target sub-IVR process and the pre-optimization target sub-IVR process included in the test result, obtain the reference correlation coefficient between the metric parameters of each node and each service parameter in the target sub-IVR process; Based on the reference correlation coefficient, optimize and update the correlation between the service parameters and the metric parameters of the obtained sub-IVR process.

13. The method according to claim 11, characterized in that, Further includes: Based on the test result, optimize and update the correspondence between the metric parameter of the node of the obtained sub-IVR process and the optimization solution.

14. A process optimization device, characterized in that, The device includes a process optimization module, The process optimization module is used to: obtain the values of at least one service parameter of each sub-IVR process in the interactive voice response IVR process during the target period, and determine the target sub-IVR process from the respective sub-IVR processes based on the value of each service parameter and the corresponding reference value of each service parameter; From among the multiple nodes included in the target sub-IVR process, determine the target node in the target sub-IVR process that is associated with the service parameter; select a target optimization solution from the set of optimization solutions, where the target optimization solution is used to optimize the target node. Based on the target optimization solution, optimize the target sub-IVR process.

15. The device according to claim 14, characterized in that, The process optimization module is configured to: Based on the pre-obtained association relationship between the service parameters of the sub-IVR process and the metric parameters of the nodes, determine the target node.

16. The device according to claim 15, characterized in that, The process optimization module is configured to: Based on the association relationship, obtain each association coefficient, where any one of the association coefficients characterizes the degree of association between any metric parameter of each node included in the target sub-IVR process and the service parameter. Based on each of the association coefficients, determine the target node.

17. The device according to claim 16, wherein The process optimization module is configured to: Based on each of the association coefficients and the values of the metric parameters of the nodes, obtain the ranking positions characterizing the superiority and inferiority between the nodes. Based on the ranking positions between the nodes, determine a set number of the target nodes from among the nodes.

18. The device according to claim 17, characterized in that, The process optimization module is configured to: For each metric parameter, based on the values of the nodes in this metric parameter, obtain the sub-ranking positions of the nodes in this metric parameter; based on each of the association coefficients and the sub-ranking positions of the nodes, obtain the ranking positions characterizing the superiority and inferiority between the nodes.

19. The device according to claim 14, wherein, The process optimization module is configured to: Based on the pre-obtained correspondence between the metric parameters of the nodes of the sub-IVR process and the optimization solutions, determine, from the set of optimization solutions, the optimization solution corresponding to the value of the metric parameter of the target node of the target sub-IVR process as the target optimization solution.

20. The device according to claim 19, characterized in that, The process optimization module is configured to: Determine at least one optimization solution corresponding to the value of the metric parameter of the target node of the target sub-IVR process in the set of optimization solutions as candidate solutions, and display each of the candidate solutions; determine the target optimization solution according to the candidate solution selected by the user or the candidate solution modified by the user.

21. The device according to claim 14, wherein The process optimization module is configured to: For each sub-IVR process, count the difference between the value of each service parameter of this sub-IVR process and the reference value of the corresponding service parameter. And count the weighted value between the preset weight of each service parameter of this sub-IVR process and the corresponding difference. Based on the weighted value of each sub-IVR process and the reference weighted value between the reference value of each service parameter and the preset weight, determine the target sub-IVR process from each sub-IVR process.

22. The device according to any one of claims 14-21, characterized in that The process optimization module is further configured to: Conduct a comparison test on the optimized target sub-IVR process and the non-optimized target sub-IVR process. If the value of the service parameter of the optimized target sub-IVR process in the test result is inferior to the value of the service parameter of the non-optimized target sub-IVR process, then restore the currently adopted target sub-IVR process to the non-optimized target sub-IVR process. Otherwise, if the value of the service parameter of the optimized target sub-IVR process in the test result is better than that of the target sub-IVR process before optimization, then the optimized target sub-IVR process is adopted.

23. The device according to claim 22, characterized in that, The process optimization module is further configured to: Based on the values of the service parameters and the index parameters of the nodes corresponding to the optimized target sub-IVR process and the target sub-IVR process before optimization included in the test result, obtain the reference correlation coefficient between the index parameters and the service parameters of each node in the target sub-IVR process; Based on the reference correlation coefficient, optimize and update the correlation between the service parameters of the obtained sub-IVR process and the index parameters of the nodes.

24. The device according to claim 22, wherein, The process optimization module is further configured to: Based on the test result, optimize and update the correspondence between the index parameters of the nodes of the obtained sub-IVR process and the optimization scheme.

25. A computing device, characterized in that, The computing device includes a processor and a memory, and the processor is configured to call computer program instructions stored in the memory to execute the method according to any one of claims 1 to 13.

Images