WO2019179043A1

WO2019179043A1 - Voice message method and apparatus, computer device and storage medium

Info

Publication number: WO2019179043A1
Application number: PCT/CN2018/102864
Authority: WO
Inventors: 吴华
Original assignee: 平安科技（深圳）有限公司
Priority date: 2018-03-22
Filing date: 2018-08-29
Publication date: 2019-09-26
Also published as: CN108632475B; CN108632475A

Abstract

Provided are a voice message method and apparatus. The method comprises: acquiring a voice message file, wherein the voice message file is generated according to a message recorded by a customer; acquiring message relevant information for identifying the message; determining a target agent to process the message, according to the voice message file and/or the message relevant information; and allocating, to the target agent, a processing task for processing the message. Intelligently determining a target agent to process a message, according to a voice message file and/or message relevant information, can realize directed allocation and improve the efficiency of voice message processing. Intelligently determining a target agent to process the message, according to a voice message file and/or message relevant information, can realize directed allocation and improve the efficiency of voice message processing.

Description

Voice message method, device, computer device and storage medium

This application claims priority to Chinese Patent Application No. 201810240134.9, entitled "Voice Message Method, Apparatus, Computer Equipment, and Storage Media", filed on March 22, 2018, the entire contents of which are incorporated by reference. In this application.

Technical field

The present application relates to the field of voice message processing technologies, and in particular, to a voice message method and apparatus.

Background technique

At present, IVR (Interactive Voice Response) messages usually use the voxmail incoming line method, which requires the allocation of IVR extension resources. After the IVR message is successfully sent, the vox message voice file is sent to the corresponding mailbox, and then the employee obtains the message voice in the mailbox and processes it.

The inventor realized that the types of services involved in the message are various, some are A business, some are B business, because the voice files are fixedly assigned to a specific mailbox, and then obtained by the customer service, then the customer service is required to master all Type of business. If the exclusive customer service is to serve the exclusive business, the targeted distribution cannot be achieved because the message is fixedly allocated.

Moreover, the inventors have found that the traditional approach adds complexity and risk of error due to the introduction of mailboxes. It is inefficient to identify the voice of the message by the staff alone.

Summary of the invention

The object of the present application is to solve at least one of the above-mentioned technical drawbacks, in particular, the technical drawback of the inability to achieve targeted allocation.

The application provides a voice message method, and the method includes the following steps:

Obtaining a message voice file, the message voice file is generated according to a message input by the client; obtaining message related information for identifying the message; determining, according to the message voice file and/or the message related information, a target for processing the message An agent; a processing task of processing the message is assigned to the target agent.

The application further provides a voice message device, the device includes: an acquisition module, configured to acquire a message voice file, the message voice file is generated according to a message input by the client; and obtaining message related information for identifying the message; a determining module, configured to determine a target agent that processes the message according to the message voice file and/or message related information; and a task assignment module, configured to allocate the processing task to the target agent.

The application also provides a computer device comprising a memory and a processor, the memory storing computer readable instructions, the computer readable instructions being executed by the processor, causing the processor to perform a voice message The method, wherein the voice message comprises the steps of: acquiring a message voice file, the message voice file is generated according to a message input by the client; acquiring message related information for identifying the message; according to the message voice file and/or the The message related information determines a target agent that processes the message; and the processing task that processes the message is assigned to the target agent.

The present application also provides a non-volatile storage medium storing computer readable instructions that, when executed by one or more processors, cause one or more processors to perform a voice message method, The voice message includes the following steps: acquiring a message voice file, the message voice file is generated according to a message input by the client; acquiring message related information for identifying the message; and according to the message voice file and/or the message The related information determines a target agent that processes the message; assigns a processing task that processes the message to the target agent.

The above-mentioned voice message method and device can intelligently determine the target agent that processes the message according to the message voice file and/or the message related information, so that targeted distribution can be realized, and the voice message processing efficiency is improved.

DRAWINGS

1 is a schematic flow chart of a voice message method according to an embodiment;

2 is a schematic diagram of a voice message device module of an embodiment;

FIG. 3 is a schematic diagram showing the internal structure of a computer device in an embodiment.

detailed description

FIG. 1 is a schematic flow chart of a voice message method according to an embodiment.

Step S10: Acquire a message voice file, and the message voice file is generated according to a message input by the client; and obtain message related information for identifying the message.

A voice file is a voice file generated based on a customer's message. The message related information refers to information related to the message, and may include, for example, a voice path ID, a start message time, an end message time, an incoming line series, an incoming area code, a customer contact number (a phone number, a website registration account, or an instant messaging account). At least one of a message voice file storage address, a message voice file name, and a message valid flag. The voice channel ID is the unique identifier of each call, that is, each call has its own unique ID, such as a string of 16-digit hexadecimal characters, which is used to uniquely mark the call, such as: 015301cd23f3a1e1. The incoming line series is the business series of the number dialed by the customer, such as the Lujin series, the gold service series, etc., such as 4008666618 is the Lujin series. The incoming area code is the area code of the number dialed by the customer. For example, 0755 is the Shenzhen area code.

After entering the message program, the server generates a message voice file according to the message entered by the client and stores the message voice file, and records message related information identifying the message. Specifically, when the instruction of the customer message is received, the message program is entered. Record the start message time and issue a beep like “Drip” to remind the customer to start the message. After the customer's message, you can select the # key or hang up to complete the message, and generate a message voice file to record the end message time. Before the message can be broadcast, the voice prompts the customer: the message flow has been entered, and the message starts to leave a message at the beginning of the prompt tone.

If the voice is not recognized after the prompt "drop" sound starts and the customer hangs up, the message is an invalid message. Or, it hangs up before the sound of the "drop" sound starts, and is also recognized as an invalid message. A voice file will only be generated if the message that produced the voice after the start of the tone is a valid message.

Therefore, in one embodiment, after entering the message program, before the message voice file is generated, by identifying the real-time decibel value of the sound, if the sound of the real-time sound decibel value exceeding the preset threshold value is recognized within the preset time period, then Determining the recognition of the voice, generating a message voice file and storing the message voice file to the server, recording the message related information; otherwise, determining that the voice is not recognized, only the message related information is recorded.

After being recognized as a valid message, the voice file is stored separately according to the message time or storage time. It can be stored in a NAS (Network Attached Storage) on a daily or monthly folder, for example, in a folder every day. All voice files belonging to the day are stored in the corresponding day. In the folder. If it is a new day or a new month, the storage folder does not exist at this time, and the storage folder is automatically created before the message is stored.

Then, the message related information such as the session ID, the start message time, the end message time, the incoming line series, the incoming line area number, the customer contact number, the message voice file storage address, the message voice file name, and the message valid flag are stored in the database. Record, see Table 1.

Table 1

Step S20: Determine a target agent that processes the message according to the message voice file and/or the message related information.

The target agent that processes the message may be determined according to the voice message file, or the target agent that processes the message may be determined according to the message related information, or the target agent that processes the message may be jointly determined according to the message voice file and the message related information.

In some embodiments, the customer information may be determined based on the voicemail file and/or the message related information, and the target agent handling the message may be determined based on the customer information. Customer information refers to the customer's personal information, such as name, customer number, ID number, contact information, and so on.

Specifically, in one embodiment, determining customer information according to the message voice file and/or the message related information, and determining, according to the customer information, that the target agent that processes the message includes:

Determining whether the customer contact number in the message related information is stored in a preset database, if yes, determining the corresponding customer information according to the customer contact number, determining the service information that the customer has purchased according to the customer information, and determining according to the service information that the customer has purchased. The target agent of the message is processed. Or, determining whether the customer contact number in the message related information is stored in the database, or otherwise performing voice recognition on the message voice file, to determine the target agent that processes the message according to the recognition result of the voice recognition. The preset database stores various customer information and voice recognition related data information, which can be stored locally on the server.

Service information refers to information that provides services to customers. Services such as insurance services, financial services, medical services, etc. The following is an introduction to insurance services, that is, the service information that has been purchased is the policy information that has been insured.

For example, if the customer contact number is the mobile phone number 13912345678, the customer corresponding to the mobile phone number is searched in the database according to the mobile phone number. If the corresponding customer is stored, the service information that the customer has purchased is determined according to the customer information, and if there is no corresponding stored The client then performs voice recognition on the voice message to determine the target agent that processes the message based on the recognition result. The voice recognition here may be to identify the voiceprint feature of the customer, or to identify at least one keyword in the voice file of the message.

Therefore, in some embodiments, performing voice recognition on the voice voice file, and determining, according to the recognition result of the voice recognition, that the target agent that processes the message includes: determining whether the customer contact number in the message related information is stored in a database, If the voice message is otherwise voice-recognized, the voiceprint feature of the customer is identified. The voiceprint features are aligned in a database to find customer information that matches the voiceprint feature. The service information that the customer has purchased is determined according to the customer information, and the target agent that processes the message is determined according to the service information that the customer has purchased. Further, the voiceprint feature and the customer contact number map can also be stored in the database.

The advantage of identifying customers through voiceprint features is that when customers use other phone numbers to leave messages or when they call, they can identify the customer without entering any personal information or policy information, thereby improving the user experience.

In another embodiment, the voice message is voice-recognized, and determining, according to the recognition result of the voice recognition, that the target agent that processes the message includes: determining whether the customer contact number in the message-related information is stored in a database, If the voice message is otherwise voice-recognized, at least one keyword in the voice file is identified. The identified keywords are compared in the database to determine the purpose of the message. For example, if the keyword is identified as "surrender", it is determined that the purpose of the customer's message may be a question about how to surrender. The target agent who processes the message is determined according to the purpose of the message. If the purpose of the message is judged as surrender, it can be subsequently assigned to the customer service agent who specializes in surrendering and complaining.

In order to better serve the customer, in some embodiments, the customer's emotion is also recognized during the speech recognition process of the voice message to be prompted according to the customer's mood or to determine whether to preferentially process the message. That is, the customer's emotional information is identified according to the voice message file; the user is prompted according to the emotion information when assigning the processing task or whether the message is preferentially processed. Specifically, the emotion index is determined according to a speech recognition feature keyword and/or a voice tone calculation mood index; the emotion index F=M×a+V×b, where M is a feature identified in the message voice file The number of keywords, V is the average decibel value of the message sound, and a and b are preset weight coefficients greater than or equal to 0, which can be set according to the actual situation; the emotional information of the customer is determined according to the emotion index. The higher the sentiment index, the worse the customer's sentiment is, and the agent needs to be given priority.

After the customer information is determined by the customer information, the target agent who processes the message is determined according to the service information that has been purchased.

Continuing with the insurance service as an example, in some embodiments, the original sales agent or sales team of the policy may be determined based on the policy information, the original sales agent is determined to be the agent who processes the message information, or the message information is determined to be processed from the sales team. The seat. It is handled by the original sales agent or sales team. Because the original sales agent or sales team knows more about the customer and the policy, it is more targeted and improves the user experience and processing efficiency.

In still other embodiments, the policy series can be determined based on the customer policy information, and the agent handling the message is determined in the exclusive customer service agent of the policy series. For example, if the policy series is determined as the life insurance series according to the customer policy information, the agent who handles the message information is determined from the customer service agent of the life insurance series.

However, there may be cases where the customer has multiple services (such as a policy), and it is necessary to determine a feature service (feature policy) from the plurality of services for determining the customer service information (policy information) and determining the agent handling the message. the process of. The following continues to use insurance services as an example.

In one of the embodiments, the determining, by the customer information, the process of processing the target agent of the message comprises: determining service information that has been purchased based on the customer information, the service information including policy information. If the customer has purchased more than one policy, the feature policy is determined according to at least one of the start date or the contract signing date, the latest customer contact time, and the insurance amount of each policy, and the processing is determined according to the feature policy determination process. The target seat of the message.

You can determine the feature policy in one of the following ways:

Method 1: Obtain the start date or contract signing date of each policy, and determine the latest policy as a feature policy. For example, the policy corresponding to the customer has policy A and policy B. If the policy information is found to be insured on December 12, 2016 and the insured date of policy B is February 12, 2017, then policy B is selected. For feature insurance.

Method 2: Obtain the latest customer contact time of each policy, and determine the policy of the most recent customer contact as a feature policy. For example, the policy corresponding to the customer has policy A and policy B. If the daily maintenance information of the policy is found, the customer can check the policy A for the latest inquiry, consultation, complaint, etc. The contact time is on September 20, 2017, and the policy B is The latest inquiry, consultation, complaints, etc., is on January 10, 2018, and policy B is selected as the characteristic policy.

Method 3: Obtain the insurance amount of each policy, and determine the policy with the largest insurance amount as the characteristic policy. For example, the policy corresponding to the customer has policy A and policy B. If the policy amount of the policy A is higher than the policy amount of the policy B, the policy A is selected as the characteristic policy.

Method 4: According to the policy start date (or contract signing date) T, the recent customer contact time R, and the insurance amount M, at least two parameters are jointly determined to determine the feature policy. For example, a judgment function F(X) can be set for the policy X as needed, and then the size of the judgment function F(X) is determined by using T, R, and M as the judgment factors, and the policy with the largest F(X) is the feature policy. Specifically, F(X)=T×a+R×b+M×c, where a, b, and c are preset weight coefficients greater than or equal to 0, which can be set according to actual conditions.

In some embodiments, the target agent that processes the message may also be determined based on the incoming line series or the incoming line number in the message related information.

Specifically, the customer service agent matching the incoming line series is determined according to the incoming line series for processing. For example, the incoming line series is the Lujin series, and the agent who handles the message is determined from the customer service seat specially designed for the Lujin series. Specifically, the processing is performed according to the incoming line area number and the customer service agent in the corresponding area. For example, if the entry area code is 0755, it is known as Shenzhen by querying the area code of the area code, and the agent handling the message is determined from the customer service agent in the Shenzhen area.

After the target agent is determined, step S30 is performed.

Step S30: Allocating a processing task for processing the message to the target agent. The target agent obtains a processing task, and the processing task records at least one task, and the target agent obtains message related information of the corresponding message from the database according to the task. And according to the message related information, the corresponding message voice file is obtained from the NAS.

The target agent can determine whether it needs to be processed according to the message related information and the voice message file, and if necessary, process according to the message related information. If the agent has already processed it, add the corresponding processing ID to the task in the database. Specifically, the return visit identifier and/or handler information are added to the database according to the processing completion instruction of the target agent, and the database is updated. See Table 2.

Table 2

In one embodiment, during the target agent return visit, the evaluation data fed back by the customer may also be received, and the evaluation data is added to the database. Evaluation can be by rating, star rating, likes, etc. The evaluation data can be used as reference data for evaluating the working conditions of the target agent.

2 is a schematic diagram of a voice message device module of an embodiment.

Corresponding to the above-mentioned voice message method, the present application further provides a voice message device, and the device includes: an acquisition module 100, an agent determination module 200, and a task assignment module 300.

The obtaining module 100 is configured to obtain a message voice file, where the message voice file is generated according to a message input by the client, and obtain message related information for identifying the message; the agent determining module 200 is configured to use the voice file and/or the message according to the message The related information determines a target agent that processes the message; the task assignment module 300 is configured to allocate a processing task for processing the message to the target agent.

Upon entering the message program, the obtaining module 100 acquires a message voice file, and obtains message related information for identifying the message, and the message voice file is generated according to the message recorded by the client.

A voice file is a voice file generated based on a customer's message. The message related information refers to information related to the message, and may include, for example, a voice path ID, a start message time, an end message time, an incoming line series, an incoming area code, a customer contact number (a phone number, a website registration account, or an instant messaging account). At least one of a message voice file storage address, a message voice file name, and a message valid flag. The session ID is a unique identifier for each call, that is, each call has its own unique ID, such as a string of 16-digit hexadecimal characters, which is used to uniquely mark the call, such as: 015301cd23f3a1e1. The incoming line series is the business series of the number dialed by the customer, such as the Lujin series, the gold service series, etc., such as 4008666618 is the Lujin series. The incoming area code is the area code of the number dialed by the customer. For example, 0755 is the Shenzhen area code.

After entering the message program, the server generates a message voice file according to the message entered by the client and stores the message voice file, and records message related information identifying the message. When the instruction of the customer message is received, the message program is entered. The obtaining module 100 records the start message time and issues a tone similar to "drop" to prompt the customer to start the message. After the customer's message, you can select the # key or hang up to complete the message, and generate a message voice file to record the end message time. The pre-message acquisition module 100 can broadcast the voice prompting client: the message flow has been entered, and the message begins to be messaged at the beginning of the prompt tone.

Therefore, in one embodiment, after entering the message program, before the message voice file is generated, the acquisition module 100 recognizes the real-time decibel value of the sound, and if the real-time sound decibel value exceeds the preset threshold value within the preset time length, Then, it is determined that the voice is recognized, the voice voice file is generated, and the voice voice file is stored to the server, and the message related information is recorded; otherwise, it is determined that the voice is not recognized, and only the message related information is recorded.

After the obtaining module 100 recognizes the valid message, the message voice file is separately stored according to the message time or the storage time. It can be stored in a NAS (Network Attached Storage) on a daily or monthly folder, for example, in a folder every day. All voice files belonging to the day are stored in the corresponding day. In the folder. If it is a new day or a new month, the storage folder does not exist at this time, and the storage folder is automatically created before the message is stored.

Then, the obtaining module 100 stores the message related information such as the session ID, the start message time, the end message time, the incoming line series, the incoming line area number, the customer contact number, the message voice file storage address, the message voice file name, and the message valid flag. Record in the database, see Table 3.

table 3

The agent determination module 200 determines a target agent that processes the message based on the message voice file and/or message related information.

The agent determining module 200 may determine, according to the voice message file, a target agent that processes the message, or determine a target agent that processes the message according to the message related information, or jointly determine a target agent that processes the message according to the message voice file and the message related information. .

In some embodiments, the agent determination module 200 can determine customer information based on the voice message and/or message related information, and determine a target agent to process the message based on the customer information. Customer information refers to the customer's personal information, such as name, customer number, ID number, contact information, and so on.

Specifically, in one embodiment, the agent determining module 200 is configured to: determine whether the customer contact number in the message related information is stored in a preset database, and if yes, determine corresponding customer information according to the customer contact number, according to the customer information. The service information that the customer has purchased is determined, and the target agent that processes the message is determined according to the service information that the customer has purchased. Or, determining whether the customer contact number in the message related information is stored in the database, or otherwise performing voice recognition on the message voice file, to determine the target agent that processes the message according to the recognition result of the voice recognition. The preset database stores various customer information and voice recognition related data information, which can be stored locally on the server.

For example, the customer contact number is the mobile phone number 13912345678, and the agent determination module 200 searches the database for the customer corresponding to the mobile phone number according to the mobile phone number. If the corresponding customer is stored, the service information that the customer has purchased is determined according to the customer information, if not If the corresponding customer is stored, the voice voice file is voice-recognized to determine the target agent that processes the message according to the recognition result. The voice recognition here may be to identify the voiceprint feature of the customer, or to identify at least one keyword in the voice file of the message.

Therefore, in some embodiments, the agent determination module 200 is configured to: determine whether the customer contact number in the message related information is stored in a database, or otherwise perform voice recognition on the voice message to identify the voiceprint feature of the customer. The voiceprint features are aligned in a database to find customer information that matches the voiceprint feature. The service information that the customer has purchased is determined according to the customer information, and the target agent that processes the message is determined according to the service information that the customer has purchased. Further, the voiceprint feature and the customer contact number map can also be stored in the database.

The advantage of identifying customers through voiceprint features is that the customer can identify the customer without entering any personal information or policy information when using other phone numbers or when calling, thereby improving the user experience.

In another embodiment, the agent determining module 200 is configured to: determine whether the customer contact number in the message related information is stored in a database, and if otherwise, perform voice recognition on the voice file to identify at least one of the voice files. Key words. The identified keywords are compared in the database to determine the purpose of the message. For example, if the keyword is identified as "surrender", it is determined that the purpose of the customer's message may be a question about how to surrender. The target agent who processes the message is determined according to the purpose of the message. If the purpose of the message is judged as surrender, it can be subsequently assigned to the customer service agent who specializes in surrendering and complaining.

In order to better serve the customer, in some embodiments, the agent determination module 200 also identifies the customer's emotion during the voice recognition process of the voice voice file, so as to prompt or determine whether to prioritize according to the customer's emotion when the processing task is subsequently assigned. The message. That is, the customer's emotional information is identified according to the voice message file; the user is prompted according to the emotion information when assigning the processing task or whether the message is preferentially processed. Specifically, the emotion index is determined according to a speech recognition feature keyword and/or a voice tone calculation mood index; the emotion index F=M×a+V×b, where M is a feature identified in the message voice file The number of keywords, V is the average decibel value of the message sound, and a and b are preset weight coefficients greater than or equal to 0, which can be set according to the actual situation; the emotional information of the customer is determined according to the emotion index. . The higher the sentiment index, the worse the customer's sentiment is, and the agent needs to be given priority.

After determining the service information purchased by the customer through the customer information, the agent determination module 200 determines the target agent that processes the message according to the service information that has been purchased.

Continuing with the insurance service as an example, in some embodiments, the agent determination module 200 may determine the original sales agent or sales team of the policy based on the policy information, determine the original sales agent as the agent who processes the message information, or determine from the sales team. The agent who processed the message. It is handled by the original sales agent or sales team. Because the original sales agent or sales team knows more about the customer and the policy, it is more targeted and improves the user experience and processing efficiency.

In still other embodiments, the agent determination module 200 can determine the policy series based on the customer policy information, and determine the agent handling the message in the exclusive customer service agent of the policy series. For example, if the policy series is determined as the life insurance series according to the customer policy information, the agent who handles the message information is determined from the customer service agent of the life insurance series.

However, there may be cases where the customer has multiple services (such as a policy), and the agent determination module 200 needs to determine a feature service (feature policy) from the plurality of services for determining the customer service information (policy information) and determining The process of handling the agent's seat. The following continues to use insurance services as an example.

In one embodiment, the agent determination module 200 is configured to: determine service information that has been purchased based on customer information, the service information including policy information. If the customer has purchased more than one policy, the feature policy is determined according to at least one of the start date or the contract signing date, the latest customer contact time, and the insurance amount of each policy, and the processing is determined according to the feature policy determination process. The target seat of the message.

The agent determination module 200 can determine the feature policy in one of the following ways:

Method 1: Obtain the start date or contract signing date of each policy and determine the latest policy as a feature policy. For example, the policy corresponding to the customer has policy A and policy B. If the policy information is found to be insured on December 12, 2016 and the insured date of policy B is February 12, 2017, then policy B is selected. For feature insurance.

In some embodiments, the agent determination module 200 may also determine a target agent that processes the message based on the incoming line series or the incoming line number in the message related information.

Specifically, the agent determination module 200 determines, according to the incoming line series, the customer service agent that matches the incoming line series to perform processing. For example, the incoming line series is the Lujin series, and the agent who handles the message is determined from the customer service seat specially designed for the Lujin series. Specifically, the agent determination module 200 determines to process with the customer service agent of the corresponding area according to the incoming line area number. For example, if the entry area code is 0755, it is known as Shenzhen by querying the area code of the area code, and the agent handling the message is determined from the customer service agent in the Shenzhen area.

After the agent determination module 200 determines the target agent, the task assignment module 300 assigns a processing task for processing the message to the target agent. The target agent obtains a processing task, and the processing task records at least one task, and the target agent acquires message related information of the corresponding message from the database according to the task. And according to the message related information, the corresponding message voice file is obtained from the NAS.

The target agent can determine whether it needs to be processed according to the message related information and the voice message file, and if necessary, process according to the message related information. If the agent has already processed it, add the corresponding processing ID to the task in the database. Specifically, the return visit identifier and/or handler information are added to the database according to the processing completion instruction of the target agent, and the database is updated. See Table 4.

Table 4

In one of the embodiments, an evaluation module may also be included. During the return visit of the target agent, the evaluation module can also receive the evaluation data fed back by the customer and add the evaluation data to the database. Evaluation can be by rating, star rating, likes, etc. The evaluation data can be used as reference data for evaluating the working conditions of the target agent.

The application also provides a computer device comprising a memory and a processor, the memory storing computer readable instructions, the computer readable instructions being executed by the processor, causing the processor to perform any of the above implementations The steps of the voice message method are described.

FIG. 3 is a schematic diagram showing the internal structure of a computer device in an embodiment. As shown in FIG. 3, the computer device includes a processor, a non-volatile storage medium, a memory, and a network interface connected by a system bus. The non-volatile storage medium of the computer device stores an operating system, a database, and computer readable instructions. The database may store a sequence of control information. When the computer readable instructions are executed by the processor, the processor may implement a processor. The voice message method, the processor can implement the functions of the acquisition module, the agent determination module, and the task assignment module in the voice message device in the embodiment shown in FIG. 2. The processor of the computer device is used to provide computing and control capabilities to support the operation of the entire computer device. Computer readable instructions may be stored in the memory of the computer device, the computer readable instructions being executable by the processor to cause the processor to perform a voice message method. The network interface of the computer device is used to communicate with the terminal connection. It will be understood by those skilled in the art that the structure shown in FIG. 3 is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation of the computer device to which the solution of the present application is applied. The specific computer device may It includes more or fewer components than those shown in the figures, or some components are combined, or have different component arrangements.

The application also provides a non-volatile storage medium storing computer readable instructions, when executed by one or more processors, causing one or more processors to perform any of the above embodiments The steps of the voice message method.

The voice message method and device, the message voice file is obtained, the message voice file is generated according to the message recorded by the client; the message related information for identifying the message is obtained; and the message voice file and/or the message is related according to the message The information determines a target agent that processes the message; a processing task that processes the message is assigned to the target agent. The target agent that processes the message is intelligently determined according to the message voice file and/or the message related information, so that targeted allocation can be realized, and the voice message processing efficiency is improved.

A person skilled in the art can understand that all or part of the process of implementing the above embodiment method can be completed by a computer program to instruct related hardware, and the computer program can be stored in a computer readable storage medium. When executed, the flow of an embodiment of the methods as described above may be included. The foregoing computer readable storage medium may be a non-volatile storage medium such as a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM). Wait.

Claims

A voice message method, the method comprising the following steps:

Acquiring a voice file of the message, the message voice file is generated according to the message recorded by the client; and obtaining message related information for identifying the message;

Determining, according to the message voice file and/or the message related information, a target agent that processes the message;

A processing task of processing the message is assigned to the target agent.
The voice message method according to claim 1, wherein the determining, according to the message voice file and/or the message related information, the target agent that processes the message comprises:

Determining customer information according to the message voice file and/or the message related information, and determining a target agent that processes the message according to the customer information.
The voice message method according to claim 2, wherein the determining the customer information according to the message voice file and/or the message related information, and determining, according to the customer information, that the target agent processing the message comprises:

Determining whether the customer contact number in the message related information is stored in a preset database;

If yes, determining corresponding customer information according to the customer contact number, determining service information that the customer has purchased according to the customer information, and determining a target agent that processes the message according to the service information that the customer has purchased; or

If not, the voice message is voice-recognized, and the target agent who processes the message is determined according to the recognition result of the voice recognition.
The voice message method according to claim 3, wherein the voice message is voice-recognized, and the target agent that processes the message is determined according to the recognition result of the voice recognition:

Performing voice recognition on the voice file to identify the voiceprint feature of the customer;

Aligning the voiceprint features in the database to find customer information that matches the voiceprint features;

Determining, according to the customer information, service information that the customer has purchased, and determining a target agent that processes the message according to the service information.
The voice message method according to claim 3, wherein the voice message is voice-recognized, and the target agent that processes the message is determined according to the recognition result of the voice recognition:

Performing voice recognition on the voice message to identify at least one keyword in the voice file;

Identifying the identified keywords in the database to determine the purpose of the message;

A target agent that processes the message is determined according to the message purpose.
The voice message method according to claim 3, wherein the voice recognition of the voice voice file comprises:

Identifying the emotional information of the client according to the message voice file; prompting according to the emotion information when determining the processing task or determining whether to preferentially process the message.
The voice message method according to claim 6, wherein the identifying the emotion information of the customer according to the message voice file comprises:

Determining the sentiment index according to a speech recognition feature keyword and/or a voice tone calculation mood index; the emotion index F=M×a+V×b, wherein M is a feature keyword number recognized in the message voice file , V is the average decibel value of the message sound, and a and b are preset weight coefficients greater than or equal to 0;

The emotional information of the customer is determined according to a sentiment index.
A voice message device, the device comprising:

An obtaining module, configured to obtain a message voice file, where the message voice file is generated according to a message input by the client; and acquiring message related information for identifying the message;

An agent determining module, configured to determine a target agent that processes the message according to the voice message and/or the message related information;

a task assignment module for assigning a processing task to the target agent.
A computer device comprising a memory and a processor, the memory storing computer readable instructions, the computer readable instructions being executed by the processor, causing the processor to perform a voice message method;

The voice message method includes the following steps:

Acquiring a voice file of the message, the message voice file is generated according to the message recorded by the client; and obtaining message related information for identifying the message;

Determining, according to the message voice file and/or the message related information, a target agent that processes the message;

A processing task of processing the message is assigned to the target agent.
The computer apparatus according to claim 9, wherein said determining, based on said message voice file and/or said message related information, a step of processing a target agent of said message comprises the steps of:

Determining customer information according to the message voice file and/or the message related information, and determining a target agent that processes the message according to the customer information.
The computer device according to claim 10, wherein the determining the customer information according to the message voice file and/or the message related information, and determining the target agent to process the message according to the customer information comprises the following steps:

Determining whether the customer contact number in the message related information is stored in a preset database;

If yes, determining corresponding customer information according to the customer contact number, determining service information that the customer has purchased according to the customer information, and determining a target agent that processes the message according to the service information that the customer has purchased; or

If not, the voice message is voice-recognized, and the target agent who processes the message is determined according to the recognition result of the voice recognition.
The computer device of claim 11 wherein the step of speech recognition of the voice message file comprises the steps of:

Identifying the emotional information of the client according to the message voice file; prompting according to the emotion information when determining the processing task or determining whether to preferentially process the message.
The computer device according to claim 12, wherein the step of identifying the emotional information of the client according to the message voice file comprises:

Determining the sentiment index according to a speech recognition feature keyword and/or a voice tone calculation mood index; the emotion index F=M×a+V×b, wherein M is a feature keyword number recognized in the message voice file , V is the average decibel value of the message sound, and a and b are preset weight coefficients greater than or equal to 0;

The emotional information of the customer is determined according to a sentiment index.
A non-volatile storage medium storing computer readable instructions, when executed by one or more processors, causing one or more processors to perform a voice message method;

The voice message method includes the following steps:

Acquiring a voice file of the message, the message voice file is generated according to the message recorded by the client; and obtaining message related information for identifying the message;

Determining, according to the message voice file and/or the message related information, a target agent that processes the message;

A processing task of processing the message is assigned to the target agent.
A non-volatile storage medium storing computer readable instructions according to claim 14, wherein said step of determining a target seat for processing said message based on said message voice file and/or said message related information comprises the following step:

Determining customer information according to the message voice file and/or the message related information, and determining a target agent that processes the message according to the customer information.
A non-volatile storage medium storing computer readable instructions according to claim 15, wherein said determining customer information based on said message voice file and/or said message related information, and determining processing based on said customer information The step of the target agent of the message includes the following steps:

Determining whether the customer contact number in the message related information is stored in a preset database;

If yes, determining corresponding customer information according to the customer contact number, determining service information that the customer has purchased according to the customer information, and determining a target agent that processes the message according to the service information that the customer has purchased; or

If not, the voice message is voice-recognized, and the target agent who processes the message is determined according to the recognition result of the voice recognition.
A nonvolatile storage medium storing computer readable instructions according to claim 16, wherein said speech voice file is speech-recognized, and determining a target seat for processing said message based on said recognition result of said speech recognition The steps include the following steps:

Performing voice recognition on the voice file to identify the voiceprint feature of the customer;

Aligning the voiceprint features in the database to find customer information that matches the voiceprint features;

Determining, according to the customer information, service information that the customer has purchased, and determining a target agent that processes the message according to the service information.
A non-volatile storage medium storing computer readable instructions according to claim 16, wherein said step of voice-recognizing a message voice file and determining a target seat for processing said message based on said recognition result of said voice recognition Includes the following steps:

Performing voice recognition on the voice message to identify at least one keyword in the voice file;

Identifying the identified keywords in the database to determine the purpose of the message;

A target agent that processes the message is determined according to the message purpose.
A non-volatile storage medium storing computer readable instructions according to claim 16, wherein the step of speech recognition of the voice message file comprises the steps of:

Identifying the emotional information of the client according to the message voice file; prompting according to the emotion information when determining the processing task or determining whether to preferentially process the message.
A non-volatile storage medium storing computer readable instructions according to claim 19, wherein the step of identifying the emotional information of the client based on the message voice file comprises:

Determining the sentiment index according to a speech recognition feature keyword and/or a voice tone calculation mood index; the emotion index F=M×a+V×b, wherein M is a feature keyword number recognized in the message voice file , V is the average decibel value of the message sound, and a and b are preset weight coefficients greater than or equal to 0;

The emotional information of the customer is determined according to a sentiment index.