Outbound management method and device, storage medium and computer equipment

Abstract

The invention discloses an outbound call management method and device, a storage medium and computer equipment, and the method comprises the steps: 1, carrying out the outbound call of a target customer, obtaining the content of an outbound call after the outbound call succeeds, marking the call of an outbound call object as a first call, and marking the call of an operator as a second call; wherein in the step, marking is carried out according to a calling object and a called object, that is, a call of the called object is marked as a first call, and a call of an operator belonging to the calling object is marked as a second call; and the call of the automatic voice in the calling process is marked as a third call. According to the outbound management method, the outbound management device, the storage medium and the computer equipment, in the aspect of the outbound management method, the language features and the regional features of the outbound object are predicted and planned, so that an operator can adopt a dialect mode in a specific communication process; through mutual communication with the outbound object, the inquiry conflict feeling of the outbound object is reduced, and the good feeling of the outbound object to the telephone operator is improved.

Description

Technical Field

[0001] The present invention relates to the field of communication technologies, and specifically to an outbound call management method, device, storage medium, and computer device. Background Art

[0002] An outbound call is an active call initiated by a modern customer service center system to a customer. It automatically dials the customer's phone number by computer. When the system connects to the customer's phone, it also connects to the operator at the call center, thus saving the waiting time for the call.

[0003] Currently, with the development of communication technologies, people are receiving an increasing number of unknown calls. Among these unknown calls, there are many abnormal calls for fraud, sales promotion, harassment, etc., which cause great interference to people's lives, leading to the psychological hint that as long as it is an outbound call, it is a fraud, causing a great negative impact on legal and normal outbound call marketing. However, the existing technologies for outbound calls lack corresponding technical solutions for this; Secondly, China has a vast territory and a large number of local dialects. In some rural and township areas, people usually communicate in their respective dialects in daily life. Therefore, some groups can understand Mandarin but may not be able to speak it or may speak it not very standard. In the process of specific outbound call communication, it is easy for both the outbound call object and the operator to have the phenomenon of failing to convey the meaning accurately, reducing the call quality and effect. Summary of the Invention

[0004] (I) Technical Problems to be Solved Aiming at the deficiencies of the existing technologies, the present invention provides an outbound call management method, device, storage medium, and computer device, and solves the problems raised in the above background art.

[0005] (II) Technical Solutions To achieve the above object, the present invention provides the following technical solution: An outbound call management method, including: The first step: Make an outbound call to a target customer. After success, obtain the outbound call content, and mark the call to the outbound call object as the first call, and the call of the operator as the second call; In this step, this step makes marks according to the call object and the called object respectively, that is, marks the call of the called object as the first call, and marks the call of the operator belonging to the call object as the second call; and marks the call of the automatic voice during the call as the third call; The second step: Perform first real-time voice monitoring on the first call. The first real-time voice monitoring is mainly used to monitor the first language features, and the first language features include dialectal colloquialisms, voice pitch, volume, and business keywords; For example, when addressing oneself, people in Northeast China say "an", people in Shanghai say "ala", people in Guangdong have the expression "from ao", people in Xi'an simply call it "hungry", people in Wenzhou use "en". Chaoshan dialect uses "wa"; In terms of voice pitch, those with tight vocal cords and high-pitched voices are generally young people, while those with loose vocal cords and low-pitched voices are generally elderly people. The emotion judgment of the feedback from the outbound call object includes manually collecting the volume of different emotions and pre-setting a database, and using the database to analyze the volume difference and track line of the monitored sound in real time, so as to obtain the real-time emotion of the outbound call object. For example, during the call, the voice signal can be recorded in real time and then imported into the audio editing software. By analyzing the height of the waveform and the corresponding decibel scale, the decibel size of the sound can be judged. Among them, the range threshold of the volume difference representing normal emotions is 30 to 50 decibels, the range threshold of the voice volume difference representing depressed emotions is less than 30 decibels, the range threshold of the volume difference representing excited emotions is 50 to 80 decibels, and the range threshold of the volume difference representing intense emotions is greater than 80 decibels. The business keywords are determined according to the reason for the outbound call, so they vary according to different customer objects; Step 3: Performing a second real-time voice monitoring on the second call, where the second real-time voice monitoring is mainly used to monitor second language features; the second language features include speech speed, volume, feedback keywords, and taboo words; The range threshold of speech speed representing normal emotions is 170-400Hz, 160-280 words / minute, the range threshold of speech speed representing depressed emotions is less than 170-400Hz, 160-280 words / minute, and the range threshold of speech speed representing excited emotions is greater than 170-400Hz, 160-280 words / minute; the range threshold of volume difference representing normal emotions is 30-50 decibels, the range threshold of speech volume difference representing depressed emotions is less than 30 decibels, and the range threshold of volume difference representing excited emotions is greater than 50 decibels; Feedback keywords are based on the perspective of the called party, and their dialects will change accordingly. The distinction between business keywords and feedback keywords is set according to the different monitored objects. This setting makes the monitoring of the call process more accurate and can make judgments and responses efficiently. Taboo words: After determining the regional characteristics based on the dialect characteristics of the called party determined in the second step, the system collects the taboo words recognized by the local people and converts them into text through the display for the operator to browse and self-prompt; Step 4: Process the first language features according to the first language model and give corresponding reminders to the operator. The first language model conducts machine learning training based on the historical voices of the outbound call objects to obtain key dialect words that can be predicted. During the subsequent call process, the system prompts the operator to communicate in the same dialect. When the operator uses the same dialect as the outbound call object, it can conduct high-quality outbound calls with a stronger sense of intimacy.

[0006] Preferably, in the second step, it is possible to judge the dialect, geographical range, and age range of the called object, and perform further operations based on the judgment results. Judge the dialect, geographical range, and age range used by the outbound call object according to the criteria in the first and second steps, and then, based on the judgment results, enable the operator to communicate in the same dialect, thereby enriching the communication effect and improving the communication quality. Moreover, compare the voice of the current outbound call object with the previous volume, the operator's speaking speed and volume during the communication in this call, evaluate the emotional changes of the outbound call object and the operator, the system evaluates the volume changes in the sentences of the outbound call object before and after, the operator's speaking speed and volume, identify the progress effect of the call, and the emotional state changes of the outbound call object and the operator can better evaluate the emotional state and communication quality of the called object during the process when the operator communicates with the called object in the same dialect. In the next call, if the volume of the called object reaches an intense emotional level compared to the volume in the previous call, it can be judged that a sensitive topic or taboo word of the outbound call object has emerged in the current communication.

[0007] Preferably, in the second step, it is possible to identify the density of business keywords in the first call and evaluate the quality of the outbound call object. By identifying the density of business keywords in the first call, the closeness between the call of the outbound call object and the business can be obtained, thereby evaluating the quality of the outbound call object.

[0008] Preferably, in the fourth step, process the above data according to the first language model to obtain the dialect communication types of the outbound call objects, generate the same dialect version of the text of the conversation to be communicated and present it to the operator, or the system matches the operator with the same dialect in the preset operator information and performs matching transfer processing, or actively asks the outbound call object whether they can communicate in Mandarin. Collect the native dialect materials of multiple operators in advance and form a database. After an outbound call object with a high difficulty in dialect imitation appears, the system automatically matches according to the materials stored in the database and the previous operator communicates with the called object to ask. After the called object agrees, transfer to an operator with the same specific dialect.

[0009] Preferably, when it is detected that the called party is in a normal mood, the operator asks the called party whether he can communicate in Mandarin. If the called party agrees, the operator directly communicates with the called party in Mandarin to ensure smooth communication.

[0010] Preferably, the second call is a call of the operator, and the second call is significantly different from the first call. Therefore, when the second call is detected to be abnormal, it is marked as an abnormal type and processed.

[0011] An outbound call management device includes a management system, wherein the management system includes a call record module, a first call analysis module, a second call analysis module, a first call processing module, a second call processing module, an operator database, a call transfer module, and a change feedback module; The call record module is synchronized with the call process and obtains the content of the outbound call, and marks the call of the outbound callee as the first call, and the call of the operator as the second call. The first call analysis module performs first real-time voice monitoring and analysis on the first call, and the first language features included in the first real-time voice include dialect features, volume, and business keywords. Among them, the dialect features are, for example, in terms of calling oneself "I", Northeastern people say "an", Shanghai people say "ala", Guangdong has the expression "from ao", Xi'an people simply call it "hungry", and Wenzhou people use "en" to express it; A second call analysis module performs second real-time voice monitoring and analysis on the second call; the second language features included in the second real-time voice include speech speed, volume, feedback keywords, and taboo words; The first call processing module includes a first language model, which uses the first language model to perform machine learning training based on the historical voice of the external call object to obtain key dialect words that can be predicted. The second call processing module includes a second language model, and the second language model handles the anomaly of the second call.

[0012] Preferably, the call transfer module is used in combination with the change feedback module. When the operator cannot understand the dialect of the called customer, the operator will give real-time feedback to the system through the change feedback module. Subsequently, the system will automatically match the operator with the same dialect characteristics as the called party in the database to switch the call. The management device includes an AI language module and a display module. The system converts the dialect communication content between the operator and the outbound caller into text in real time through the AI ​​language module and displays it to the operator through the display module.

[0013] An outbound call management storage medium stores computer program instructions. When the computer program instructions are executed by a processor, a call restriction method described in any of the outbound call management methods is implemented, and the storage medium includes a mass storage for data or instructions.

[0014] An outbound call management computer device includes a portable computer and a non-portable computer, such as a laptop computer and a desktop computer. The computer itself includes a processor, a communication interface, and a bus. The processor, the memory, and the communication interface are connected through the bus and communicate with each other to satisfy the operation of the computer program.

[0015] (III) Beneficial effects The present invention provides an outbound call management method, device, storage medium, and computer device, having the following beneficial effects: (1) In terms of the outbound call management method, device, storage medium, and computer device, in the outbound call management method, by predicting and planning the language characteristics and regional characteristics of the outbound call object, in the specific communication process, the operator can use a dialectal way to communicate with the outbound call object, reducing the resistance of the outbound call object to being questioned and enhancing the favorability of the outbound call object towards the operator.

[0016] (2) In terms of the outbound call management method, device, storage medium, and computer device, in the outbound call management device, through the set call transfer module and replacement feedback module, when the operator cannot understand the dialect of the called customer, the operator can use the replacement feedback module to provide real-time feedback to the system. Subsequently, the system will automatically match an operator with the same dialect characteristics as the outbound call object in the database for switched answering, fully ensuring a good communication effect with the outbound call object. (3) In terms of the outbound call management method, device, storage medium, and computer device, in the outbound call management device, through the set AI language module, the dialect communication content between the operator and the outbound call object is converted into real-time text and displayed to the operator through the display module, for the operator to confirm in real time that there is no disagreement in the communication intention between the two parties, avoiding communication misalignment in some sentences due to dialect pronunciation problems and fully ensuring the quality of outbound call communication.

[0017] (4) In terms of the storage medium and computer device of the outbound call management method, device, storage medium, and computer device, it can meet the usage requirements such as large storage and high-performance computing. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a schematic diagram of the principle of the outbound call management method of the present invention; Figure 2 It is a schematic diagram of the principle of the outbound call management device of the present invention. DETAILED DESCRIPTION

[0019] Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in the field without making any creative work shall fall within the scope of protection of the present invention.

[0020] This embodiment provides an outbound call management method The management prevention, the first step, is to make an outbound call to the target customer, obtain the outbound call content after success, and mark the call of the outbound call object as the first call, and the call of the operator as the second call. In this step, the call object and the called object are marked separately, that is, the call of the called object is marked as the first call, the call of the operator belonging to the calling object is marked as the second call, and the call of the automatic voice in the calling process is marked as the third call; During outbound calls, especially predictive outbound calls, the system automatically calls the outbound call recipient, answers the call, and automatically processes and transfers the call to the operator. Therefore, voices from different sources are distinguished and identified, which can better identify the dialect and regional characteristics of the called party, and the service status of the operator, so that corresponding measures can be taken in time.

[0021] The second step is to perform a first real-time voice monitoring on the first call. In this step, the first real-time voice monitoring is mainly used to monitor the first language features, which include dialect features, voice tones, and business keywords. For example, when addressing oneself, people in Northeast China say "an", people in Shanghai say "ala", people in Guangdong have the expression "from ao", people in Xi'an simply call it "hungry", people in Wenzhou use "en". Chaoshan dialect uses "wa"; In terms of voice pitch, those with tight vocal cords and high-pitched voices are generally young people, while those with loose vocal cords and low-pitched voices are generally elderly people. This embodiment uses the dialect, spoken language, and voice tone of the called party to evaluate whether the called party can speak Mandarin and where the called party is from. The emotion judgment of the feedback from the external call object includes first manually collecting the volume of different emotions and pre-setting a database, and using the database to analyze the volume difference and track line of the monitored sound in real time, so as to obtain the real-time emotion of the external call object. For example, during the call, the voice signal can be recorded in real time and then imported into the audio editing software, and the height of the waveform and the corresponding decibel scale are analyzed to judge the decibel size of the sound. Among them, the emotion judgment of the feedback from the external call object is that the range threshold of the volume difference representing normal emotions is 30 to 50 decibels, the range threshold of the volume difference representing depressed emotions is less than 30 decibels, the range threshold of the volume difference representing excited emotions is 50 to 80 decibels, and the range threshold of the volume difference representing intense emotions is greater than 80 decibels; The business keywords are determined according to the reason for the outbound call, so they vary according to the customer object. For example, for communication customers, the business keywords include the name of the debtor, the platform involved, "phone bills", "suspended service", "credit investigation" and other related words. Business keywords can be obtained from the statistics of multiple outbound call records in the same business, and the words with high frequency are the keywords; In the identification of business keywords, common words need to be manually eliminated to prevent them from being misidentified as business-related words.

[0022] The third step is to perform a second real-time voice monitoring on the second call.

[0023] In this step, the second real-time voice monitoring is mainly used to monitor the second language features; the second language features include speech speed, volume, feedback keywords, and taboo words; The range threshold of speech speed representing normal emotions is 170-400Hz, 160-280 words / minute, the range threshold of speech speed representing depressed emotions is less than 170-400Hz, 160-280 words / minute, and the range threshold of speech speed representing excited emotions is greater than 170-400Hz, 160-280 words / minute; the range threshold of volume difference representing normal emotions is 30-50 decibels, the range threshold of speech volume difference representing depressed emotions is less than 30 decibels, and the range threshold of volume difference representing excited emotions is greater than 50 decibels; Feedback keywords are based on the perspective of the person being called, and their dialect terms will change accordingly. For example, when the operator is communicating with a callee who speaks a different dialect, the same dialect communication words can be used as feedback keywords. On the other hand, the feedback keywords can also be used to confirm the current call status based on the emotions of the callee. For example, when the callee hears the same dialect feedback keywords, the emotions of the callee should be in the range of emotional excitement, that is, the volume difference range threshold is 50-80 decibels; The distinction between business keywords and feedback keywords is set according to the different monitored objects. This setting makes the monitoring of the call process more accurate and can make judgments and responses efficiently; Taboo words: After determining the regional characteristics based on the dialect characteristics of the called party determined in the second step, the system collects the taboo words recognized by the local people and converts them into text through the display for the operator to browse and self-prompt, so as to avoid the communication being terminated or the communication willingness of the called party being reduced due to the repeated appearance of taboo words; The fourth step is to process the first language features according to the first language model and remind the operator accordingly.

[0024] In this step, the first language model is trained by machine learning based on the historical speech of the outbound call object to obtain key dialect words that can be predicted. For example, if the dialect colloquialism "俺" appears in the first two minutes of communication, the system will prompt the operator to use the same dialect to communicate in the subsequent process of the call. When the operator uses the same dialect as the outbound call object for initial communication, if the volume of the outbound call object reaches the defined range of emotional excitement, it also shows that the operator's communication dialect is correct and the dialect should continue to be used for communication. Through dialect-based mutual communication, the resistance of the outbound call object to being questioned can be reduced, and the outbound call object's favorability towards the operator can be improved; Determine the dialect, region, and age range of the person being called, and perform further operations based on the determination results.

[0025] In this step, the dialect, geographical range, and age range used by the outbound call recipient are judged according to the standards in the first and second steps, and then based on the judgment results, the operator communicates in the same dialect, thereby enriching the communication effect and improving the quality of communication.

[0026] The voice of the outbound call recipient is compared with the previous volume during the call, the speaking speed and volume of the operator during the communication process, and the emotional changes of the outbound call recipient and the emotional changes of the operator are evaluated.

[0027] In this step, the system monitors the volume changes in the caller's sentences before and after the call, the operator's speaking speed and volume, and identifies the progress of the call. The emotional state changes of the caller and the operator can better evaluate the emotional state and communication quality of the callee when the operator uses the same dialect as the callee. If the volume of the called party in the next call is more emotional than that in the previous call, it can be judged that sensitive topics or taboo words of the outgoing caller have appeared in the current communication; Using the above examples, it is possible to judge the preference characteristics of the external call object, such as which factors are liked and which factors are disliked, and then use them as communication suggestions for the operator to improve the effect of external calls; Identify the business keyword density in the first call and evaluate the quality of the external call object. By identifying the business keyword density in the first call, the closeness between the call of the external call object and the business can be obtained, so as to evaluate the quality of the external call object. For example, if the average number of occurrences of business keywords within a 1-minute length in the first call is 8 times, 4 times, 2 times, and 0 times respectively, the called objects are divided into 5 levels. The more occurrences there are later, the higher the quality of the external call object; Process the above data according to the first language model to obtain the dialect communication types of the external call object, generate the same dialect version of the conversation text to be communicated and present it to the operator, so that the operator can communicate in dialect, or the system matches the operator with the same dialect in the preset operator information and performs matching transfer processing, or actively asks the external call object whether they can communicate in Mandarin; In this step, the conversation text generated according to the dialect characteristics of the external call object should be simple and clear and presented to the operator. The data presented to the operator is only helpful for the communication content between the operator and the external call object, and the evaluation of the quality of the external call object is not presented to the operator during the communication process to avoid affecting the operator's handling of the call; The own dialect information of multiple operators is collected in advance and formed into a database. After an external call object with a high difficulty in dialect imitation appears, the system automatically matches according to the information stored in the database and the previous operator communicates with the called object to ask. After the called object agrees, transfer to the operator with the same specific dialect. It is also possible to detect that the called object is in a normal mood in the second step, and the operator asks the called object whether they can communicate in Mandarin. If the called object agrees, the operator directly communicates with the called object in Mandarin to ensure smooth communication; The second call is the call of the operator. Since the operators are all trained and use fixed conversation texts, the second call is significantly different from the first call. Therefore, when an abnormality in the second call is detected, it is processed according to the type of abnormality. For example, if the operator's voice is too loud, prompt them to lower the volume; if the operator's voice is too small, prompt them to increase the volume. Another example is that when the operator says taboo words recognized by the geographical characteristics of the called object in the process of communication with the called object, the system will promptly prompt the operator to apologize to the called object actively, so as to ensure the smooth progress of subsequent communication.

[0028] The external call management device provided by this example: Including a management system, the management system includes a call record module, a first call analysis module, a second call analysis module, a first call processing module, a second call processing module, an operator database, a call transfer module, and a replacement feedback module The call record module is synchronized with the call process and obtains the content of the outbound call, and marks the call of the outbound caller as the first call and the call of the operator as the second call, thereby facilitating the analysis of different objects, and the call record module also marks the call of the automatic voice as the third call; a first call analysis module, performing first real-time voice monitoring and analysis on the first call, wherein the first language features included in the first real-time voice include dialect features, volume, and business keywords, wherein the dialect features include, for example, in terms of calling oneself "I", people in Northeast China say "an", people in Shanghai say "ala", people in Guangdong have the expression "from ao", people in Xi'an simply call it "hungry", people in Wenzhou use "en", and Chaoshan dialect uses "wa"; The second call analysis module performs a second real-time voice monitoring and analysis on the second call; the second language features included in the second real-time voice include speech speed, volume, feedback keywords, and taboo words. The first call processing module includes a first language model, which uses the first language model to perform machine learning training based on the historical voice of the external call object to obtain key dialect words that can be predicted. For example, if the dialect colloquialism "俺" appears within the first two minutes of the communication, the system will prompt the operator to use the same dialect to communicate in the subsequent process of the call; The second call processing module includes a second language model, and the second language model processes the abnormality of the second call, for example, if the operator's voice is too loud, the operator is prompted to lower the volume; if the operator's voice is too soft, the operator is prompted to increase the volume. For example, when the operator and the called party are communicating, if the operator says a taboo word recognized by the called party in multiple regional characteristics, the system will promptly prompt the operator to take the initiative to apologize to the called party; The operator database collects the dialect features of multiple operators in advance and forms a database, which makes it easier for the system to find suitable operators during real-time calls. The call transfer module and the change feedback module provide real-time feedback to the system through the change feedback module when the operator cannot understand the dialect of the called customer. Subsequently, the system automatically matches the operator with the same dialect features as the called party in the database to switch the call. The device includes an AI language module and a display module: The system performs real-time text conversion on the dialect communication content between the operator and the called party through the AI language module and displays it to the operator through the display module, so that the operator can confirm in real time that there are no differences in the communication intentions of both parties, avoid communication misalignments in some sentences caused by dialect pronunciation problems, and fully ensure the quality of outbound calls; The outbound call management storage medium provided in this example: The storage medium stores computer program instructions, which, when executed by a processor, implement the call restriction method described in any one of the outbound call management methods. The storage medium includes a mass storage device for data or instructions, such as a hard disk drive, a floppy disk drive, a flash memory, an optical disc, a magneto-optical disc, a magnetic tape, or a universal serial bus drive, or a combination of two or more; The memory includes a medium that can be inside or outside the integrated gateway disaster recovery device and can be removed or not removed (or fixed). The memory is a non-volatile solid-state memory; The outbound call management computer device provided in this example: It includes a portable computer, such as a laptop, and a non-portable computer, such as a desktop computer. The computer itself includes a processor, a communication interface, and a bus. Among them, the bus includes hardware, software, or both, and couples the components of the online data flow metering device together. The processor, the memory, and the communication interface are connected through the bus and complete communication with each other to meet the operation of the computer program.

[0029] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device.

[0030] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions, and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. An outbound call management method, characterized in that: include: first step: Make an outbound call to the target customer, obtain the outbound call content after success, and mark the call of the outbound caller as the first call, and the call of the operator as the second call; Step 2: Performing first real-time voice monitoring on the first call, where the first real-time voice monitoring is mainly used to monitor first language features, where the first language features include dialect, voice tone, volume, and business keywords; Step 3: Performing a second real-time voice monitoring on the second call, where the second real-time voice monitoring is mainly used to monitor second language features; the second language features include speech speed, volume, feedback keywords, and taboo words; Step 4: The first language features are processed according to the first language model, and the operator is reminded accordingly. The first language model is trained through machine learning based on the historical voice of the outbound call object to obtain key dialect words that can be predicted. In the subsequent call process, the system prompts the operator to communicate in the same dialect. When the operator uses the same dialect as the outbound call object for high-quality outbound communication, a stronger sense of intimacy is achieved.

2. The outbound call management method according to claim 1, characterized in that: In the second step, the dialect, geographical range, and age range of the called party can be determined, and further operations can be performed based on the determination results. The voice of the current outbound caller can be compared with the previous volume in the current call, and the speaking speed and volume of the operator during the communication process can be compared to evaluate the emotional changes of the outbound caller and the operator. If, in the next call, the volume of the called party reaches an emotionally intense level compared to the volume in the previous call, it can be determined that sensitive topics or taboo words of the outbound caller have appeared in the current communication.

3. The outbound call management method according to claim 1, characterized in that: In the second step, the density of business keywords in the first call can be identified, and the quality of the outbound call object can be evaluated. By identifying the density of business keywords in the first call, the closeness between the call and the business of the outbound call object can be obtained, thereby evaluating the quality of the outbound call object.

4. The outbound call management method according to claim 1, characterized in that: The fourth step is to process the data according to the first language model to obtain the dialect communication type of the outbound call object, generate the text of the to-be-communicated speech into the same dialect version and present it to the operator, or the system matches the operator with the same dialect in the preset operator information and performs matching and transfer processing, or actively asks the outbound call object whether it can communicate in Mandarin; The dialect data of multiple operators are collected in advance and formed into a database. When an external caller appears whose dialect is difficult to imitate, the system automatically matches the caller based on the data stored in the database and the previous operator communicates and inquires with the called party. If the called party agrees, the call is transferred to an operator with the same specific dialect.

5. An outbound call management method according to claim 4, characterized in that: The sound volumes of different emotions are collected and a database is pre-set. The volume difference and audio track of the monitored sound are analyzed in real time through the database, so as to obtain the real-time emotion of the outbound caller. When it is detected that the called person is in a normal mood, the operator will ask the called person whether he can communicate in Mandarin. If the called person agrees, the operator will directly communicate with the called person in Mandarin to ensure smooth communication.

6. An outbound call management method according to claim 5, characterized in that: The second call is a call of the operator, and the second call is obviously different from the first call. Therefore, when the second call is detected to be abnormal, it is marked as an abnormal type and processed.

7. An outbound call management device, characterized in that: It includes a management system, which includes a call record module, a first call analysis module, a second call analysis module, a first call processing module, a second call processing module, an operator database, a call transfer module, and a change feedback module; The call recording module is synchronized with the call process and obtains the content of the outbound call, and marks the call of the outbound callee as the first call, and the call of the operator as the second call. The first call analysis module performs first real-time voice monitoring and analysis on the first call, and the first language features included in the first real-time voice include dialect features, volume, and business keywords; A second call analysis module performs second real-time voice monitoring and analysis on the second call; the second language features included in the second real-time voice include speech speed, volume, feedback keywords, and taboo words; The first call processing module includes a first language model, which uses the first language model to perform machine learning training based on the historical voice of the external call object to obtain key dialect words that can be predicted. The second call processing module includes a second language model, and the second language model handles the anomaly of the second call.

8. An outbound call management device according to claim 7, characterized in that: The call transfer module is used in combination with the change feedback module. When the operator cannot understand the dialect of the called customer, the operator will provide real-time feedback to the system through the change feedback module. Subsequently, the system will automatically match an operator with the same dialect characteristics as the called party in the database to switch the call. The management device includes an AI language module and a display module. The system converts the dialect communication content between the operator and the outbound caller into text in real time through the AI ​​language module and displays it to the operator through the display module.

9. An outbound call management storage medium, characterized in that: The storage medium stores computer program instructions, which, when executed by a processor, implement the call restriction method according to any one of claims 1 to 6, and the storage medium includes a large-capacity memory for data or instructions.

10. An outbound call management computer device, characterized in that: The computer includes portable computers and non-portable computers, and the computer itself includes a processor, a communication interface, and a bus.

Images