Customer label mining method and device based on dialogue scene, equipment and medium
By using dialogue records and tag mining models in the outgoing call management system, we can identify and label customers' business scenarios, problems, processes and attitudes in real time, and solve the problem of insufficient singleness and accuracy of customer tags in the existing technology, and improve the outgoing call efficiency and business benefits.
Patent Information
- Application Number
- CN202510159054.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-02-13
- Publication Date
- 2025-06-03
AI Technical Summary
In the existing outbound call management system, customer tags are usually determined based on registration information, resulting in insufficient accuracy of outbound call and low efficiency and business benefits.
By obtaining the current conversation record and entering a pre-established tag mining model when talking to the target customer, identifying business scenario tags, customer problem tags, business process tags and customer attitude tags, and digging more customer tags in real time and accurately.
It realizes real-time understanding and labeling of key information in the dialogue records during the dialogue process, improves the accuracy and richness of customer tags, and thus improves outbound call efficiency and business benefits.
Smart Images
Figure CN120086336A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of artificial intelligence technology. Specifically, it relates to a method, apparatus, device, and medium for mining customer tags based on a dialogue scenario. Background Art
[0002] With the continuous development of artificial intelligence technology, more and more enterprises have begun to apply an outbound management system to call customers, conduct voice calls with customers, and promote and sell enterprise services to customers, such as financial insurance services.
[0003] The existing outbound management system supports selecting target customers with specific customer tags from the outbound list for precise calling. However, in fact, the customer tags of each customer in the outbound list are usually determined only based on the registration information of each customer on the enterprise service platform. For example, the customer tag of customer A is the region where customer A is located, and the tag content is relatively single, which still limits the outbound precision to a certain extent, resulting in low outbound efficiency and business benefits. Summary of the Invention
[0004] The purpose of the embodiments of this application is to provide a method, apparatus, device, and medium for mining customer tags based on a dialogue scenario, so as to achieve the technical effect of accurately mining more customer tags in real time during the dialogue with customers.
[0005] In a first aspect, the embodiments of this application provide a method for mining customer tags based on a dialogue scenario, including:
[0006] When having a dialogue with a target customer, obtain the current dialogue record of the target customer;
[0007] Input the current dialogue record into a pre-established tag mining model to obtain potential tags of the target customer; wherein, the tag mining model is used to identify one or more of business scenario tags, customer problem tags, business process tags, and customer attitude tags from the current dialogue record as the potential tags;
[0008] Determine that the customer tags of the target customer include the potential tags.
[0009] In the above implementation process, in the case of having a conversation with the target customer, the current conversation record with the target customer is input into a pre-established label mining model, enabling the label mining model to identify one or more of the business scenario labels, customer problem labels, business process labels, and customer attitude labels from the current conversation record as potential labels, obtaining the potential labels of the target customer, and determining that the customer labels of the target customer include the potential labels, which can continuously understand and label potential key information such as business scenarios, customer problems, business processes, and customer attitudes in the conversation record during the conversation with the customer, thereby accurately and in real-time mining more customer labels.
[0010] Further, the obtaining of the current conversation record with the target customer includes:
[0011] Obtaining the current round of question-and-answer statements sent by the target customer and the historical conversation record with the target customer;
[0012] Combining the current round of question-and-answer statements and the historical conversation record to obtain the current conversation record.
[0013] In the above implementation process, by combining the current round of question-and-answer statements sent by the target customer and the historical conversation record with the target customer in each round of question-and-answer during the conversation with the target customer to obtain the current conversation record with the target customer, it can ensure the real-time and accurate acquisition of the current conversation record with the target customer.
[0014] Further, the label mining model includes one or more of a business scenario label mining model for identifying the business scenario labels from the current conversation record, a customer problem label mining model for identifying the customer problem labels from the current conversation record, a business process label mining model for identifying the business process labels from the current conversation record, and a customer attitude label mining model for identifying the customer attitude labels from the current conversation record.
[0015] In the above implementation process, by pre-establishing one or more of a business scenario label mining model for identifying business scenario labels from the current conversation record, a customer problem label mining model for identifying customer problem labels from the current conversation record, a business process label mining model for identifying business process labels from the current conversation record, and a customer attitude label mining model for identifying customer attitude labels from the current conversation record as the label mining model, it can use various label mining models in the label mining model to specifically handle their respective label identification tasks, effectively improving the identification accuracy of the label mining model for various potential labels.
[0016] Further, the label mining model includes the business scenario label mining model and the customer problem label mining model;
[0017] Inputting the current conversation record into a pre - established tag mining model to obtain potential tags of the target customer, including:
[0018] Inputting the current conversation record into the business scenario tag mining model to obtain the business scenario tags;
[0019] Inputting the current conversation record and the business scenario tags into the customer problem tag mining model to obtain the customer problem tags.
[0020] In the above implementation process, on the premise that the tag mining model includes a business scenario tag mining model and a customer problem tag mining model, first input the current conversation record of the target customer into the business scenario tag mining model to obtain the business scenario tags of the target customer, and then input the current conversation record of the target customer and the business scenario tags of the target customer into the customer problem tag mining model to obtain the customer problem tags of the target customer, which can take into account the correlation between customer problems and business scenarios and more accurately mine customer problem tags.
[0021] Further, the tag mining model includes the business scenario tag mining model and the business process tag mining model;
[0022] Inputting the current conversation record into a pre - established tag mining model to obtain potential tags of the target customer, including:
[0023] Inputting the current conversation record into the business scenario tag mining model to obtain the business scenario tags;
[0024] Inputting the current conversation record and the business scenario tags into the business process tag mining model to obtain the business process tags.
[0025] In the above implementation process, on the premise that the tag mining model includes a business scenario tag mining model and a business process tag mining model, first input the current conversation record of the target customer into the business scenario tag mining model to obtain the business scenario tags of the target customer, and then input the current conversation record of the target customer and the business scenario tags of the target customer into the business process tag mining model to obtain the business process tags of the target customer, which can take into account the correlation between business processes and business scenarios and more accurately mine business process tags.
[0026] Further, the method further includes:
[0027] When the business process label is the target business process label, optimize the business process corresponding to the business process label; wherein, the target business process label includes the business process label identified from the historical conversation record with the target customer.
[0028] In the above implementation process, by optimizing the business process corresponding to the business process label when the business process label is the target business process label such as the business process label identified from the historical conversation record with the target customer, it is possible to timely discover the bottlenecks in the business process for optimization and improve the business sales efficiency.
[0029] Further, the potential label includes the customer attitude label;
[0030] The method further includes:
[0031] When the customer attitude label is a negative attitude label, send a target reply statement to the target customer; and / or,
[0032] End the conversation with the target customer.
[0033] In the above implementation process, by sending a target reply statement to the target customer when the customer attitude label of the target customer is a negative attitude label, and / or ending the conversation with the target customer, it is possible to timely perceive the negative attitude of the customer for appeasement and avoid customer complaints.
[0034] In a second aspect, an embodiment of the present application provides a customer label mining device based on a conversation scenario, including:
[0035] A conversation record acquisition module, configured to acquire the current conversation record with the target customer when having a conversation with the target customer;
[0036] A potential label mining module, configured to input the current conversation record into a pre-established label mining model to obtain the potential labels of the target customer; wherein, the label mining model is used to identify one or more of business scenario labels, customer problem labels, sales process labels, and customer attitude labels from the current conversation record as the potential labels;
[0037] A customer label determination module, configured to determine that the customer label of the target customer includes the potential labels.
[0038] In a third aspect, an embodiment of the present application provides an electronic device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor; when the processor executes the computer program, the method described above is implemented.
[0039] Fourthly, an embodiment of the present application provides a computer-readable storage medium, which includes a stored computer program; wherein, when the computer program runs, it controls the device where the computer-readable storage medium is located to execute the method described above. Description of the Drawings
[0040] To more clearly illustrate the technical solutions of the embodiments of the present application, the following will briefly introduce the drawings required to be used in the embodiments of the present application. It should be understood that the following drawings only show some embodiments of the present application, and therefore should not be regarded as a limitation of the scope. For those of ordinary skill in the art, without creative efforts, other related drawings can also be obtained based on these drawings.
[0041] Figure 1 It is a schematic flowchart of a method for mining customer tags based on a dialogue scenario provided by the first embodiment of the present application;
[0042] Figure 2 It is a schematic structural diagram of a device for mining customer tags based on a dialogue scenario provided by the second embodiment of the present application;
[0043] Figure 3 It is a schematic structural diagram of an electronic device provided by the third embodiment of the present application. Detailed Embodiments
[0044] The following will describe the technical solutions in the embodiments of the present application in conjunction with the drawings in the embodiments of the present application.
[0045] It should be noted that: in the description of the present application, terms such as "first" and "second" are only used for distinguishing descriptions and cannot be understood as indicating or implying relative importance. At the same time, the step numbers in the text are only for the convenience of explaining the embodiments of the present application and do not serve as a function of limiting the execution order of the steps.
[0046] With the continuous development of artificial intelligence technology, more and more enterprises have begun to use an outbound management system to call customers, make voice calls with customers, and promote and sell enterprise business to customers, such as financial insurance business.
[0047] In the related art, the existing outbound management system supports selecting target customers with specific customer tags from the outbound list for accurate calling. However, in fact, the customer tags of each customer in the outbound list are usually determined only based on the registration information of each customer on the enterprise business platform. For example, the customer tag of customer A is the area where customer A is located, and the tag content is relatively single, which still limits the outbound accuracy to a certain extent, resulting in low outbound efficiency and business benefits.
[0048] To this end, the present application proposes a method for mining customer tags based on a dialogue scenario. When having a dialogue with a target customer, the current dialogue record with the target customer is input into a pre-established tag mining model. The tag mining model identifies one or more of business scenario tags, customer problem tags, business process tags, and customer attitude tags from the current dialogue record as potential tags, obtaining the potential tags of the target customer, so as to determine that the customer tags of the target customer include the potential tags, and can continuously understand and label potential key information such as business scenarios, customer problems, business processes, and customer attitudes in the dialogue record during the dialogue with the customer, thereby accurately mining more customer tags in real time.
[0049] Next, the technical solutions in the embodiments of the present application will be described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments.
[0050] The method provided by the embodiments of the present application can be executed by relevant terminal devices, and hereinafter, a terminal device installed with an outbound call management system, such as a user terminal, is taken as an example of the execution subject for description.
[0051] It should be noted that the user terminal includes terminal devices such as mobile phones, tablets, or computers held by enterprise staff.
[0052] Please refer to Figure 1 , Figure 1 which is a schematic flowchart of a method for mining customer tags based on a dialogue scenario provided by the first embodiment of the present application. The first embodiment of the present application provides a method for mining customer tags based on a dialogue scenario, including steps S101 to S103:
[0053] S101. When having a dialogue with a target customer, obtain the current dialogue record with the target customer.
[0054] As an example, when enterprise staff needs to conduct business promotion, according to the actual business promotion requirements, obtain the outbound call list of the target business, input the outbound call list into the user terminal. There are multiple customers in the outbound call list, and each customer carries their own basic tags, such as customer gender tags, customer age tags, and customer region tags, etc., so that the user terminal can select a target customer carrying the basic tags specified by the enterprise staff from the outbound call list for calling and have a dialogue with the target customer.
[0055] The user terminal obtains the current dialogue record with the target customer when having a dialogue with the target customer.
[0056] In practical applications, the mining period can be set in advance. During the process of the user terminal having a conversation with the target customer, when the mining period arrives, the current conversation record with the target customer is obtained.
[0057] S102: Input the current conversation record into a pre-established label mining model to obtain the potential labels of the target customer; among them, the label mining model is used to identify one or more of the business scenario label, customer problem label, business process label, and customer attitude label from the current conversation record as potential labels.
[0058] Exemplarily, a label mining model is pre-established. The label mining model is used to identify one or more of the business scenario label, customer problem label, business process label, and customer attitude label from the current conversation record with the target customer as the potential labels of the target customer.
[0059] In practical applications, the label mining model can be established by training a large language model.
[0060] A large language model (LLM) refers to a deep learning model trained with a large amount of text data, which can generate natural language text or understand the meaning of natural language text. The large language model can handle various natural language tasks, such as text classification, question answering, dialogue, etc.
[0061] For example, collect a large number of human-machine conversation records, each of which is labeled with one or more of the business scenario label, customer problem label, business process label, and customer attitude label. Input these human-machine conversation records into the large language model, and adjust the parameters of the large language model with the minimization of the loss function as the optimization goal to obtain the label mining model. Among them, for classification problems, the cross-entropy loss function can be selected as the loss function.
[0062] After the user terminal obtains the current conversation record with the target customer, it inputs the current conversation record into the pre-established label mining model, enabling the label mining model to identify one or more of the business scenario label, customer problem label, business process label, and customer attitude label from the current conversation record as potential labels, and obtaining the potential labels of the target customer.
[0063] It should be noted that the business scenario tags identified from the current conversation record are used to indicate the current business scenario, such as the scenario of promoting a certain business to a certain group of people, such as the scenario of promoting financial insurance business to students. The customer problem tags identified from the current conversation record are used to indicate the type and content of the problems currently raised by the target customer, such as the business cost problem of "what discounts are available for students to purchase financial insurance business". The business process tags identified from the current conversation record are used to indicate the type and nodes of the current business process, such as the business introduction node in the financial insurance business sales process. The customer attitude tags identified from the current conversation record are used to indicate the attitude currently shown by the target customer, such as a positive or negative attitude.
[0064] S103. Determine that the customer tags of the target customer include potential tags.
[0065] Exemplarily, after obtaining the potential tags of the target customer, the user terminal determines that the customer tags of the target customer include potential tags, so that at this time, the customer tags of the target customer include basic tags such as customer gender tags, customer age tags, and customer region tags, as well as one or more types of potential tags among business scenario tags, customer problem tags, business process tags, and customer attitude tags.
[0066] In practical applications, since the current conversation records obtained by the user terminal from the target customer in different mining cycles may not be exactly the same, the potential tags of the target customer obtained by the user terminal in different mining cycles may also not be exactly the same. After obtaining the potential tags of the target customer in each mining cycle, the user terminal can only retain the latest obtained various potential tags of the target customer in the customer tags of the target customer. For example, assume that the potential tags of the target customer obtained by the user terminal in a certain mining cycle include business scenario tag a and customer problem tag b1, and the potential tags of the target customer obtained by the user terminal in the next mining cycle of this mining cycle include customer problem tag b2, business process tag c, and customer attitude tag d. Then, the customer tags determined by the user terminal for the target customer include basic tags such as customer gender tags, customer age tags, and customer region tags, as well as business scenario tag a, customer problem tag b2, business process tag c, and customer attitude tag d.
[0067] In the embodiment of the present application, when having a conversation with a target customer, the current conversation record with the target customer is input into a pre-established label mining model, so that the label mining model identifies one or more of business scenario labels, customer problem labels, business process labels, and customer attitude labels from the current conversation record as potential labels, obtaining the potential labels of the target customer, and determining that the customer labels of the target customer include the potential labels, which can continuously understand and label potential key information such as business scenarios, customer problems, business processes, and customer attitudes in the conversation record during the conversation with the customer, thereby accurately mining more customer labels in real time.
[0068] In an alternative embodiment, the obtaining of the current conversation record with the target customer includes: obtaining the current round of question-and-answer statements sent by the target customer, and the historical conversation record with the target customer; combining the current round of question-and-answer statements and the historical conversation record to obtain the current conversation record.
[0069] Exemplarily, during the process of the user terminal having a conversation with the target customer, it usually conducts one or more rounds of question-and-answer with the target customer. During the process of the user terminal having a conversation with the target customer, once it detects that the target customer sends the current round of question-and-answer statements, it obtains the current round of question-and-answer statements sent by the target customer, and the historical conversation record with the target customer, and combines the current round of question-and-answer statements and the historical conversation record to obtain the current conversation record with the target customer.
[0070] In practical applications, a prompt template can be set in advance, and the user terminal writes the current round of question-and-answer statements sent by the target customer, and the historical conversation record with the target customer into the prompt template to obtain the current conversation record with the target customer.
[0071] It should be noted that the prompt template is a text template used to guide an artificial intelligence system to complete a specific task or provide information.
[0072] In the embodiment of the present application, by combining the current round of question-and-answer statements sent by the target customer and the historical conversation record with the target customer during each round of question-and-answer in the conversation with the target customer, the current conversation record with the target customer is obtained, which can ensure the real-time and accurate acquisition of the current conversation record with the target customer.
[0073] In an alternative embodiment, the label mining model includes one or more of a business scenario label mining model for identifying business scenario labels from the current conversation record, a customer problem label mining model for identifying customer problem labels from the current conversation record, a business process label mining model for identifying business process labels from the current conversation record, and a customer attitude label mining model for identifying customer attitude labels from the current conversation record.
[0074] Exemplarily, in order to improve the recognition accuracy of the label mining model for various potential labels, a business scenario label mining model for identifying business scenario labels from the current conversation record can be established in advance for business scenario labels, a customer problem label mining model for identifying customer problem labels from the current conversation record can be established in advance for customer problem labels, a business process label mining model for identifying business process labels from the current conversation record can be established in advance for business process labels, and a customer attitude label mining model for identifying customer attitude labels from the current conversation record can be established in advance for customer attitude labels, so that the label mining model includes one or more of the business scenario label mining model, the customer problem label mining model, the business process label mining model, and the customer attitude label mining model.
[0075] In practical applications, any one of the business scenario label mining model, the customer problem label mining model, the business process label mining model, and the customer attitude label mining model can be established by training a large language model.
[0076] In the embodiment of the present application, by establishing in advance one or more of a business scenario label mining model for identifying business scenario labels from the current conversation record, a customer problem label mining model for identifying customer problem labels from the current conversation record, a business process label mining model for identifying business process labels from the current conversation record, and a customer attitude label mining model for identifying customer attitude labels from the current conversation record as the label mining model, various label mining models in the label mining model can be used to specifically handle their respective label recognition tasks, effectively improving the recognition accuracy of the label mining model for various potential labels.
[0077] In an optional embodiment, the label mining model includes a business scenario label mining model and a customer problem label mining model; the inputting the current conversation record into the pre-established label mining model to obtain potential labels of the target customer includes: inputting the current conversation record into the business scenario label mining model to obtain business scenario labels; and inputting the current conversation record and the business scenario labels into the customer problem label mining model to obtain customer problem labels.
[0078] Exemplarily, after the user terminal obtains the current conversation record with the target customer, since the label mining model includes a business scenario label mining model and a customer problem label mining model, the current conversation record is first input into the business scenario label mining model to enable the business scenario label mining model to understand and label the business scenario in the current conversation record, identify the business scenario label from the current conversation record, obtain the business scenario label of the target customer, and then input the current conversation record and the business scenario label of the target customer into the customer problem label mining model to enable the customer problem label mining model to determine the business scenario indicated by the business scenario label of the target customer, understand and label the customer problems related to the business scenario in the current conversation record, identify the customer problem label from the current conversation record, and obtain the customer problem label of the target customer.
[0079] For example, assume that the label mining model includes a business scenario label mining model and a customer problem label mining model, and the current conversation record obtained by the user terminal with the target customer is as follows:
[0080] Enterprise staff: Hello, this call is to recommend an electronic policy for supplementary medical insurance to you. After the medical insurance and social insurance reimbursement, the remaining part can be reimbursed for you again. Can I explain it to you in 2 minutes?
[0081] Target customer: I'm a student and I have no money.
[0082] Enterprise staff: Don't reject me in a hurry. The minimum monthly premium for this policy is 6 yuan, and the coverage is very comprehensive, covering all major and minor diseases and accidents, and it can reimburse imported drugs and self-paid drugs. Can you please learn about our product first?
[0083] Target customer: Are there any discounts?
[0084] Then the user terminal first inputs the current conversation record into the business scenario label mining model to obtain the business scenario label of the target customer. At this time, the business scenario label of the target customer is used to indicate the scenario of currently promoting the supplementary medical insurance policy to students. Then, the current conversation record and the business scenario label of the target customer are input into the customer problem label mining model to obtain the customer problem label of the target customer. At this time, the customer problem label of the target customer is used to indicate the business expense-related problem of "what discounts are there for students to purchase the supplementary medical insurance policy" currently raised by the target customer.
[0085] In the embodiment of the present application, on the premise that the label mining model includes a business scenario label mining model and a customer problem label mining model, first, the current conversation record with the target customer is input into the business scenario label mining model to obtain the business scenario label of the target customer, and then the current conversation record with the target customer and the business scenario label of the target customer are input into the customer problem label mining model to obtain the customer problem label of the target customer, which can take into account the correlation between the customer problem and the business scenario and more accurately mine the customer problem label.
[0086] In an alternative embodiment, the label mining model includes a business scenario label mining model and a business process label mining model; the step of inputting the current conversation record into the pre-established label mining model to obtain the potential label of the target customer includes: inputting the current conversation record into the business scenario label mining model to obtain the business scenario label; inputting the current conversation record and the business scenario label into the business process label mining model to obtain the business process label.
[0087] Exemplarily, after the user terminal obtains the current conversation record with the target customer, since the label mining model includes a business scenario label mining model and a business process label mining model, the current conversation record is first input into the business scenario label mining model to enable the business scenario label mining model to understand and label the business scenario in the current conversation record, identify the business scenario label from the current conversation record to obtain the business scenario label of the target customer, and then the current conversation record and the business scenario label of the target customer are input into the business process label mining model to enable the business process label mining model to determine the business scenario indicated by the business scenario label of the target customer, understand and label the business process related to the business scenario in the current conversation record, identify the business process label from the current conversation record to obtain the business process label of the target customer.
[0088] For example, assume that the label mining model includes a business scenario label mining model and a business process label mining model, and the current conversation record with the target customer obtained by the user terminal is:
[0089] "Enterprise staff: Hello, the call is to recommend an e-policy for supplementary medical insurance to you. After the medical insurance and social insurance are reimbursed, the remaining part can be reimbursed for you again. Can I give you a 2-minute explanation? ......
[0091] Enterprise staff: I have sent you a link to the product introduction of the supplementary medical insurance policy. You can click on the link to view the detailed information."
[0092] The user terminal first inputs the current conversation record into the business scenario label mining model to obtain the business scenario label of the target customer. At this time, the business scenario label of the target customer is used to indicate the scenario of selling supplementary medical insurance policies to students currently. Then, the current conversation record and the business scenario label of the target customer are input into the business process label mining model to obtain the business process label of the target customer. At this time, the business process label of the target customer is used to indicate the product introduction node in the supplementary medical insurance policy sales process currently.
[0093] In the embodiment of the present application, on the premise that the label mining model includes a business scenario label mining model and a business process label mining model, the current conversation record with the target customer is first input into the business scenario label mining model to obtain the business scenario label of the target customer. Then, the current conversation record with the target customer and the business scenario label of the target customer are input into the business process label mining model to obtain the business process label of the target customer, which can take into account the correlation between the business process and the business scenario and more accurately mine the business process label.
[0094] In an optional embodiment, the method further includes step S104:
[0095] S104. When the business process label is the target business process label, optimize the business process corresponding to the business process label; wherein, the target business process label includes the business process label identified from the historical conversation record with the target customer.
[0096] As an example, during the conversation between the user terminal and the customer, if the business process labels of the target customer obtained in multiple consecutive mining cycles are exactly the same, it means that it is continuously stuck at the business process node indicated by this business process label.
[0097] In order to timely discover the bottlenecks in the business process for optimization and improve the business sales efficiency, the target business process label can be preset, where the target business process label includes the business process label identified from the historical conversation record with the target customer, such as the business process label of the target customer obtained in the previous mining cycle of the current mining cycle.
[0098] After the user terminal obtains the business process label of the target customer in the current mining cycle, it determines whether this business process label is the target business process label. If this business process label is the target business process label, it is considered that the business process node indicated by this business process label is a bottleneck, and at this time, the business process corresponding to the business process label is optimized.
[0099] For example, assume that the business process tag indicates that the user terminal is currently at the product introduction node in the supplementary medical insurance policy sales process. In this case, the user terminal will optimize the supplementary medical insurance policy sales process. The optimization content includes re-obtaining the product introduction link of the supplementary medical insurance policy and sending the new product introduction link to the target customer.
[0100] In the embodiment of the present application, when the business process tag is a target business process tag such as a business process tag identified from the historical conversation record with the target customer, the business process corresponding to the business process tag is optimized. This can timely detect the bottlenecks in the business process and optimize them, thereby improving the business sales efficiency.
[0101] In an alternative embodiment, the potential tag includes a customer attitude tag.
[0102] The method further includes step S105:
[0103] S105. When the customer attitude tag is a negative attitude tag, send a target response statement to the target customer; and / or, end the conversation with the target customer.
[0104] Exemplarily, if the customer has no intention of purchasing the business promoted in this conversation and even has dissatisfaction with the behavior of promoting the business in this conversation, then the customer will inevitably show a negative attitude.
[0105] To timely perceive the customer's negative attitude and perform appeasement processing to avoid customer complaints, a target response statement can be pre-set. The target response statement is a statement customized by enterprise staff for appeasing customers, such as "Sorry to bother you. Thank you for answering. Wish you a happy life!" When the user terminal determines that the customer attitude tag is a negative attitude tag, it sends the target response statement to the customer and / or ends the conversation with the target customer.
[0106] It should be noted that the negative attitude tag is used to indicate the customer's tendency towards a negative attitude.
[0107] After obtaining the customer attitude tag of the target customer, the user terminal determines whether the customer attitude tag of the target customer is a negative attitude tag. If the customer attitude tag of the target customer is a negative attitude tag, it is considered that the target customer has no intention of purchasing the business promoted in this conversation and even has dissatisfaction with the behavior of promoting the business in this conversation. At this time, a target response statement is sent to the target customer and / or the conversation with the target customer is ended.
[0108] It should be noted that there are mainly the following three situations where the user terminal sends a target response statement to the target customer and / or ends the conversation with the target customer: The user terminal only sends a target response statement to the target customer; The user terminal only ends the conversation with the target customer; The user terminal first sends a target response statement to the target customer and then ends the conversation with the target customer.
[0109] In the embodiment of the present application, when the customer attitude label of the target customer is a negative attitude label, by sending a target response statement to the target customer and / or ending the conversation with the target customer, the negative attitude of the customer can be timely perceived and appeased, avoiding customer complaints.
[0110] Please refer to Figure 2 , Figure 2 which is a schematic structural diagram of a customer label mining device based on a dialogue scenario provided by the second embodiment of the present application. The second embodiment of the present application provides a customer label mining device based on a dialogue scenario, including: a dialogue record acquisition module 201, configured to acquire the current dialogue record with the target customer when having a conversation with the target customer; a potential label mining module 202, configured to input the current dialogue record into a pre-established label mining model to obtain potential labels of the target customer; wherein, the label mining model is used to identify one or more of business scenario labels, customer problem labels, sales process labels, and customer attitude labels from the current dialogue record as potential labels; a customer label determination module 203, configured to determine that the customer label of the target customer includes the potential labels.
[0111] In an optional embodiment, the acquisition of the current dialogue record with the target customer includes: acquiring the current round of question-and-answer statements sent by the target customer and the historical dialogue record with the target customer; combining the current round of question-and-answer statements and the historical dialogue record to obtain the current dialogue record.
[0112] In an optional embodiment, the label mining model includes one or more of a business scenario label mining model for identifying business scenario labels from the current dialogue record, a customer problem label mining model for identifying customer problem labels from the current dialogue record, a business process label mining model for identifying business process labels from the current dialogue record, and a customer attitude label mining model for identifying customer attitude labels from the current dialogue record.
[0113] In an optional embodiment, the label mining model includes a business scenario label mining model and a customer problem label mining model; the inputting the current dialogue record into the pre-established label mining model to obtain potential labels of the target customer includes: inputting the current dialogue record into the business scenario label mining model to obtain business scenario labels; inputting the current dialogue record and the business scenario labels into the customer problem label mining model to obtain customer problem labels.
[0114] In an alternative embodiment, the label mining model includes a business scenario label mining model and a business process label mining model; the step of inputting the current conversation record into the pre-established label mining model to obtain potential labels of the target customer includes: inputting the current conversation record into the business scenario label mining model to obtain business scenario labels; and inputting the current conversation record and the business scenario labels into the business process label mining model to obtain business process labels.
[0115] In an alternative embodiment, the apparatus further includes: a business process optimization module, configured to optimize the business process corresponding to the business process label when the business process label is a target business process label; wherein the target business process label includes a business process label identified from the historical conversation record with the target customer.
[0116] In an alternative embodiment, the potential labels include customer attitude labels; the apparatus further includes: a negative attitude perception module, configured to send a target reply statement to the target customer when the customer attitude label is a negative attitude label; and / or end the conversation with the target customer.
[0117] The implementation processes of the functions and roles of the various modules in the above apparatus are specifically described in the implementation processes of the corresponding steps in the above method, and will not be elaborated here.
[0118] Please refer to Figure 3 , Figure 3 , which is a schematic structural diagram of an electronic device provided in the third embodiment of the present application. The third embodiment of the present application provides an electronic device 30, including a processor 301, a memory 302, and a computer program stored in the memory 302 and configured to be executed by the processor 301; when the processor 301 executes the computer program, it implements the method described in the first embodiment of the present application and can achieve the same beneficial effects.
[0119] Among them, when the processor 301 reads the computer program from the memory 302 through the bus 303 and executes the computer program, it can implement the method described in any embodiment included in the method described in the first embodiment of the present application.
[0120] The processor 301 can process digital signals and can include various computing architectures. For example, a complex instruction set computer architecture, a reduced instruction set computer architecture, or an architecture that implements a combination of multiple instruction sets. In some examples, the processor 301 may be a microprocessor.
[0121] The memory 302 can be used to store instructions executed by the processor 301 or data related to the instruction execution process. These instructions and / or data may include code for implementing some or all of the functions of one or more modules described in the embodiments of the present application. The processor 301 in the embodiments of the present disclosure can be used to execute the instructions in the memory 302 to implement the method as described in the first embodiment of the present application. The memory 302 includes a dynamic random access memory, a static random access memory, a flash memory, an optical memory, or other memories well known to those skilled in the art.
[0122] The fourth embodiment of the present application provides a computer-readable storage medium, which includes a stored computer program; wherein, when the computer program runs, it controls the device where the computer-readable storage medium is located to execute the method as described in the first embodiment of the present application, and can achieve the same beneficial effects.
[0123] In summary, the embodiments of the present application provide a method, apparatus, device, and medium for mining customer tags based on a dialogue scenario. The method for mining customer tags based on a dialogue scenario includes: in the case of having a dialogue with a target customer, obtaining the current dialogue record with the target customer; inputting the current dialogue record into a pre-established tag mining model to obtain potential tags of the target customer; wherein, the tag mining model is used to identify one or more of business scenario tags, customer problem tags, business process tags, and customer attitude tags from the current dialogue record as potential tags; determining that the customer tags of the target customer include the potential tags. By inputting the current dialogue record with the target customer into a pre-established tag mining model in the case of having a dialogue with the target customer, the embodiments of the present application enable the tag mining model to identify one or more of business scenario tags, customer problem tags, business process tags, and customer attitude tags from the current dialogue record as potential tags, and obtain the potential tags of the target customer, so as to determine that the customer tags of the target customer include the potential tags, which can continuously understand and label potential key information such as business scenarios, customer problems, business processes, and customer attitudes in the dialogue record during the dialogue process with the customer, thereby mining more customer tags accurately in real time.
[0124] In several embodiments provided by the present application, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely illustrative. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architectures, functions, and operations of devices, methods, and computer program products according to multiple embodiments of the present application. In this regard, each block in the flowchart or block diagram may represent a module, a program segment, or a part of code, and the module, program segment, or part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order from that marked in the accompanying drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, as well as the combination of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system for performing the specified functions or actions, or can be implemented by a combination of dedicated hardware and computer instructions.
[0125] In addition, the functional modules in each embodiment of the present application can be integrated together to form an independent part, or each module can exist alone, or two or more modules can be integrated to form an independent part.
[0126] If the above functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present application. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.
[0127] The above are only the specific implementation manners of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed in the present application, and all should be covered by the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.
Claims
1. A customer tag mining method based on conversation scenarios, characterized in that: include: In the case of a conversation with a target customer, obtaining a current conversation record with the target customer; Input the current conversation record into a pre-established label mining model to obtain potential labels of the target customer; wherein the label mining model is used to identify one or more of business scenario labels, customer problem labels, business process labels and customer attitude labels from the current conversation record as the potential labels; Determine that the customer tag of the target customer includes the potential tag.
2. The method according to claim 1, characterized in that The obtaining of the current conversation record with the target customer includes: Obtaining the current round of question and answer statements sent by the target customer, as well as historical conversation records with the target customer; The current round of question and answer statements and the historical conversation records are combined to obtain the current conversation record.
3. The method according to claim 1, characterized in that The label mining model includes one or more of a business scenario label mining model for identifying the business scenario label from the current conversation record, a customer problem label mining model for identifying the customer problem label from the current conversation record, a business process label mining model for identifying the business process label from the current conversation record, and a customer attitude label mining model for identifying the customer attitude label from the current conversation record.
4. The method according to claim 3, characterized in that The label mining model includes the business scenario label mining model and the customer question label mining model; The step of inputting the current conversation record into a pre-established tag mining model to obtain potential tags of the target customer includes: Inputting the current conversation record into the business scenario label mining model to obtain the business scenario label; The current conversation record and the business scenario label are input into the customer question label mining model to obtain the customer question label.
5. The method according to claim 3, characterized in that: The label mining model includes the business scenario label mining model and the business process label mining model; The step of inputting the current conversation record into a pre-established tag mining model to obtain potential tags of the target customer includes: Inputting the current conversation record into the business scenario label mining model to obtain the business scenario label; The current conversation record and the business scenario label are input into the business process label mining model to obtain the business process label.
6. The method according to claim 5, characterized in that The method further comprises: In the case where the business process tag is a target business process tag, the business process corresponding to the business process tag is optimized; wherein the target business process tag includes a business process tag identified from a historical conversation record with the target customer.
7. The method according to any one of claims 1 to 6, characterized in that: The potential tags include the customer attitude tags; The method further comprises: When the customer attitude tag is a negative attitude tag, sending a target reply statement to the target customer; and / or, End the conversation with said target customer.
8. A customer tag mining device based on conversation scenarios, characterized in that: include: A conversation record acquisition module is used to acquire the current conversation record with the target customer when having a conversation with the target customer; A potential label mining module is used to input the current conversation record into a pre-established label mining model to obtain the potential label of the target customer; wherein the label mining model is used to identify one or more of business scenario labels, customer problem labels, sales process labels and customer attitude labels from the current conversation record as the potential label; The customer tag determination module is used to determine that the customer tag of the target customer includes the potential tag.
9. An electronic device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor; when the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program; wherein, when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the method according to any one of claims 1 to 7.