Talk skill recommendation method and device, storage medium and electronic equipment

Abstract

The invention relates to a verbal skill recommendation method and device, a storage medium and electronic equipment, and relates to the field of artificial intelligence, and the method comprises the steps: firstly obtaining the real-time conversation content of customer service personnel and a customer; semantic recognition is carried out on the real-time dialogue content, and a core problem of the client is extracted; analyzing according to the core problem to obtain a corresponding reply verbal skill; and then the reply verbal skill is recommended to the customer service staff. By applying the technical scheme of the invention, semantic recognition is performed on the real-time dialogue content between the customer service staff and the customer, and the corresponding reply verbal skill is recommended to the customer service staff, so that the professionality and uniformity of the reply verbal skill are improved, the working efficiency of the customer service staff is improved, the customer service staff can reply the problem consulted by the customer in time, and the user experience is improved. And the waiting time of the customer is reduced, so that the experience feeling of the customer is improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and specifically to a method, device, storage medium, and electronic device for recommending speech skills. Background Art

[0002] When customers want to know more about a product or service, they usually ask customer service staff for detailed information about the product or service. At this time, the customer service staff needs to help customers solve the product or service problems they encounter.

[0003] Currently, when answering customer inquiries, customer service personnel may need to switch between several interfaces in order to find relevant content of the customer's inquiries and lock in the content required by the customer among the query results of multiple interfaces.

[0004] However, customer service staff need to repeatedly search the knowledge base based on customer inquiries. Authority restrictions and knowledge blind spots may cause manual queries to be inaccurate and unable to respond to customers in a timely manner, which in turn leads to low work efficiency, long waiting times for customers, and a poor experience. Summary of the invention

[0005] In view of this, the present application provides a method, device, storage medium and electronic device for recommending speech skills, the main purpose of which is to improve the technical problem that the manual query method in the current existing technology is inefficient, resulting in long waiting time for customers and poor user experience.

[0006] In a first aspect, the present application provides a method for recommending a speech technique, comprising:

[0007] Get the real-time conversation between customer service staff and customers;

[0008] Perform semantic recognition on the real-time conversation content to extract the core issues of the customer;

[0009] Analyze the core issues and get corresponding reply words;

[0010] Recommend the reply script to the customer service staff.

[0011] Optionally, performing semantic recognition on the real-time conversation content to extract the core questions of the customer includes:

[0012] Performing semantic recognition on the real-time conversation content to obtain customer intention, and / or context emotion, and / or the customer's purchase history, and / or the customer's browsing behavior;

[0013] The core problem of the customer is determined based on the customer intention, and / or the contextual sentiment, and / or the purchase history, and / or the browsing behavior.

[0014] Optionally, before recommending the reply script to the customer service personnel, the method further includes:

[0015] Identify customer emotions based on the real-time conversation content;

[0016] Adjust the response based on the identified customer sentiment;

[0017] The step of recommending the reply script to the customer service personnel includes:

[0018] Recommend the adjusted response words to customer service staff.

[0019] Optionally, the analyzing the core question to obtain a corresponding reply script includes:

[0020] Input the core question into a target knowledge base for retrieval, wherein the target knowledge base includes a professional knowledge base and / or a general knowledge base, and the target knowledge base is constructed based on a tree structure management of specialized knowledge points and / or general knowledge points;

[0021] The reply words are determined according to the search results.

[0022] Optionally, the analyzing the core question to obtain a corresponding reply script includes:

[0023] If the search result corresponding to the core question cannot be obtained in the target knowledge base, the large model is called to search outside the scope of the target knowledge base and generate a reply script. The large model is used to process language-related tasks.

[0024] Optionally, the method further includes: if a preset keyword is identified in the real-time conversation content, based on the target knowledge base, recommending knowledge content related to the preset keyword to the customer service personnel.

[0025] Optionally, after recommending the reply script to the customer service staff, the method further includes:

[0026] Determine the type of business that the customer intends to handle based on the customer's core issues;

[0027] Based on the target knowledge base, generating a text message template according to the business type;

[0028] The SMS template is recommended to the customer service personnel.

[0029] In a second aspect, the present application provides a speech recommendation device, comprising:

[0030] An acquisition module is configured to acquire the real-time conversation content between the customer service staff and the customer;

[0031] A recognition module configured to perform semantic recognition on the real-time conversation content and extract the core questions of the customer;

[0032] An analysis module is configured to obtain corresponding reply words according to the core question analysis;

[0033] The recommendation module is configured to recommend the reply script to the customer service staff.

[0034] In a third aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the method described in the first aspect is implemented.

[0035] In a fourth aspect, the present application provides an electronic device, comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor implements the method described in the first aspect when executing the computer program.

[0036] By means of the above technical solution, the present application provides a method, device, storage medium and electronic device for recommending words. Specifically, firstly, the real-time conversation content between the customer service staff and the customer is obtained; then, semantic recognition is performed on the real-time conversation content to extract the customer's core question; the corresponding reply words are obtained according to the core question analysis; and then the reply words are recommended to the customer service staff. Compared with the current existing technology, by applying the technical solution of the present application, semantic recognition is performed on the real-time conversation content between the customer service staff and the customer, and the corresponding reply words are recommended to the customer service staff, which improves the professionalism and uniformity of the reply words, and at the same time improves the work efficiency of the customer service staff, so that the customer service staff can respond to the customer's inquiries in a timely manner, reducing the customer's waiting time, thereby improving the customer's experience.

[0037] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0039] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0040] Figure 1 A schematic diagram of a flow chart of a method for recommending speech skills provided in an embodiment of the present application is shown;

[0041] Figure 2 A schematic diagram showing a flow chart of another method for recommending speech skills provided in an embodiment of the present application is shown;

[0042] Figure 3 A schematic diagram showing an application example provided by an embodiment of the present application;

[0043] Figure 4 A structural schematic diagram of a speech recommendation device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0044] In order to more clearly understand the above-mentioned purposes, features and advantages of the present application, the scheme of the present application will be further described below. It should be noted that the embodiments of the present application and the features in the embodiments can be combined with each other without conflict.

[0045] In order to improve the inefficiency of the manual query method in the existing technology, which leads to long waiting time and poor user experience for customers. This embodiment provides a method for recommending speech skills, such as Figure 1 As shown, the method includes:

[0046] Step 101: Obtain the real-time conversation content between the customer service staff and the customer.

[0047] For example, the conversation content between customer service personnel and customers can be monitored in real time through the conversation management platform.

[0048] Step 102: Perform semantic recognition on the real-time conversation content to extract the customer's core questions.

[0049] In some examples, semantic recognition can include aspects such as semantic parsing, context understanding, entity recognition, sentiment analysis, intent recognition, and multi-round dialogue management.

[0050] For example, deep learning models can be used to capture the complex structure and patterns of language, analyze the customer's purpose or intention based on the understanding of the real-time conversation content, and then extract the core issues from the real-time conversation content so that the next step can be taken to deal with the core issues. Among them, the customer's core issues may include product information, service information, operation instructions, consulting suggestions, etc.

[0051] Step 103: Analyze the core issues and get corresponding reply words.

[0052] For example, it is possible to understand the customer's core issues in real time, analyze the core issues, obtain the Frequently Asked Questions (FAQs) that best match the core issues, and generate response scripts based on the real-time conversation content.

[0053] Step 104: Recommend the reply script to the customer service staff.

[0054] In some examples, the recommendation of reply words can support recommending words for customer service staff to choose from after recognizing keywords such as products, or proactively recommending reply words if it is recognized that the customer service staff has not responded for a period of time after a round of conversation. If no suitable reply words are returned, standard polite words can be recommended for customer service staff to choose from.

[0055] By applying the method of this embodiment, it is possible to intelligently provide customer service personnel with professional and unified words to reply to customers, thereby improving the efficiency of knowledge point query, reducing the burden on the query system, and improving the work efficiency of customer service personnel.

[0056] This embodiment first obtains the real-time conversation content between the customer service staff and the customer; then performs semantic recognition on the real-time conversation content to extract the customer's core question; obtains the corresponding reply script based on the core question analysis; and then recommends the reply script to the customer service staff. Compared with the current existing technology, this embodiment improves the professionalism and uniformity of the reply script by performing semantic recognition on the real-time conversation content between the customer service staff and the customer, and recommends the corresponding reply script to the customer service staff, while improving the work efficiency of the customer service staff, enabling the customer service staff to respond to the customer's inquiries in a timely manner, reducing the customer's waiting time, and thus improving the customer's experience.

[0057] Further, as a refinement and extension of the above embodiment, in order to fully illustrate the specific implementation process of the method of this embodiment, this embodiment provides the following Figure 2 The specific method shown includes:

[0058] Step 201: Obtain the real-time conversation content between the customer service staff and the customer.

[0059] For example, the real-time conversation may include "How is the insurance premium calculated?", "How to renew the insurance?", etc.

[0060] Step 202: Perform semantic recognition on the real-time conversation content to extract the customer's core questions.

[0061] For example, semantic recognition technology can be used to identify and extract core issues in multiple rounds of customer consultations, such as specific issues related to insurance coverage and compensation, so that recommended answers can be provided quickly and accurately later.

[0062] Optionally, step 202 may specifically include: performing semantic recognition on the real-time conversation content to obtain customer intent, and / or contextual emotions, and / or customer purchase history, and / or customer browsing behavior; determining the customer's core issue based on customer intent, and / or contextual emotions, and / or purchase history, and / or browsing behavior.

[0063] For example, the application of semantic recognition technology in real-time speech recommendation scenarios is an advanced artificial intelligence interaction method that can help customer service personnel provide more accurate and personalized responses or suggestions by deeply understanding the conversation content, contextual emotions and customer intentions. It can also recommend products that meet customer needs, product terms and other content to customer service personnel by analyzing the customer's purchase history, browsing behavior and customer intentions corresponding to the real-time conversation content.

[0064] Step 203: Input the core question into the target knowledge base for retrieval.

[0065] The target knowledge base includes a professional knowledge base and / or a general knowledge base, and the target knowledge base is constructed based on the tree structure management of specialized knowledge points and / or general knowledge points.

[0066] In some examples, building and maintaining an efficient knowledge base usually involves the following steps: maintenance of knowledge points, review of knowledge points, batch import of knowledge points, etc. Specifically, the maintenance of knowledge points can mainly include regular checks on the accuracy and timeliness of knowledge points to ensure that the content is up to date; adjust the structure of the knowledge base as needed, such as adding new categories or deleting categories that are no longer used, and maintain the logic and consistency of the structure so that users can easily find the information they need; continuously optimize the expression of content to make it clearer and more concise, and continuously improve the quality of content based on user feedback and usage data. The review of knowledge points can include the review of content accuracy, language clarity, and format consistency to ensure that the content of the knowledge points is accurate and free of factual errors, that the expression of the knowledge points is clear and easy to understand, and that the knowledge points follow a consistent format and style. Prepare the knowledge point data to be imported before batch importing. These data usually exist in the form of spreadsheets or databases. Then the knowledge point data can be imported into the knowledge base platform, and the data in the knowledge base can be reviewed and corrected.

[0067] For example, you can collect the most frequently asked questions by customers through customer service records, user feedback, online surveys, etc., organize the collected questions by topic or category, provide concise and clear answers for each question, ensure that the answers are accurate, provide practical information as much as possible, and review them regularly to ensure the accuracy and timeliness of the content.

[0068] Exemplarily, the target knowledge base is comprehensive, wherein the professional knowledge base contains relevant specialized knowledge points, and the general knowledge base contains relevant general knowledge points. Knowledge point retrieval based on the target knowledge base can be divided into keyword retrieval, full-text retrieval and other retrieval methods, and the management of the knowledge base adopts tree structure management.

[0069] Step 204: Determine the reply words based on the search results.

[0070] For example, the knowledge search scope can be set, and the professional knowledge base can be called first to retrieve the corresponding reply scripts for the professional knowledge points, thereby improving the retrieval efficiency. If no results are found in the professional knowledge base, the general knowledge base can be further called for retrieval and reply, and finally the recommended scripts can be provided to the customer service staff.

[0071] Optionally, the method of this embodiment may further include: if the search result corresponding to the core question cannot be obtained in the target knowledge base, calling the large model to search outside the scope of the target knowledge base and generate a reply script. The large model can be used to process language-related tasks.

[0072] For example, if the customer's core question is the compensation status of a certain insurance product, the knowledge base can retrieve the product terms and common problems with the product, and promptly return the customer service staff with recommended response words and cited sources; if the customer's core question is about a certain operation activity of the company, and the professional product knowledge base cannot meet the retrieval requirements, the knowledge base can retrieve and search company-wide documents, and promptly return the customer service staff with recommended response words and cited sources; if the customer's core question is a daily knowledge question, the knowledge base can retrieve general knowledge and related content, and promptly return the customer service staff with recommended response words and cited sources; if the general knowledge base can no longer meet the retrieval requirements, the knowledge base can retrieve the big model capabilities to complete the interaction with the big model, and promptly return the customer service staff with the big model recommended response words; if it involves standardized customer service etiquette, opening remarks, closing remarks, appeasement, apology and other scenarios, the customer service staff can be recommended how to communicate with customers based on basic etiquette standards to enhance customer experience.

[0073] For example, if the customer's question is "How to apply for insurance", you can search the knowledge base for the methods of applying for insurance. If the search results are online application on the official website, telephone application, email application, and offline application at the insurance company, the corresponding response can be "You can apply online through the insurance company's official website, or by phone, email, or in person at the insurance company"; if the customer's question is "What types of insurance products are there", you can search the knowledge base for the types of insurance products. If the search results are life insurance and property insurance, and there are further detailed categories, the corresponding response can be "Insurance products can be roughly divided into two categories: life insurance and property insurance. Among them, life insurance includes life insurance, health insurance, accidental injury insurance, etc.; property insurance includes home insurance, car insurance, commercial property insurance, etc.."

[0074] Step 205: Recommend the reply script to the customer service staff.

[0075] In some examples, the recommendation of real-time reply scripts can support manual and / or automatic recommendations. Among them, manual recommendation can support customer service personnel to select a certain dialogue to provide recommended scripts, and automatic recommendation will recommend scripts for customer service personnel to refer to after identifying keywords such as products, or actively provide recommended scripts after a certain round of dialogue when the customer service personnel do not respond for a period of time. If there is no suitable answer, it will support the recommendation of standard scripts to politely reply to the customer, and the customer service staff will follow up in a timely manner.

[0076] Optionally, the method of this embodiment may further include: identifying customer emotions based on the real-time conversation content; adjusting the reply language based on the identified customer emotions; accordingly, step 205 may specifically include: recommending the adjusted reply language to the customer service staff.

[0077] In some examples, customer emotions can be recognized based on the customer's real-time conversation information, and sales strategies can be adjusted based on the recognized customer emotions. The reply language can be adjusted according to the sales strategy to provide customers with instant responses and improve customer experience.

[0078] For example, semantic understanding and / or sentiment analysis techniques can be used to analyze the customer's text content, understand the meaning of the customer's question, and identify the customer's emotional state. Based on the identified customer emotions, the system can adjust the response language. For example, if the customer expresses anger or dissatisfaction, the system may add more soothing and understanding response words to the response language; if the customer appears anxious or confused, the response language can be adjusted to one that can answer the question more directly, and express the intention to try to provide further help to the customer to better meet the customer's needs.

[0079] For example, when recommending a product to a customer and receiving feedback from the customer on the recommended product, the customer's preference and intention for the product can be determined through customer emotion recognition. For example, if the customer's conversation content is "This is too expensive, it's only about 2,000 yuan", then based on the customer's statement, the customer service staff can search for products with a price of about 2,000 yuan and similar performance in the knowledge base, and then recommend the corresponding product to the customer to improve the customer experience. Furthermore, a comparison of various information of previous products and current products can be given to make it easier for customers to make choices.

[0080] Optionally, the method of this embodiment may further include: if a preset keyword is identified in the real-time conversation content, then based on the target knowledge base, recommending knowledge content related to the preset keyword to the customer service personnel.

[0081] In some examples, the search scope corresponding to the real-time reply script recommended to customer service personnel can be set. For example, specific conversation content can be selected, keywords can be identified for knowledge base retrieval, and knowledge content and reply scripts can be recommended. When the customer service personnel open the smart assistant, knowledge content related to the customer and the products they purchase is displayed to improve the work efficiency of the customer service personnel while reducing the query frequency of the system.

[0082] For example, the preset keywords can be words with high consultation frequency related to products and / or services, such as insurance terms, claims procedures, etc. After the keywords are identified, the content or processes related to the keywords can be automatically displayed to help customer service personnel quickly understand the products and / or services mentioned by customers, enrich the knowledge reserves of customer service personnel, and further help customer service personnel organize and select words, improve the accuracy and professionalism of responses in different scenarios, so as to bring a better experience to customers.

[0083] Optionally, the method of this embodiment may also include: determining the type of business that the customer intends to handle based on the customer's core question; generating a text message template based on the business type based on the target knowledge base; and recommending the text message template to customer service personnel.

[0084] For example, Figure 3 As shown, the intelligent assistant in this embodiment can realize functions such as real-time quality inspection and reminder, FAQ and speech recommendation, SMS template recommendation, work order template recommendation, etc., to improve the work efficiency and work quality of customer service personnel.

[0085] In some examples, by applying the method of this embodiment, the most likely SMS template to be sent can be recommended to the customer service staff based on the content of the conversation between the customer and the customer service staff, including recommending commonly used SMS templates or personalized SMS templates based on different service scenarios, historical behaviors and / or preferences. SMS replies can also be automatically generated to customers, such as product introductions, claims notifications, event invitations, etc. If the customer frequently purchases insurance products, the system may recommend related product event information templates.

[0086] For example, during a call, the customer service staff's service attitude and service language can be quality checked in real time, and the customer service staff can be reminded after any violation, so as to ensure the quality of the conversation.

[0087] For example, by understanding the customer's questions in real time, when the system extracts the customer's core question through understanding the context of the real-time conversation content or recognizes that the customer's question contains specific keywords, it will automatically recommend relevant FAQ items or scripts, and recommend the most matching FAQ to the customer service staff. For example, if the customer mentions "claims", the system will recommend related FAQs about the claims policy.

[0088] For example, the most likely work order template to be filled out can be recommended to the customer service staff based on the content of the conversation between the customer and the customer service staff and the nature of the customer's question, such as technical issues, returns and refunds, invoice issues, etc.

[0089] Compared with the current existing technology, this embodiment performs semantic recognition on the real-time conversation content between customer service personnel and customers, and recommends professional and unified reply words to customer service personnel, thereby eliminating the step of customer service personnel repeatedly searching the knowledge base, increasing the accuracy of query results and the professionalism of reply words, and to a great extent avoiding reply delays and errors caused by insufficient grasp of knowledge points, thereby improving the work efficiency of customer service personnel, enabling customer service personnel to respond to customer inquiries in a timely manner, reducing customer waiting time, and thus improving customer experience.

[0090] Further, as Figure 1 The specific implementation of the method shown in the embodiment provides a speech recommendation device, such as Figure 4 As shown, the device includes: an acquisition module 31, an identification module 32, an analysis module 33, and a recommendation module 34.

[0091] An acquisition module 31 is configured to acquire the real-time conversation content between the customer service staff and the customer;

[0092] The recognition module 32 is configured to perform semantic recognition on the real-time conversation content and extract the core questions of the customer;

[0093] An analysis module 33 is configured to obtain corresponding reply words according to the core question analysis;

[0094] The recommendation module 34 is configured to recommend the reply script to the customer service staff.

[0095] In some examples, the recognition module 32 is specifically configured to perform semantic recognition on the real-time conversation content to obtain customer intent, and / or contextual sentiment, and / or the customer's purchase history, and / or the customer's browsing behavior; and determine the customer's core issue based on the customer intent, and / or the contextual sentiment, and / or the purchase history, and / or the browsing behavior.

[0096] In some examples, the recognition module 32 is further configured to recognize customer emotions based on the real-time conversation content; adjust the reply language based on the recognized customer emotions; correspondingly, the recommendation module 34 is further configured to recommend the adjusted reply language to the customer service staff.

[0097] In some examples, the analysis module 33 is specifically configured to input the core question into a target knowledge base for retrieval, wherein the target knowledge base includes a professional knowledge base and / or a general knowledge base, and the target knowledge base is constructed based on a tree structure management of specialized knowledge points and / or general knowledge points; and determine the reply words based on the retrieval results.

[0098] In some examples, the analysis module 33 is further specifically configured to call the big model to perform a search outside the scope of the target knowledge base and generate a reply script if the search result corresponding to the core question cannot be obtained in the target knowledge base. The big model is used to process language-related tasks.

[0099] In some examples, the recommendation module 34 is further configured to recommend knowledge content related to the preset keyword to the customer service personnel based on the target knowledge base if a preset keyword is identified in the real-time conversation content.

[0100] In some examples, the recommendation module 34 is specifically configured to determine the type of business that the customer intends to handle based on the customer's core issue; generate a text message template based on the business type based on the target knowledge base; and recommend the text message template to the customer service staff.

[0101] Based on the above Figure 1 and Figure 2 The method shown in the embodiment also provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the above-mentioned Figure 1 and Figure 2 The method shown.

[0102] Based on this understanding, the technical solution of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, USB flash drive, mobile hard disk, etc.), including a number of instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods of various implementation scenarios of the present application.

[0103] Based on the above Figure 1 and Figure 2 The method shown, and Figure 4 In order to achieve the above-mentioned purpose, the embodiment of the present application further provides an electronic device, which may include a storage medium and a processor; the storage medium is used to store a computer program; the processor is used to execute the computer program to achieve the above-mentioned Figure 1 and Figure 2 The method shown.

[0104] Optionally, the above-mentioned physical device may also include a user interface, a network interface, a camera, a radio frequency (RF) circuit, a sensor, an audio circuit, a WI-FI module, etc. The user interface may include a display, an input unit such as a keyboard, etc., and the optional user interface may also include a USB interface, a card reader interface, etc. The network interface may optionally include a standard wired interface, a wireless interface (such as a WI-FI interface), etc.

[0105] Those skilled in the art will appreciate that the above-mentioned physical device structure provided in this embodiment does not constitute a limitation on the physical device, and may include more or fewer components, or a combination of certain components, or different arrangements of components.

[0106] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages the hardware and software resources of the above-mentioned physical device, and supports the operation of the information processing program and other software and / or programs. The network communication module is used to realize the communication between the components inside the storage medium, and the communication with other hardware and software in the information processing physical device.

[0107] Through the description of the above implementation methods, the technical personnel in this field can clearly understand that the present application can be implemented by means of software plus the necessary general hardware platform, or by hardware. By applying the scheme of this embodiment, the real-time conversation content between the customer service staff and the customer is semantically recognized, and professional and unified reply words are recommended to the customer service staff, which saves the customer service staff from repeatedly searching the knowledge base, increases the accuracy of the query results and the professionalism of the reply words, and greatly avoids the reply delays and errors caused by insufficient knowledge points, improves the work efficiency of the customer service staff, enables the customer service staff to reply to the customer's questions in a timely manner, reduces the customer's waiting time, and thus improves the customer's experience.

[0108] 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 such actual relationship or order between these entities or operations. Moreover, the term "comprising" 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 includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprising a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.

[0109] The above is only a specific implementation of the present application, so that those skilled in the art can understand or implement the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to these embodiments herein, but will conform to the widest scope consistent with the principles and novel features applied for herein.

Claims

1. A method for recommending speech skills, characterized in that: include: Get the real-time conversation between customer service staff and customers; Perform semantic recognition on the real-time conversation content to extract the core issues of the customer; Analyze the core issues and get corresponding reply words; Recommend the reply script to the customer service staff.

2. The method according to claim 1, characterized in that The semantic recognition of the real-time conversation content is performed to extract the core questions of the customer, including: Performing semantic recognition on the real-time conversation content to obtain customer intention, and / or context emotion, and / or the customer's purchase history, and / or the customer's browsing behavior; The core problem of the customer is determined based on the customer intention, and / or the contextual sentiment, and / or the purchase history, and / or the browsing behavior.

3. The method according to claim 1, characterized in that Before recommending the reply script to the customer service personnel, the method further includes: Identify customer emotions based on the real-time conversation content; Adjust the response based on the identified customer sentiment; The step of recommending the reply script to the customer service personnel includes: Recommend the adjusted response words to customer service staff.

4. The method according to claim 1, characterized in that: The corresponding reply words obtained according to the analysis of the core question include: Input the core question into a target knowledge base for retrieval, wherein the target knowledge base includes a professional knowledge base and / or a general knowledge base, and the target knowledge base is constructed based on a tree structure management of specialized knowledge points and / or general knowledge points; The reply words are determined according to the search results.

5. The method according to claim 4, characterized in that The corresponding reply words obtained according to the analysis of the core question include: If the search result corresponding to the core question cannot be obtained in the target knowledge base, the large model is called to search outside the scope of the target knowledge base and generate a reply script. The large model is used to process language-related tasks.

6. The method according to claim 4, characterized in that The method further comprises: If a preset keyword is identified in the real-time conversation content, knowledge content related to the preset keyword is recommended to the customer service personnel based on the target knowledge base.

7. The method according to claim 4, characterized in that After recommending the reply script to the customer service personnel, the method further includes: Determine the type of business that the customer intends to handle based on the customer's core issues; Based on the target knowledge base, generating a text message template according to the business type; The SMS template is recommended to the customer service personnel.

8. A speech recommendation device, characterized in that: include: An acquisition module is configured to acquire the real-time conversation content between the customer service staff and the customer; A recognition module configured to perform semantic recognition on the real-time conversation content and extract the core questions of the customer; An analysis module is configured to obtain corresponding reply words according to the core question analysis; The recommendation module is configured to recommend the reply script to the customer service staff.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

10. An electronic device comprising a storage medium, a processor, and a computer program stored in the storage medium and executable on the processor, characterized in that: When the processor executes the computer program, the method according to any one of claims 1 to 7 is implemented.

Images