CN117633199A

CN117633199A - Recommended content generation method and electronic equipment

Info

Publication number: CN117633199A
Application number: CN202311516442.7A
Authority: CN
Inventors: 陈杰; 金阳春; 余昭润; 蒋炯明; 蔡加佳; 庞皓翰
Original assignee: Zhejiang Tmall Technology Co Ltd
Current assignee: Zhejiang Tmall Technology Co Ltd
Priority date: 2023-11-14
Filing date: 2023-11-14
Publication date: 2024-03-01

Abstract

The embodiment of the application discloses a recommended content generation method and electronic equipment, wherein the method comprises the following steps: in the process of communicating with a user by a client service person in a target service scene, acquiring voice content; converting the voice content into text content, and identifying the sender of the voice content so as to add sender identification information to the text content; and constructing a prompt word text for carrying out dialogue with an artificial intelligence AI large language model according to the text content with the sender identification information and the candidate appeal type information in the target service scene so as to generate recommended content on at least one information element field required in the appeal list by the AI large language model. Through this application embodiment, can promote and complain processing efficiency, reduce cost.

Description

Recommended content generation method and electronic equipment

Technical Field

The present invention relates to the field of information processing technologies, and in particular, to a recommended content generation method and an electronic device.

Background

In the commodity information service system, after-sale problems often occur, some problems cannot be agreed between a consumer and a merchant, even when disputes and the like occur, the consumer can complain to the platform side, and then the consumer's appeal can be known by the manual customer service of the platform side in a telephone manner, and a solution is provided for the consumer.

In the above process, if the manual customer service can give a specific solution and obtain the approval of the user, other operators can further perform specific landing execution on the solution; alternatively, for some more complex problems, especially those involving multiple different departments to cooperatively resolve, a customer service person who specifically communicates with the customer may not be able to directly give a solution, but need to transfer to other departments to resolve, etc. In any of the above cases, it is generally necessary to generate a corresponding order by manually taking a customer service during or after the call is completed, and the order may be used as an information carrier in the subsequent scenario of performing or transferring the specific solution to the floor. In the prior art, such a requirement list is usually manually created by a manual customer service responsible for wiring, including specific requirement points, prediction of a requirement scene, and the like, which are all performed empirically by the manual customer service, and then filled into corresponding option boxes, and the like. However, this places a relatively high demand on the service capabilities of the human service, and therefore, results in a high labor cost. In addition, the manual induction summarizing, scene prediction and other processes may have the problems of errors, low efficiency and the like.

Disclosure of Invention

The application provides a recommended content generation method and electronic equipment, which can improve the appeal processing efficiency and reduce the cost.

The application provides the following scheme:

a recommended content generation method, comprising:

in the process of communicating with a user by a client service person in a target service scene, acquiring voice content;

converting the voice content into text content, and identifying the sender of the voice content so as to add sender identification information to the text content;

and constructing a prompt word text for carrying out dialogue with an artificial intelligence AI large language model according to the text content with the sender identification information and the candidate appeal type information in the target service scene so as to generate recommended content on at least one information element field required in the appeal list by the AI large language model.

Wherein, still include:

before constructing the prompt word text, word segmentation and/or sentence breaking processing is carried out on the text content.

The recommended content generated by the AI large language model comprises the following components: and determining the matched appeal type from the candidate appeal types by the AI large language model through model understanding of the text content.

Wherein, still include:

and carrying out preliminary prediction on the type of the appeal to which the text content belongs in a keyword matching mode, and adding the preliminary predicted type of appeal information into the prompt word text so that the AI large language model can carry out prediction on the type of appeal by combining the preliminary predicted type of appeal information.

The recommended content generated by the AI large language model comprises the following components: and by summarizing the text content, the generated user problem gist descriptive content, user appeal gist descriptive content and solution content given by the customer service personnel are used for selectively filling in the appeal list.

The text of the prompt word also comprises format information of output content, so that the AI large language model generates recommended content on at least one information element field required in the resort list according to the format information.

Wherein, still include:

responding to an operation request of the recommended content storage appeal list submitted by the customer service personnel and generated according to the AI large language model, analyzing the recommended content to generate formatted appeal information, and generating the appeal list.

A recommended content generation device comprising:

the voice content acquisition unit is used for acquiring voice content in the process of communicating with the user by the client service personnel in the target service scene;

a voice recognition unit for converting the voice content into text content and recognizing an issuer of the voice content so as to add issuer identification information to the text content;

and the prompt word text construction unit is used for constructing a prompt word text used for carrying out dialogue with the artificial intelligence AI large language model according to the text content with the sender identification information and the candidate appeal type information in the target service scene so as to generate recommended content on at least one information element field required in the appeal list by the AI large language model.

A computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of the method of any of the preceding claims.

An electronic device, comprising:

one or more processors; and

a memory associated with the one or more processors, the memory for storing program instructions that, when read for execution by the one or more processors, perform the steps of the method of any of the preceding claims.

According to a specific embodiment provided by the application, the application discloses the following technical effects:

according to the embodiment of the application, in a specific service scene, in the process of carrying out voice conversation between a customer service person and a user, voice content generated by conversation between the customer service person and the user can be obtained, then voice recognition is carried out, the voice content is converted into text content, the sender of the voice content is recognized, and sender identification information is added for the text content. And then, according to the text content with the sender identification information and the candidate appeal type information in the specific service scene, constructing a prompt word text for carrying out dialogue with an artificial intelligence AI large language model so as to generate the content on the specific component fields in the appeal list for the voice dialogue process by the AI large language model for selection or reference by the customer service personnel. By the method, intelligent generation of the resort list content can be realized through the AI large language model, and dependence on experience of customer service personnel and capabilities of summarization, scene prediction and the like is reduced, so that efficiency is improved, and labor cost is reduced.

Of course, not all of the above-described advantages need be achieved at the same time in practicing any one of the products of the present application.

Drawings

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings that are needed in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.

FIG. 1 is a schematic diagram of a system architecture provided by an embodiment of the present application;

FIG. 2 is a flow chart of a method provided by an embodiment of the present application;

FIG. 3 is a schematic illustration of an interface provided by an embodiment of the present application;

FIG. 4 is a schematic illustration of another interface provided by an embodiment of the present application;

FIG. 5 is a schematic diagram of an apparatus provided by an embodiment of the present application;

fig. 6 is a schematic diagram of an electronic device provided in an embodiment of the present application.

Detailed Description

The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are only some, but not all, of the embodiments of the present application. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments in the present application are within the scope of the protection of the present application.

Firstly, as described in the background art section, in the scenes of after-sales consultation/complaints and the like, the complaint list is mainly used for bearing several elements such as user information, order information, complaint type information and the like, and in addition, the description of problem points fed back by the user, description of points of user complaint points, solutions and the like can be selectively included. In the prior art, the contents of the plurality of information element fields need to be summarized and recorded by a human customer service. In addition, in order to facilitate classification management or viewing of user requirements, a plurality of requirement types may be defined in advance in the system, and by classifying a plurality of different requirements of a user into corresponding requirement types, a requirement list classification management or display processing based on the requirement types is realized. At this time, in the process of communicating with the user, the manual customer service is required to predict the type of the specific requirement according to own experience, and when creating the requirement list, the manual selection filling and other processes are required to be performed in the corresponding option boxes.

Aiming at the problems, in the embodiment of the application, the intelligent resort single content generation based on the manual customer service and user communication process can be realized by combining an AI (Artificial Intelligence ) large language model (Large Language Model, LLM, simply referred to as AI large model) with voice recognition and other technologies, so that the efficiency is improved, the dependence on the scene prediction capability of customer service personnel is reduced, and the labor cost is reduced.

Among them, for ease of understanding, the AI large language model is briefly described below. The AI large Model may refer to a basic Model (Foundation Model), and in particular, may refer to a Model that is trained using massive data, has a huge amount of parameters, and can adapt to a series of downstream tasks. For the AI large model, there is a characteristic that the parameter amount is huge (along with the continuous iteration of the model, the parameter amount generally increases exponentially, from one hundred million to one trillion to one million, and even more) on the parameter scale, and from the mode support, the AI large model gradually develops to support multiple tasks in multiple modes from supporting a single task in a single mode such as picture, image, text, voice, video and the like. That is, the large model generally has high-efficiency understanding capability of multi-mode information, cross-mode sensing capability, migration and execution capability of cross-differentiation tasks, and the like, and may even have multi-mode information sensing capability as embodied by human brain.

From another perspective, the AI large model is a short for an artificial intelligence pre-training large model, and comprises two layers of meanings of the pre-training and the large model, and the two layers of meanings are combined to generate a new artificial intelligence mode, namely, the model can support various downstream applications without fine adjustment after the pre-training is completed on a large-scale data set or with fine adjustment of a small amount of data. That is, the AI large model benefits from its paradigm of "large-scale pretraining plus fine tuning," which can adapt well to different downstream tasks, exhibiting its powerful versatility. The large AI model with universality can obtain excellent performance only by carrying out corresponding fine adjustment in different downstream application scenes under the condition of sharing parameters, and breaks through the limitation that the traditional AI model is difficult to generalize to other tasks.

From the viewpoint of the processing results, the above-described AI large Model also belongs to a Generative Model. Such models not only can "understand" how the data was generated based on the feature predictions, but can also "create" new data based thereon.

In terms of a content generation mode, what content needs to be generated by the AI large model can be "told" through constructing Prompt word (Prompt) text (a language used for interacting with the artificial intelligent model), and the content generated by the AI large model can be enabled to accord with the required expectations and requirements through writing the special form of the Prompt word text.

Because the AI large model has the capability of creating the content, the AI large model can be applied to the embodiment of the application to assist customer service personnel to complete tasks such as the creation of a resort sheet in the voice call process. That is, in the process that the customer service personnel and the user carry out voice dialogue, the recommended content on at least one information element field required in the resort list can be generated by the AI large model, the customer service personnel can select whether to adopt the suggestion of the AI large model, if so, the recommended content generated by the AI large model can be automatically filled into the option frame corresponding to the resort list, and therefore, the processes of manually carrying out resort scene prediction, summarization of resort points and the like by the customer service personnel are omitted, and the efficiency is improved.

In specific implementation, some targeted training may be performed on the AI large model first, for example, the history call record and the corresponding type of the appeal may be used as training data and input into the AI large model, so that the AI large model obtains the capability of generating the foregoing.

Of course, when the AI large model is applied to the scene in the embodiment of the present application, for example, first, the input of the AI large model is usually mainly based on text, for example, the foregoing prompt word text, so that the output content of the AI large model can be controlled relatively well. However, in the embodiment of the present application, particularly, in a scenario where a customer service person performs a voice dialogue with a user, if voice content is directly input into the AI large model, although the AI large model may also have an ability to understand the voice content, the content output by the AI large model may be difficult to be effectively controlled. Therefore, in specific implementation, the related technology of voice-to-text can be combined to construct the prompt word text for dialogue with the AI large model. In the process of constructing the prompt word text, some preprocessing and the like can be performed on text contents converted from the voice in advance, so that the quality of contents generated by the AI large model is further improved.

In addition, since the content of the conversation may have a relatively strong randomness or spoken language in the process of performing a voice conversation with the customer service personnel, for example, some users may speak a lot of content irrelevant to the current appeal in addition to stating the own appeal in the process of performing the conversation with the customer service personnel, and at this time, if the summary of the appeal types is freely performed by the AI big model, the output result may deviate from the after-sales service scene, and the like. Therefore, in concrete implementation, various types of appeal that may be possessed by a user in scenes such as after-sales service can be summarized in advance, and in concrete construction of a prompt word text for dialogue with the AI large model, besides the text content converted from voice content, the various types of appeal can be input into the AI large model as candidates, so that the AI large model can predict content related to the types of appeal in the candidates, and therefore limit on the freedom of generating the content of the AI large model is achieved.

From the view of system architecture, referring to fig. 1, the embodiment of the present application may be applied to a merchandise information service system, and specifically may be a service function for providing specific intelligent generation of a claim list content in a workstation system of a customer service staff. In particular, when a user sends out requirements such as complaints to the platform side, customer service personnel of the platform side can actively initiate a call to the user under the necessary condition, so that the user requirements are known in more detail, and solutions and the like are determined in the call process. In the voice call process of customer service personnel and users, services such as 'resort recommendation' can be provided for the customer service personnel, specifically, processing such as voice-to-text conversion and generation of prompt word texts can be automatically performed, recommended content is generated by an AI large model based on the specific prompt word texts, then a resort list can be automatically generated according to the recommended content, and working such as resort type prediction in the call process of the customer service personnel is not needed, so that efficiency can be improved.

Specific embodiments provided in the embodiments of the present application are described in detail below.

First, the embodiment of the application provides a content generation method, referring to fig. 2, the method may specifically include:

s201: and in the process of communicating with the user by the client service personnel in the target service scene, acquiring voice content.

The target service scenario may specifically include an after-sales service scenario in a merchandise information service system, where the after-sales service scenario is usually an after-sales service provided by a customer service person on a platform side for a consumer user (for example, specifically may occur after the consumer user communicates with an after-sales person of a specific merchant and is invalid, or when a dispute occurs between the consumer user and the merchant, etc.); of course, other service scenarios may also be included in the specific implementation, such as, for example, an on-the-fly service scenario, and so forth.

In specific implementation, after a user initiates a request for complaints or other related after-sales services to a platform side, customer service personnel can receive specific request information in a workbench interface of the customer service personnel and then can enter a detail interface corresponding to the request, for example, an interface shown in fig. 3, and content related to information submitted when the user initiates the request can be displayed in the detail interface; if the customer service personnel consider that the detailed requirements of the user need to be further known, the call process with the user can be initiated through options such as 'call initiation' provided in the detail interface, for example, the call process can specifically comprise voice call, video call, and the like. For example, after the call is initiated, a window as shown in fig. 3 may be popped up to show information related to the current call, including the called number and the calling number. After the user side answers successfully, customer service personnel can conduct dialogue with the user so as to know the detailed requirements of the user and communicate specific solutions with the user.

In the prior art, customer service personnel are required to summarize the problems, the requirement points and the like of the user while thinking how to provide a proper solution for the user in the conversation process, and also determine the specific requirement type so as to complete the task of creating the requirement list. In the embodiment of the application, the AI large model can assist the customer service personnel to complete the process of creating the resort, in addition, the problem key points, the resort key points and the like of the user can be summarized, and the customer service personnel only need to think about whether to adopt the suggestion given by the AI large model. Naturally, before the AI large model generates the content, in order to enable control of the AI large model generated content, the voice content of both parties of the call may be processed first, and therefore, the voice content generated during the call of both parties may be acquired first. Specifically, because the specific call function can be provided by the customer service personnel workbench, the generated voice content can be directly collected in the call process.

S202: the voice content is converted into text content, and an issuer of the voice content is identified to add issuer identification information to the text content.

After the voice content is collected, the voice content may first be converted to text content for the purpose of facilitating construction of specific alert word text, which may be accomplished by voice recognition techniques. Of course, in the embodiment of the present application, specific voice content is continuously generated, so the voice recognition process may also be performed in real time, that is, new text content may be continuously recognized.

In addition, in addition to identifying text content, the sender corresponding to the specific voice content may be identified, that is, who each specific sentence is spoken may be identified, and then the sender identifier may be added to the converted text content. The specific sender identifier may represent the number of a specific customer service person, the nickname of the user, or may represent only the identity distinction of the customer service person or the user.

There are various ways to identify the sender information corresponding to the voice content. For example, recognition of voice features may be included to distinguish them, or keywords or the like contained in specifically recognized text content may be combined to recognize them, or the like.

In the preferred embodiment, besides performing voice recognition and adding the sender identifier, the text content obtained by conversion can be subjected to word segmentation, sentence breaking and other processing, so that the AI large model can more accurately understand specific input content, and meanwhile, the accuracy of the generated content is improved.

For example, for a certain call process, the specific speech is converted into text, sentence breaking is performed, and after the sender identifier is added, the speech can be displayed as:

1 x 2\n. N is equal to n is you good, i is equal to two is hot line, you buy the toilet brush of 7.9 yuan before. Hiccup, the second party applies for after-sale intervention woolen before seeing you, feedback that goods are not received, and the subsequent me sees that express display is sent to you in the morning today, and what is the goods are received or not? N1 2\n I am not there, I say that the address was written wrongly, let him change, he say nothing, then … …

Wherein, "1 x 2" represents a user, "lux" represents a customer service person, "n" represents a line feed, it can be seen that specific text content is sentence-broken, punctuation marks are added, and in addition, whether each sentence is spoken by the customer service person or the user can be displayed, and the like.

S203: and constructing a prompt word text for carrying out dialogue with an artificial intelligence AI large language model according to the text content with the sender identification information and the candidate appeal type information in the target service scene so as to generate recommended content on at least one information element field required in the appeal list by the AI large language model.

After converting the voice content into text content and performing sentence breaking, adding information such as a sender identification, a Prompt word text (Prompt) for dialog with the AI large model may be constructed. Specifically, the prompt word text may include a start dialogue content to be used for "telling" what content the AI big model needs to generate, for example, "you are an official customer service, you need to give a appeal scene according to the problem of the consumer, and summarize the points of the appeal. Such an open dialog may be generic, i.e. the text of the prompting word may be started with this sentence for a call between different users and different customer service personnel. In addition, the text of the prompt word can also comprise format information of output content, so that the AI big model carries out specific content output according to the format information. For example, "directly summarize dialog content in the format of \\problem: xxxx\n complaint: xxxx\n solution: xxxx\n complaint scene: xxxx\n', etc.

Furthermore, as described above, the candidate appeal type information in the target service scene may be provided in the prompt word text, that is, in the same target service scene, there may be multiple different types of appeal of the user, the possible appeal types may be summarized in advance, the candidate appeal types are provided, the AI big model may select a specifically matched appeal type from the candidate appeal types, and then the dialogue gist may be summarized on the basis, so as to provide a suggested solution, and avoid too random content output by the AI big model.

Among these, there may be one-to-many relationships between the target service scenario and the type of appeal, for example, for an after-sales service scenario, specific possible types of appeal may include: merchant shipping and logistics problems, door-to-door installation, return logistics, shipping logistics other problems, and the like; for an off-the-shelf service scenario, the type of appeal may include the acquisition of coupons, the reimbursement use of coupons, and so forth. In specific implementation, the correspondence between the target service scenario and the requirement type may be saved in a form of a data table, for example, as shown in table 1:

TABLE 1

After determining the current target service scene, determining which candidate appeal types correspond to through querying the data table and other modes, and further, using the appeal types as candidate items to be embodied in a prompt word text, namely, determining the appeal type matched with the current dialogue content from the candidate appeal types by using an AI large model, and summarizing the problem key points, the appeal key points and the like of the user on the basis.

In a specific implementation, the start dialogue, format information, candidate appeal types and the like of the prompt word text can be universal and can exist in the form of a prompt word template, so that after processing such as voice-to-text processing and the like is completed for different conversation processes, the prompt word text specific to the current conversation process can be generated only by adding specific dialogue content into the template.

For example, for a certain dialog process, the specifically generated prompt text may be:

"you are an official customer service, you need to give a appeal scene according to the consumer's problem, and provide a solution. The following is a chat record of you and consumers, directly summarizing the dialog content in the format of \\ 'question: xxxx\n complaint: xxxx\n solution: xxxx\n complaint scenario: xxxx\n'. The alternatives of the scene are: merchant's shipment and commodity circulation problem, the installation of going to the door, the commodity returned commodity circulation, other problems of shipment commodity circulation, refund, change goods, maintenance, repair send, freight … …, the dialogue is: 1.2 n 03-28:04:37 n. N is your own, i is your hot line little two, you have purchased the brand shampoo before. After-market intervention was applied before you see hiccup, feedback that no hand pair was received? N1 2\n pairs, i have purchased the shampoo before, none of the small samples i say in the details are received. N. times. Shu/n-kappa. N1 2\n i wish to restock these children. Can n? . One can, one can help you complement delivery one by n \n1 \ 2\n en, good, and thank you'

In addition, in the concrete implementation, before the prompt word text is input into the AI large model, the primarily predicted type of the appeal of the text content obtained by voice recognition can be further predicted in a keyword matching mode, so that the primarily predicted type of the appeal information is further included in the prompt word text. Thus, the primarily predicted type of the appeal by keyword matching or the like can be provided to the AI large model as a reference. That is, the AI large model mainly analyzes the input information in a natural language understanding manner, but may supplement the judgment result obtained in a keyword matching manner, so that the AI large model may synthesize various information and generate more accurate content.

It should be noted that, in a specific implementation, the dialog content may be generated in real time and gradually added to the prompt word text template, where, every time a part of new dialog content is added, the AI big model may make some corresponding summary or recommend the content in the specific claim list. Of course, when the dialog content is relatively small, the summarized content or recommended content may not be accurate enough, and as new dialog content increases, the AI large model may gradually generate more accurate content.

For example, for a certain dialog process, the user says: "hello, i recently purchased a refrigerator on the same side, but i found that buying was expensive, because now there was a coupon, the price became much lower after the coupon was superimposed", at this time, the AI big model could generate recommended complaint list content according to these dialogue contents, and could pop up the prompt message of "system recommended complaint for your intelligence", and could also provide a "click view" option, after the user clicks on this option, the recommended complaint content could be presented by creating a tab page or the like, for example, as shown in fig. 4 (a), which could include:

"description of problem: how to obtain platform virtual currency

User appeal: virtual currency of the platform

The solution is as follows: obtained by participating in a specified activity, completing a task, etc.

Of course, since the content of the user statement is also limited at this time, the content of the AI large model generation may not be accurate enough. Thereafter, the user then speaks again: "I want to refund me", customer service personnel say: "good help you look at you, you operate refund directly in order, then me gives coupon to you, you can place a new order again", at this time, AI big model can regenerate recommended appeal content according to the above-mentioned dialogue content, for example, as shown in fig. 4 (B):

"description of problem: goods and money returned

User appeal: the return of goods is desired

The solution is as follows: providing a refund service.

At this time, if the customer service personnel consider that the generated content is more accurate, the "save appeal" option in the interface can be clicked directly, and correspondingly, the content generated by the large AI model can be automatically filled into the specific option box in the appeal list.

Here, the content related to the solution output by the AI large model may be summarized from the dialogue content, that is, the customer service personnel gives the solution and speaks the solution to the user in a dialogue manner, and in this process, the AI large model may summarize the solution given by the customer service personnel and output the summarized solution as a content generated by the AI large model. Alternatively, solutions recommended by AI large models may be included in another manner, and so on. The contents of the problems, the requirements and the solutions generated by the AI large model can be modified by customer service personnel, so that the contents can be displayed to the customer service personnel through the interface, and if the customer service personnel consider that the contents generated by the AI large model are inaccurate, the contents can be manually modified. Of course, the modified procedure would still be somewhat more efficient than a completely manual input. In addition, regarding the type of the complaints, since the accuracy of the prediction is generally high after the prediction by the AI large model, the complaint form can be automatically filled in without being presented to the user in the interface shown in fig. 4 (a) or 4 (B) described above.

Through the mode, in the process of voice conversation between the customer service personnel and the user, the AI large model can summarize the conversation content key points in real time or quasi-real time, and recommend specific appeal types, so that dependence on experience of the customer service personnel, manual input speed and other aspects is reduced. The content to be filled in with respect to the order typically includes user information, order type information, and the content to be filled in with respect to the problem point, the order point, the solution, and the like. In addition, the AI large model may summarize the content of the problem point, the requirement point, the solution, and the like, so if the filling is needed, the filling may be completed based on the content generated by the AI large model.

The procedure of filling out the complaint form may also be automatically completed, for example, after the customer service personnel initiates an operation request through the option of "save complaint" as in fig. 4 (a) or 4 (B), the recommended content generated by the AI large model may be parsed to generate formatted complaint information for generating the complaint form, and so on. The generated order may be pushed to a specific order table or other system for further processing by subsequent operators, including generating an order, or, in the event that no solution has been provided, may be forwarded to other customer service operators to provide a specific solution, and so on.

For example, after the order is generated, if solution information is already included therein, a specific job order may be generated according to the order so that a specific worker performs landing processing for a specific user order according to the job order. For example, if a customer service person determines that a specific solution is "help you pay" during a call with a user, the solution can be recorded by a claim list, and a special operator is subsequently used to complete the pay operation. For example, the above-mentioned claim list is first transferred to the operator who actually pays for the claim to process, the operator checks the information to determine whether the seller needs to be notified, whether the seller complains, whether the seller has a plurality of jobs such as a deposit, etc., and each job may be generated when each job is executed. After each job ticket is executed, the user's appeal is resolved.

In summary, through the embodiment of the present application, in a specific service scenario, during a voice dialogue between a customer service person and a user, a voice content generated by a call between two parties may be obtained, and then voice recognition is performed, the voice content is converted into text content, and an issuer of the voice content is identified, and issuer identification information is added to the text content. And then, according to the text content with the sender identification information and the candidate appeal type information in the specific service scene, constructing a prompt word text for carrying out dialogue with an artificial intelligence AI large language model so as to generate the content on the specific component fields in the appeal list for the voice dialogue process by the AI large language model for selection or reference by the customer service personnel. By the method, intelligent generation of the resort list content can be realized through the AI large language model, and dependence on experience of customer service personnel and capabilities of summarization, scene prediction and the like is reduced, so that efficiency is improved, and labor cost is reduced.

It should be noted that, in the embodiments of the present application, the use of user data may be involved, and in practical applications, user specific personal data may be used in the schemes described herein within the scope allowed by applicable legal regulations in the country where the applicable legal regulations are met (for example, the user explicitly agrees to the user to actually notify the user, etc.).

Corresponding to the above method embodiment, the present application further provides a recommended content generating device, referring to fig. 5, where the device may include:

a voice content obtaining unit 501, configured to obtain voice content during a call between a customer service person and a user in a target service scenario;

a voice recognition unit 502 for converting the voice content into text content and recognizing an issuer of the voice content so as to add issuer identification information to the text content;

a prompt word text construction unit 503, configured to construct a prompt word text for performing a dialogue with an artificial intelligence AI big language model according to text content with sender identification information and candidate appeal type information in the target service scene, so as to generate recommended content on at least one information element field required in a appeal list by the AI big language model.

Wherein the apparatus may further comprise:

and the text processing unit is used for carrying out word segmentation and/or sentence breaking processing on the text content before constructing the prompt word text.

In addition, the apparatus may further include:

the preliminary prediction unit is used for carrying out preliminary prediction on the type of the appeal of the text content in a keyword matching mode, and adding the preliminarily predicted type of appeal information into the prompt word text so that the AI large language model can carry out prediction on the type of appeal by combining the preliminarily predicted type of appeal information.

In addition, the recommended content generated by the AI large language model further comprises: and by summarizing the text content, the generated user problem gist descriptive content, user appeal gist descriptive content and solution content given by the customer service personnel are used for selectively filling in the appeal list.

In addition, the apparatus may further include:

and the analysis unit is used for responding to an operation request of storing the appeal list of the recommended content generated according to the AI large language model submitted by the customer service personnel, analyzing the recommended content and generating formatted appeal information so as to generate the appeal list.

In addition, the embodiment of the application further provides a computer readable storage medium, on which a computer program is stored, which when executed by a processor, implements the steps of the method of any one of the foregoing method embodiments.

And an electronic device comprising:

one or more processors; and

a memory associated with the one or more processors for storing program instructions that, when read for execution by the one or more processors, perform the steps of the method of any of the preceding method embodiments.

Fig. 6, among other things, illustrates an architecture of an electronic device, for example, device 600 may be a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, an exercise device, a personal digital assistant, an aircraft, and so forth.

Referring to fig. 6, device 600 may include one or more of the following components: a processing component 602, a memory 604, a power component 606, a multimedia component 608, an audio component 610, an input/output (I/O) interface 612, a sensor component 614, and a communication component 616.

The processing component 602 generally controls overall operation of the device 600, such as operations associated with display, telephone calls, data communications, camera operations, and recording operations. The processing component 602 may include one or more processors 620 to execute instructions to perform all or part of the steps of the methods provided by the disclosed subject matter. Further, the processing component 602 can include one or more modules that facilitate interaction between the processing component 602 and other components. For example, the processing component 602 may include a multimedia module to facilitate interaction between the multimedia component 608 and the processing component 602.

The memory 604 is configured to store various types of data to support operations at the device 600. Examples of such data include instructions for any application or method operating on device 600, contact data, phonebook data, messages, pictures, videos, and the like. The memory 604 may be implemented by any type or combination of volatile or nonvolatile memory devices such as Static Random Access Memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disk.

The power supply component 606 provides power to the various components of the device 600. The power components 606 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the device 600.

The multimedia component 608 includes a screen between the device 600 and the user that provides an output interface. In some embodiments, the screen may include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensor may sense not only the boundary of a touch or sliding action, but also the duration and pressure associated with the touch or sliding operation. In some embodiments, the multimedia component 608 includes a front camera and/or a rear camera. The front-facing camera and/or the rear-facing camera may receive external multimedia data when the device 600 is in an operational mode, such as a shooting mode or a video mode. Each front camera and rear camera may be a fixed optical lens system or have focal length and optical zoom capabilities.

The audio component 610 is configured to output and/or input audio signals. For example, the audio component 610 includes a Microphone (MIC) configured to receive external audio signals when the device 600 is in an operational mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signals may be further stored in the memory 604 or transmitted via the communication component 616. In some embodiments, audio component 610 further includes a speaker for outputting audio signals.

The I/O interface 612 provides an interface between the processing component 602 and peripheral interface modules, which may be a keyboard, click wheel, buttons, etc. These buttons may include, but are not limited to: homepage button, volume button, start button, and lock button.

The sensor assembly 614 includes one or more sensors for providing status assessment of various aspects of the device 600. For example, the sensor assembly 614 may detect the on/off state of the device 600, the relative positioning of the components, such as the display and keypad of the device 600, the sensor assembly 614 may also detect a change in position of the device 600 or a component of the device 600, the presence or absence of user contact with the device 600, the orientation or acceleration/deceleration of the device 600, and a change in temperature of the device 600. The sensor assembly 614 may include a proximity sensor configured to detect the presence of nearby objects in the absence of any physical contact. The sensor assembly 614 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor assembly 614 may also include an acceleration sensor, a gyroscopic sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

The communication component 616 is configured to facilitate communication between the device 600 and other devices, either wired or wireless. The device 600 may access a wireless network based on a communication standard, such as WiFi, or a mobile communication network of 2G, 3G, 4G/LTE, 5G, etc. In one exemplary embodiment, the communication component 616 receives broadcast signals or broadcast-related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, the communication component 616 further includes a Near Field Communication (NFC) module to facilitate short range communications. For example, the NFC module may be implemented based on Radio Frequency Identification (RFID) technology, infrared data association (IrDA) technology, ultra Wideband (UWB) technology, bluetooth (BT) technology, and other technologies.

In an exemplary embodiment, the apparatus 600 may be implemented by one or more Application Specific Integrated Circuits (ASICs), digital Signal Processors (DSPs), digital Signal Processing Devices (DSPDs), programmable Logic Devices (PLDs), field Programmable Gate Arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic elements for executing the methods described above.

In an exemplary embodiment, a non-transitory computer readable storage medium is also provided, such as memory 604, including instructions executable by processor 620 of device 600 to perform the methods provided by the disclosed subject matter. For example, the non-transitory computer readable storage medium may be ROM, random Access Memory (RAM), CD-ROM, magnetic tape, floppy disk, optical data storage device, etc.

From the above description of embodiments, it will be apparent to those skilled in the art that the present application may be implemented in software plus a necessary general purpose hardware platform. Based on such understanding, the technical solutions of the present application may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a storage medium, such as a ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to perform the methods described in the embodiments or some parts of the embodiments of the present application.

In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are all referred to each other, and each embodiment mainly describes differences from other embodiments. In particular, for a system or system embodiment, since it is substantially similar to a method embodiment, the description is relatively simple, with reference to the description of the method embodiment being made in part. The systems and system embodiments described above are merely illustrative, wherein the elements illustrated as separate elements may or may not be physically separate, and the elements shown as elements may or may not be physical elements, may be located in one place, or may be distributed over a plurality of network elements. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art will understand and implement the present invention without undue burden.

The above describes the recommended content generation method and the electronic device provided by the present application in detail, and specific examples are applied to describe the principles and embodiments of the present application, and the description of the above examples is only used to help understand the method and core ideas of the present application; also, as will occur to those of ordinary skill in the art, many modifications are possible in view of the teachings of the present application, both in the detailed description and the scope of its applications. In view of the foregoing, this description should not be construed as limiting the application.

Claims

1. A recommended content generation method, characterized by comprising:

2. The method as recited in claim 1, further comprising:

3. The method of claim 1, wherein the step of determining the position of the substrate comprises,

4. A method according to claim 3, further comprising:

5. The method of claim 1, wherein the step of determining the position of the substrate comprises,

6. The method according to any one of claim 1 to 5, wherein,

the text of the prompt word also comprises format information of output content, so that the AI large language model generates recommended content on at least one information element field required in the resort sheet according to the format information.

7. The method according to any one of claims 1 to 5, further comprising:

8. A recommended content generation device characterized by comprising:

9. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the steps of the method according to any one of claims 1 to 7.

10. An electronic device, comprising:

one or more processors; and

a memory associated with the one or more processors for storing program instructions that, when read for execution by the one or more processors, perform the steps of the method of any of claims 1 to 7.