Method for optimizing service providing mode and electronic equipment

Abstract

The embodiment of the invention discloses a method for optimizing a service providing mode and electronic equipment, and the method comprises the steps: collecting dialogue data generated between a plurality of buyers and sellers, carrying out the formatting processing of the dialogue data through taking each round of dialogue as a unit, and inputting the formatted dialogue data into an artificial intelligence AI large model for analysis; receiving question description content needing to be answered by the AI large model, and converting the question description content into a second vector; screening out at least one target information point text related to the question description content from the plurality of information point texts; and according to dialogue data corresponding to the at least one target information point text and the question description content, constructing prompt information used for carrying out dialogue with the AI large model. According to the embodiment of the invention, the method can achieve the optimization of a value-added service providing mode, and enables the value-added service to play a role in improving the conversion rate of commodities, improving the user experience, and the like more effectively.

Description

Technical Field

[0001] The present application relates to the field of information processing technology, and in particular to a method and electronic device for optimizing a service provision method. Background Art

[0002] With the rapid development of e-commerce, consumers have higher and higher requirements for shopping experience. In order to improve the shopping experience of users, some value-added services will be provided to merchants in some commodity information service systems. For merchants who have purchased such value-added services, the platform can provide a series of safeguards. For example, a platform can provide merchants with a "safe purchase" service, which packages multiple service items, including return shipping, seven-day no-reason return, etc., in order to enhance buyers' shopping confidence and satisfaction, and the corresponding expenses will be borne by the platform. For example, a merchant has opened the "safe purchase" service provided by the platform. Since the specific service items include "return shipping", after a consumer purchases a certain product from the merchant, if a return occurs, the consumer does not need to bear the shipping costs incurred during the return process, and even the merchant does not need to pay this shipping cost, but the platform will bear it. Of course, according to the pre-agreed rules, the platform may charge the merchant a certain percentage of service fees, etc. In this way, for consumers, since they do not have to bear the shipping costs incurred during the return process, their shopping risks can be reduced. Therefore, it can increase consumers' sense of trust, allowing consumers to purchase goods with greater peace of mind, thereby improving the conversion rate of goods. For merchants, since they do not have to bear all the shipping costs for each return, it is also beneficial to improve merchants' satisfaction with the platform and their willingness to launch services such as "Safe Shopping".

[0003] However, in actual applications, although value-added services such as "Safe Shopping" can improve user experience to a certain extent, there are still shortcomings in some aspects, so that merchants are not willing to open such value-added services. Correspondingly, the role of such value-added services in improving product conversion rates, user experience, etc. is difficult to play more fully. Summary of the invention

[0004] The present application provides a method and electronic device for optimizing the service provision method, which can optimize the value-added service provision method so that it can more effectively play a role in improving product conversion rate, user experience, etc.

[0005] This application provides the following solutions:

[0006] A method for optimizing a service provision method, comprising:

[0007] Collecting conversation data generated between multiple buyer and seller users, formatting the conversation data according to each round of conversation, and inputting the data into an artificial intelligence (AI) big model for analysis, so that the AI ​​big model generates corresponding information point text for each round of conversation, and converting the information point text into a first vector;

[0008] Receiving a description of a question that needs to be answered by the AI ​​big model, and converting the description of the question into a second vector; the description of the question is related to summarizing the needs or demands of the buyer and seller in the process of using the value-added service;

[0009] By calculating the distances between the second vector and the first vector respectively, at least one target information point text related to the problem description content is selected from the plurality of information point texts;

[0010] Based on the dialogue data corresponding to the at least one target information point text and the question description content, prompt information for communicating with the AI ​​big model is constructed, so that the AI ​​big model can generate content for answering the corresponding question by performing natural language understanding on the dialogue data corresponding to the target information point text, so as to optimize the way of providing value-added services according to the content generated by the AI ​​big model.

[0011] The content generated by the AI ​​big model includes: content related to the differences in the attention paid by different types of buyer users to multiple service items in the value-added service;

[0012] The optimizing the provision method of the value-added service according to the content generated by the AI ​​big model includes:

[0013] When the information of the value-added service is disclosed to the buyer user, according to the type of the buyer user, the information of the service items that the buyer user of this type is more concerned about is disclosed preferentially.

[0014] When transmitting the information of the value-added service to the buyer user, according to the type of the buyer user, the information of the service items that the buyer user of this type is more concerned about is preferentially transmitted, including:

[0015] In the process of displaying the information page of a certain product to the buyer user, if the seller user associated with the product has activated the value-added service, then based on the type of the buyer user, information on service items that the buyer user of this type is more concerned about will be displayed preferentially in the information page.

[0016] When transmitting the information of the value-added service to the buyer user, according to the type of the buyer user, the information of the service items that the buyer user of this type is more concerned about is preferentially transmitted, including:

[0017] After receiving a request from a buyer user to have a conversation with a seller user, if the seller user has activated the value-added service, information about service items that the buyer user of this type is more concerned about will be displayed preferentially in the dialog interface based on the type of the buyer user.

[0018] The content generated by the AI ​​big model includes: content related to differences in demand for multiple service items in the value-added service by seller users in different industries;

[0019] The optimizing the provision method of the value-added service according to the content generated by the AI ​​big model includes:

[0020] When providing the information of the value-added service to the seller user, the service items that the seller user needs more are provided preferentially among the value-added services according to the industry to which the seller user belongs.

[0021] Among them, it also includes:

[0022] Service items that are not needed or have low demand from seller users are deleted from the value-added services so as to reduce the service price of the value-added services.

[0023] The content generated by the AI ​​big model includes: questions commonly held by buyer users related to the value-added service;

[0024] The optimizing the provision method of the value-added service according to the content generated by the AI ​​big model includes:

[0025] Add explanatory information corresponding to the question to the information page of the value-added service.

[0026] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of any of the methods described above.

[0027] An electronic device, comprising:

[0028] one or more processors; and

[0029] A memory associated with the one or more processors, the memory being used to store program instructions, wherein the program instructions, when read and executed by the one or more processors, execute the steps of any of the methods described above.

[0030] A computer program product comprises a computer program / computer executable instructions, wherein the computer program / computer executable instructions implement the steps of any of the aforementioned methods when executed by a processor in an electronic device.

[0031] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0032] Through the embodiments of the present application, advanced AI big model technology can be used to conduct in-depth analysis of the conversation data between buyers and sellers generated in the system. Since the AI ​​big model can understand complex language structures and their contexts, it can more accurately capture the key information and implicit intentions in the buyer-seller conversations, and generate answers to questions related to summarizing the needs or demands of buyers and sellers in the process of using the value-added services, thereby providing strong support for optimizing product design. Among them, since the number of dialogue data collected is usually relatively large, in order to meet the AI ​​big model's limit on the length of input word units, the dialogue data can first be formatted according to each round of dialogue and then input into the artificial intelligence AI big model for analysis, so that the AI ​​big model generates corresponding information point text for each round of dialogue, and converts the information point text into a first vector. After receiving the description of the question that needs to be answered by the AI ​​big model, the question description content can be converted into a second vector. After that, at least one target information point text related to the question description content can be screened out from the multiple information point texts by calculating the distance between the second vector and the first vector respectively, and according to the dialogue data corresponding to the at least one target information point text and the question description content, a prompt information for dialogue with the AI ​​big model is constructed, so that the AI ​​big model generates content for answering the corresponding question by performing natural language understanding on the dialogue data corresponding to the target information point text. In this way, the length of the data input into the AI ​​big model can meet the requirements of the AI ​​big model for the length of input word units.

[0033] Of course, any product implementing the present application does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] 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 will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0035] Figure 1 It is a schematic diagram of the system architecture provided by the embodiment of the present application;

[0036] Figure 2 is a flow chart of the method provided in the embodiment of the present application;

[0037] Figure 3 is a schematic diagram of a first word cloud diagram provided in an embodiment of the present application;

[0038] Figure 4 is a schematic diagram of a second word cloud diagram provided in an embodiment of the present application;

[0039] Figure 5 It is a schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0040] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments in the present application belong to the scope of protection of this application.

[0041] First of all, it should be noted that the inventors of this application found in the process of implementing the embodiments of this application that the deficiencies of the value-added services such as "safe shopping" provided by the platform to merchants are mainly reflected in the following aspects:

[0042] 1. Existing value-added services usually package multiple service items and provide them to merchants in the form of "packages". In addition, the service items in the "packages" provided to all merchants are the same, which means that all merchants need to pay for all the service items in the "package" before they can activate the value-added service. However, some merchants may only need to use some of the service items, while other service items are rarely or never used. If all these service items are packaged together and the merchants pay the service fee, the merchants' willingness to pay may be affected. For example, the "Safe Shopping" service provides the "Return with Shipping Included" service to all merchants. However, suppose a merchant mainly sells handicrafts, and the orders of this type of merchant are characterized by the fact that there may be many products in one order and the amount is large. If buyers encounter problems during the purchase process, they generally tend to pay compensation or exchange the goods, and rarely use the return service. At this time, if the "Return with Shipping Included" service is provided to this type of merchant, the "Safe Shopping" service will not be attractive to this merchant because there is no such demand. Or for this type of merchant, the service fee of the "Safe Shopping" service includes the cost of the "Return with Shipping Included" service, which will lead to the merchant's lack of willingness to open the corresponding service.

[0043] 2. After a merchant has opened a value-added service such as "Safe Shopping", the system can provide consumers with the "Safe Shopping" logo when displaying information about specific products associated with the merchant to consumers, or when consumers initiate a dialogue request with the merchant, so that consumers have more trust in specific products or merchants. However, since the "Safe Shopping" service may include multiple different service items, and different consumers may pay different attention to different service items, if the "Safe Shopping" logo is provided in general, the role of the logo may be relatively limited.

[0044] 3. Merchants or consumers may have some common questions when using specific value-added services, such as specific fee-bearing rules, service scope, etc. In the existing technology, these questions may need to be resolved through communication with the platform customer service, but this will cause a huge workload for the platform customer service.

[0045] In order to solve the above problems, it is necessary to obtain the following information: which services are in demand by merchants in different industries, and which services are not in demand or have very low demand; which services are of greater concern to different types of consumers; what common questions merchants or consumer users have when using value-added services, etc.

[0046] In order to obtain the above information, one way is to configure it through manual review or customer service feedback. For example, in some cases, manual review can be used to collect user feedback and solve problems through the customer service team. This method can obtain high-quality feedback information and handle complex user problems. However, there are high labor costs, difficulty in scalability, slow feedback processing speed, inability to respond to user needs in real time, and consistency issues (different customer service personnel may have different processing methods, etc.), etc.

[0047] In order to overcome the above problems, another way is to obtain it through data mining. Specifically, since the commodity information service system usually provides a dialogue consultation system for buyers and sellers, the buyer user can initiate consultations with the seller user on issues arising from the purchase, order placement, after-sales, etc. During this process, the system will record the dialogue data between the two parties, and the above information can be obtained by analyzing and mining these dialogue data. Therefore, the dialogue data between buyers and sellers can be collected and the above information can be obtained from these dialogue data.

[0048] In order to analyze and mine the conversation data between buyers and sellers, one way is to use a rule-based text analysis system to analyze such conversation data, that is, to use a rule-based method to identify specific information in the text through predefined keywords and phrases. This method is easy to implement and does not require a large amount of training data. However, its disadvantage is that it cannot capture complex user intentions or contextual information, and the rules need to be maintained and updated manually, which is difficult to adapt to rapidly changing user needs and market trends. In addition, the processing ability for non-standardized text is limited, and important information is easily missed. Another way is to use traditional machine learning algorithms (such as support vector machines, decision trees, etc.) to classify or regress text data. This method is a relatively mature technology and is suitable for structured data. However, it requires manual feature extraction, and feature engineering is complex and time-consuming. The performance of natural language processing tasks is limited, especially in processing long texts or context understanding. In addition, it may not capture the subtle differences in the text, resulting in low accuracy.

[0049] With the emergence of AI (Artificial Intelligence) models, especially AI large-scale parameter models (AI big models for short), they have been applied in multiple scenarios and also provide a new idea for conversation data analysis. The so-called AI big model refers to a deep learning model containing a large number of parameters. Due to its large scale, this AI big model can store and process a large amount of information, thereby achieving higher performance in various tasks. Therefore, relying on the powerful natural language text understanding and logical thinking ability of the pre-trained AI big model, the conversation data between buyers and sellers can be analyzed, and a variety of conclusive information content required for the optimization of value-added services can be generated.

[0050] However, when using the AI ​​big model to process the conversation data between the buyer and seller, a problem may be encountered: in order to generate the various conclusive information content required in the value-added service optimization process, it is usually necessary to collect a large amount of conversation content generated in the system over a period of time and analyze it by the AI ​​big model. This results in that there may be a lot of conversation data collected, for example, it may include tens of millions of conversation data, etc.; however, the input token length supported by the AI ​​big model is usually limited, and all the collected conversation data cannot be directly input into the AI ​​big model for analysis.

[0051] In view of the above problems, the embodiments of the present application provide corresponding solutions. In this solution, after collecting multiple dialogue data, the dialogue data can first be formatted in units of each round of dialogue, and then input into the AI ​​large model for analysis, so as to convert each round of dialogue into one or more information points for expression. Among them, each round of dialogue can refer to a dialogue between a buyer and a seller on a topic, etc., which may have two sentences or many sentences. For example, a buyer user initiates a round of dialogue, and the two parties end the dialogue after multiple dialogue contents. The system may also provide the buyer user with an evaluation option when it is determined that the dialogue has ended, for evaluating the service of the seller's customer service. At this time, the operation of the buyer user making an evaluation can also be regarded as the end of a round of dialogue, etc. Among them, since a round of dialogue may include multiple dialogue contents, it may only include a few main information points, such as what questions the buyer raised, whether there is a dispute between the two parties, and what solutions the seller provides. Therefore, in an embodiment of the present application, the multiple dialogue data of each round of dialogue can first be analyzed using the AI ​​big model to summarize the information points included therein. In this way, each round of dialogue can be expressed through shorter text, thereby reducing the amount of dialogue data.

[0052] However, after aggregating each round of dialogue as described above, the amount of data processed may still exceed the AI ​​big model's limit on the maximum input token length, so further processing can be performed. Specifically, the information points corresponding to each round of dialogue can be mathematically expressed through processing methods such as Embedding to obtain multiple vectors (referred to as "first vectors" here for ease of distinction). In addition, the description of the question that needs to be answered by the AI ​​big model in the future can also be mathematically expressed to obtain the corresponding vector, and accordingly, the vector can be called the "second vector". Then, the distance between the second vector and the multiple first vectors can be calculated, and the distance can represent the correlation between the information point of the specific dialogue and the above-mentioned question description problem. Therefore, through distance calculation, at least one target information point text related to the question description content can be screened out from the multiple information point texts. Then, the original text of the dialogue data corresponding to the screened target information point text and the aforementioned question description information can be used together to construct a prompt message and input it into the AI ​​big model. In this way, the AI ​​big model can understand the dialogue data corresponding to the target information point text in natural language and generate content for answering the corresponding question. For example, the specific question description information may be "What are the demands of merchants in the clothing industry when using the 'Safe Shopping' service?" After filtering out the conversation data related to the question through vector matching, the length of the input data can be effectively shortened to meet the AI ​​big model's requirements for the maximum length of the input token. Then, the AI ​​big model can output the content used to answer the above question based on the specific input data.

[0053] After obtaining the answer content generated by the AI ​​big model, the provision of value-added services such as "Safe Shopping" can be optimized based on the content generated by the AI ​​big model. Among them, there are many specific optimization methods. For example, for seller users, the AI ​​big model can be used to summarize the specific demands of merchants in multiple different industries when using specific value-added services, and then a more personalized service item combination can be provided to seller users in different industries in the "package". For example, the service items that seller users in a specific industry need more can be provided to seller users first, making them more attractive to seller users. In addition, service items that seller users in a certain industry will not use or rarely use can be deleted from the "package", so as to reduce the price of the "package" and make seller users more willing to activate specific value-added services, and so on.

[0054] In addition, for buyer users, AI big models can also be used to discover the differences in the needs of different types of buyer users for value-added services, and provide more personalized information about specific service items to different types of buyer users. For example, buyer users can be divided into Class B buyers and Class C buyers. Class B buyers usually refer to bulk buyers or corporate buyers, who purchase large quantities of goods and may use them for commercial purposes, such as wholesale, resale, etc.; Class C buyers refer to individual consumers or terminal buyers, who purchase goods mainly for personal consumption. For different types of buyers, the specific needs may be different when using value-added services. For example, Class B buyers may pay more attention to batch issues and customized needs, and the solutions tend to be bulk replacement, customized modification, etc., while Class C buyers pay more attention to personal consumption experience, such as color and style discrepancies, and the solutions tend to be the return and exchange of individual goods, personalized compensation plans, etc. Therefore, when the information of value-added services is revealed through the product information page or the dialogue interface between buyers and sellers, it can be specific to the service items that buyer users are more concerned about, so as to further improve the user experience.

[0055] Furthermore, the AI ​​big model can be used to summarize the common questions that buyers or sellers have about value-added services, such as the scope of service, freight payment rules, etc. These questions can be answered by providing explanatory text in the relevant interface of the value-added service, avoiding the platform customer service frequently explaining the same questions to different users, etc.

[0056] From the perspective of system architecture, see Figure 1 , the embodiment of the present application can provide a system for optimizing the way of providing value-added services. The system can collect the conversation data between the buyer and the seller, and after sorting it out in units of each round of conversation, input it into the AI ​​big model for information aggregation, and convert the multiple conversation data contained in each round of conversation into a more concise way of expression through information points, and convert it into a first vector. After that, the question description information used to ask the AI ​​big model in the future can also be expressed as a second vector, and the conversation data related to the specific question can be screened out by the distance between the vectors, and constructed as prompt information together with the question description information, so as to be input into the AI ​​big model for analysis and summary, and then based on the answer results of the AI ​​big model, the way of providing value-added services is optimized. Specifically, it can include personalized processing of the service items included in the "package" for seller users in different industries, and for different types of buyer users, when transmitting information such as labels of value-added services, differentiated expressions can be made in the service item dimension, and information about service items that buyer users are more concerned about can be given priority to improve user experience, and so on.

[0057] The specific implementation scheme provided in the embodiments of the present application is described in detail below.

[0058] First, the present application embodiment provides a method for optimizing the service provision mode, see Figure 2 , the method may include:

[0059] S201: Collecting conversation data generated between multiple buyer and seller users, formatting the conversation data in units of each conversation round, and inputting the data into an artificial intelligence (AI) big model for analysis, so that the AI ​​big model generates corresponding information point text for each conversation round, and converts the information point text into a first vector.

[0060] In specific implementation, the conversation records between buyers and sellers in the system regarding returns, refunds or other related issues in the past six months (or other time lengths) can be collected to obtain specific conversation data. In an optional way, the context information of the specific conversation data can also be obtained, such as information about related products, information about related orders, and so on.

[0061] After the conversation data is collected, the conversation data can be formatted in units of each round of conversation and then input into the artificial intelligence AI model for analysis, so that the AI ​​model generates corresponding information point text for each round of conversation, thereby realizing the aggregation of specific conversation data through information points. For example, a conversation between user A and merchant B and a conversation between user C and merchant D have the same main information points, which can be aggregated into the same information point text for expression, and so on. Afterwards, multiple information point texts can be converted into multiple first vectors. Specifically, when generating vectors, it can be achieved by means such as Embedding, which will not be described in detail here.

[0062] Before the conversation data of each round of conversation is input into the AI ​​big model for aggregation, the conversation data of each round of conversation can also be formatted so that the specific data can be expressed in a unified format. For example, the conversation data can be expressed in the format of the initiator, conversation content, etc. corresponding to each conversation data, and each conversation data can be separated by a separator such as a semicolon, etc.

[0063] S202: Receive a question description that needs to be answered by the AI ​​big model, and convert the question description into a second vector; the question description is related to summarizing the needs or demands of the buyer and seller in the process of using the value-added service.

[0064] After the first vector is converted, the question description content that needs to be answered by the AI ​​big model can also be received. This question description content is related to summarizing the needs or demands of the buyer and seller in the process of using the value-added service, that is, the specific instructions issued to the AI ​​big model. Of course, this instruction can be expressed through natural language text content. After that, the question description content can be converted into a second vector.

[0065] It should be noted that, in the actual implementation, there may be multiple questions that need to be answered by the AI ​​large model. Therefore, there may be multiple specific question descriptions, which can be converted into different second vectors respectively.

[0066] S203: Filter out at least one target information point text related to the question description content from the plurality of information point texts by calculating the distances between the second vector and the first vector respectively.

[0067] After obtaining the second vector and multiple first vectors, the distance between the second vector and the multiple first vectors can be calculated respectively. Then, at least one target information point text related to the problem description content can be screened out from the multiple information point texts based on the distance calculation results.

[0068] S204: Construct prompt information for communicating with the AI ​​big model based on the dialogue data corresponding to the at least one target information point text and the question description content, so that the AI ​​big model can generate content for answering the corresponding question by performing natural language understanding on the dialogue data corresponding to the target information point text, so as to optimize the way of providing value-added services according to the content generated by the AI ​​big model.

[0069] After obtaining at least one target information point text, the original dialogue data corresponding to these target information point texts can be determined, and then together with the question description content, a prompt for interacting with the AI ​​big model can be constructed, and then it can be input into the AI ​​big model, so that the AI ​​big model can understand the dialogue data corresponding to the target information point text in natural language and generate content for answering the corresponding question. Among them, depending on the specific question description content, the specific selected dialogue data will be different, and accordingly, the AI ​​big model will also give different answer content accordingly.

[0070] After obtaining the content generated by the AI ​​big model, the provision method of the value-added service can be optimized according to these contents. Among them, the specific optimization method can have multiple aspects. For example, the content generated by the AI ​​big model can include: content related to the difference in the attention of different types of buyer users to multiple service items in the value-added service. At this time, when optimizing, it can be that when the information of the value-added service is revealed to the buyer user, according to the type of the buyer user, the information of the service items that the buyer user of this type is more concerned about is preferentially revealed. Specifically, it can be in the process of showing the information page of a certain product (for example, the product details page, etc.) to the buyer user, if the seller user associated with the product has opened the value-added service, then according to the type of the buyer user, the information of the service items that the buyer user of this type is more concerned about is preferentially displayed in the information page. Alternatively, after receiving a request initiated by the buyer user to have a conversation with the seller user, if the seller user has opened the value-added service, it can also be based on the type of the buyer user. The type of service items that the buyer user of this type is more concerned about can be preferentially displayed in the dialog interface. Among them, when displaying the information of specific service items in the dialog box interface, it can be displayed in the form of information cards, etc. By clicking on the information card, the user can view the detailed information of the specific service item, which may include explanatory information about the specific service scope, fee-bearing rules, etc.

[0071] Alternatively, if the content generated by the AI ​​big model includes: content related to the differences in the demands of seller users in different industries for multiple service items in the value-added service. Then the specific optimization method at this time may be, when providing the information of the value-added service to the seller user, according to the industry to which the seller user belongs, giving priority to providing the service items that the seller user needs more in the value-added service, thereby improving the seller user's satisfaction with the value-added service. Alternatively, in a preferred implementation, service items that the seller user does not need or has a low demand may also be deleted from the value-added service, so as to reduce the service price of the value-added service.

[0072] In addition, the content generated by the AI ​​big model can also include: the buyer users’ common questions related to the value-added service. For example, the sellers’ most concerned issues include the exit process, fee deduction rules, the impact of the service on the store, etc. The buyers’ concerns focus on whether the order enjoys return shipping, the scope of service, whether it affects the price of the goods, etc. At this time, the specific optimization method can be to add explanatory information corresponding to the questions in the information page of the value-added service. Among them, for the most feedback from sellers, such as excessive fees, you can consider adjusting the fee structure and providing more attractive price packages; or, for the most feedback from buyers about the scope of services, you can reduce misunderstandings by optimizing the description of the service page, simplify the return process, clarify the rules for bearing freight during the return process, and improve service transparency, etc.

[0073] Through the above methods, the AI ​​big model can be used to analyze the behavioral differences between category B buyers and category C buyers, and the differences in demands of sellers in different industries, so that different optimization strategies can be proposed. This differentiated service can better meet the needs of different types of buyers and sellers, and improve the effectiveness and satisfaction of the service.

[0074] In addition, in specific implementation, the AI ​​big model can also generate word cloud diagrams corresponding to different types of questions based on the frequency of different phrases in the buyer-seller dialogue. For example, for the question of "buyer-seller dispute points", the specific word cloud diagram can be as follows: Figure 3 As shown, the larger the font size, the higher the probability of occurrence of the term. For example, Figure 3 The example shown shows that the disputes between buyers and sellers are more concentrated on shipping costs, refunds, returns, etc. Or, for the question "Solutions provided by sellers", the specific word cloud diagram can be as follows Figure 4 As shown in the figure, the solutions provided by sellers are more focused on "refund", "return", "shipping", etc. By identifying common points of dispute between buyers and sellers and the solutions provided by sellers, it is helpful to gain a deeper understanding of buyers' concerns and sellers' methods of dealing with problems, providing a basis for further product optimization.

[0075] In summary, through the embodiments of the present application, advanced AI big model technology can be used to conduct in-depth analysis of the conversation data between buyers and sellers generated in the system. Since the AI ​​big model can understand complex language structures and their contexts, it can more accurately capture the key information and implicit intentions in the buyer-seller conversations, and generate answers to questions related to summarizing the needs or demands of buyers and sellers in the process of using the value-added services, thereby providing strong support for optimizing product design. Among them, since the number of dialogue data collected is usually relatively large, in order to meet the AI ​​big model's limit on the length of input word units, the dialogue data can first be formatted according to each round of dialogue and then input into the artificial intelligence AI big model for analysis, so that the AI ​​big model generates corresponding information point text for each round of dialogue, and converts the information point text into a first vector. After receiving the description of the question that needs to be answered by the AI ​​big model, the question description content can be converted into a second vector. After that, at least one target information point text related to the question description content can be screened out from the multiple information point texts by calculating the distance between the second vector and the first vector respectively, and according to the dialogue data corresponding to the at least one target information point text and the question description content, a prompt information for dialogue with the AI ​​big model is constructed, so that the AI ​​big model generates content for answering the corresponding question by performing natural language understanding on the dialogue data corresponding to the target information point text. In this way, the length of the data input into the AI ​​big model can meet the requirements of the AI ​​big model for the length of input word units.

[0076] It should be noted that the embodiments of the present application may involve the use of user data. In actual applications, user-specific personal data can be used in the scheme described herein within the scope permitted by applicable laws and regulations, subject to the requirements of applicable laws and regulations of the country where the user is located (for example, with the user's explicit consent, effective notification to the user, etc.).

[0077] Corresponding to the above method embodiment, the embodiment of the present application further provides a device for optimizing the service provision mode, which may include:

[0078] An information point text generation and conversion unit, used to collect the conversation data generated between multiple buyer and seller users, and format the conversation data according to each round of conversation as a unit, and then input it into the artificial intelligence AI big model for analysis, so that the AI ​​big model generates corresponding information point text for each round of conversation, and converts the information point text into a first vector;

[0079] A question description content conversion unit, configured to receive question description content that needs to be answered by the AI ​​big model, and convert the question description content into a second vector; the question description content is related to summarizing the needs or demands of the buyer and seller in the process of using the value-added service;

[0080] a target text determination unit, configured to select at least one target information point text related to the question description content from the plurality of information point texts by calculating the distances between the second vector and the first vector respectively;

[0081] A prompt information construction unit is used to construct prompt information for communicating with the AI ​​big model based on the dialogue data corresponding to the at least one target information point text and the question description content, so that the AI ​​big model can generate content for answering the corresponding question by performing natural language understanding on the dialogue data corresponding to the target information point text, so as to optimize the way of providing value-added services according to the content generated by the AI ​​big model.

[0082] The content generated by the AI ​​big model includes: content related to the differences in the attention paid by different types of buyer users to multiple service items in the value-added service;

[0083] Specifically, when optimizing the provision of value-added services based on the content generated by the AI ​​big model, it includes:

[0084] When the information of the value-added service is disclosed to the buyer user, according to the type of the buyer user, the information of the service items that the buyer user of this type is more concerned about is disclosed preferentially.

[0085] Specifically, in the process of displaying the information page of a certain product to the buyer user, if the seller user associated with the product has activated the value-added service, then based on the type of the buyer user, information on service items that the buyer user of this type is more concerned about will be displayed preferentially in the information page.

[0086] Alternatively, after receiving a request from a buyer user to have a conversation with a seller user, if the seller user has activated the value-added service, information about service items that the buyer user of this type is more concerned about will be displayed preferentially in the dialog interface based on the type of the buyer user.

[0087] The content generated by the AI ​​big model includes: content related to differences in demand for multiple service items in the value-added service by seller users in different industries;

[0088] Specifically, when optimizing the provision method of value-added services based on the content generated by the AI ​​big model, it may include:

[0089] When providing the information of the value-added service to the seller user, the service items that the seller user needs more are provided preferentially among the value-added services according to the industry to which the seller user belongs.

[0090] In addition, the device may also include:

[0091] The deleting unit is used to delete service items that are not needed or have low demand from the value-added services, so as to reduce the service price of the value-added services.

[0092] The content generated by the AI ​​big model includes: questions commonly held by buyer users related to the value-added service;

[0093] Specifically, when optimizing the provision method of value-added services based on the content generated by the AI ​​big model, it may include:

[0094] Add explanatory information corresponding to the question to the information page of the value-added service.

[0095] In addition, an embodiment of the present application further provides a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the steps of any one of the methods in the aforementioned method embodiments are implemented.

[0096] And an electronic device, comprising:

[0097] one or more processors; and

[0098] A memory associated with the one or more processors, the memory being used to store program instructions, wherein the program instructions, when read and executed by the one or more processors, execute the steps of the method described in any one of the aforementioned method embodiments.

[0099] A computer program product includes a computer program / computer executable instructions, which, when executed by a processor in an electronic device, implement the steps of the method described in the aforementioned method embodiment.

[0100] in, Figure 5 The architecture of the electronic device is shown as an example, which may include a processor 510, a video display adapter 511, a disk drive 512, an input / output interface 513, a network interface 514, and a memory 520. The processor 510, the video display adapter 511, the disk drive 512, the input / output interface 513, the network interface 514, and the memory 520 may be communicatively connected via a communication bus 530.

[0101] Among them, the processor 510 can be implemented by a general-purpose CPU (Central Processing Unit, processor), a microprocessor, an application specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, etc., to execute relevant programs to implement the technical solution provided in this application.

[0102] The memory 520 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 520 can store an operating system 521 for controlling the operation of the electronic device 500, and a basic input and output system (BIOS) for controlling the low-level operation of the electronic device 500. In addition, a web browser 523, a data storage management system 524, and a service optimization processing system 525, etc. can also be stored. The above-mentioned service optimization processing system 525 can be an application program that specifically implements the operations of the aforementioned steps in the embodiment of the present application. In short, when the technical solution provided by the present application is implemented by software or firmware, the relevant program code is stored in the memory 520 and is called and executed by the processor 510.

[0103] The input / output interface 513 is used to connect the input / output module to realize information input and output. The input / output module can be configured in the device as a component (not shown in the figure), or it can be externally connected to the device to provide corresponding functions. The input device may include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device may include a display, a speaker, a vibrator, an indicator light, etc.

[0104] The network interface 514 is used to connect to a communication module (not shown) to realize communication interaction between the device and other devices. The communication module can realize communication through a wired mode (such as USB, network cable, etc.) or a wireless mode (such as mobile network, WIFI, Bluetooth, etc.).

[0105] The bus 530 comprises a pathway for transmitting information between the various components of the device (eg, the processor 510, the video display adapter 511, the disk drive 512, the input / output interface 513, the network interface 514, and the memory 520).

[0106] It should be noted that, although the above device only shows a processor 510, a video display adapter 511, a disk drive 512, an input / output interface 513, a network interface 514, a memory 520, a bus 530, etc., in the specific implementation process, the device may also include other components necessary for normal operation. In addition, it can be understood by those skilled in the art that the above device may also only include components necessary for implementing the solution of the present application, and does not necessarily include all the components shown in the figure.

[0107] It can be known from the description of the above implementation methods that those skilled in the art can clearly understand that the present application can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application can be essentially or partly contributed to the prior art in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments of the present application or certain parts of the embodiments.

[0108] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can refer to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiment. The system and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without creative work.

[0109] The above is a detailed introduction to the method and electronic device for optimizing the service provision mode provided by the present application. The principles and implementation methods of the present application are described in detail using specific examples. The description of the above embodiments is only used to help understand the method and core ideas of the present application. At the same time, for those skilled in the art, according to the ideas of the present application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present application.

Claims

1. A method for optimizing a service provision method, characterized in that: include: Collecting conversation data generated between multiple buyer and seller users, formatting the conversation data according to each round of conversation, and inputting the data into an artificial intelligence (AI) big model for analysis, so that the AI ​​big model generates corresponding information point text for each round of conversation, and converting the information point text into a first vector; Receiving a description of a question that needs to be answered by the AI ​​big model, and converting the description of the question into a second vector; the description of the question is related to summarizing the needs or demands of the buyer and seller in the process of using the value-added service; By calculating the distances between the second vector and the first vector respectively, at least one target information point text related to the problem description content is selected from the plurality of information point texts; According to the dialogue data corresponding to the at least one target information point text and the question description content, prompt information for communicating with the AI ​​big model is constructed, so that the AI ​​big model can generate content for answering the corresponding question by performing natural language understanding on the dialogue data corresponding to the target information point text, so as to optimize the way of providing value-added services according to the content generated by the AI ​​big model.

2. The method according to claim 1, characterized in that The content generated by the AI ​​big model includes: content related to the differences in the attention paid by different types of buyer users to multiple service items in the value-added service; The optimizing the provision method of the value-added service according to the content generated by the AI ​​big model includes: When the information of the value-added service is disclosed to the buyer user, according to the type of the buyer user, the information of the service items that the buyer user of this type is more concerned about is disclosed preferentially.

3. The method according to claim 2, characterized in that When transmitting the information of the value-added service to the buyer user, according to the type of the buyer user, the information of the service items that the buyer user of this type is more concerned about is preferentially transmitted, including: In the process of displaying the information page of a certain product to the buyer user, if the seller user associated with the product has activated the value-added service, then based on the type of the buyer user, information on service items that the buyer user of this type is more concerned about will be displayed preferentially in the information page.

4. The method according to claim 2, characterized in that: When transmitting the information of the value-added service to the buyer user, according to the type of the buyer user, the information of the service items that the buyer user of this type is more concerned about is preferentially transmitted, including: After receiving a request from a buyer user to have a conversation with a seller user, if the seller user has activated the value-added service, information about service items that the buyer user of this type is more concerned about will be displayed preferentially in the dialog interface based on the type of the buyer user.

5. The method according to claim 1, characterized in that The content generated by the AI ​​big model includes: content related to differences in demand for multiple service items in the value-added service by seller users in different industries; The optimizing the provision method of the value-added service according to the content generated by the AI ​​big model includes: When providing the information of the value-added service to the seller user, the service items that the seller user needs more are provided preferentially among the value-added services according to the industry to which the seller user belongs.

6. The method according to claim 1, characterized in that Also includes: Service items that are not needed or have low demand from seller users are deleted from the value-added services so as to reduce the service price of the value-added services.

7. The method according to claim 1, characterized in that The content generated by the AI ​​big model includes: the buyer users’ common questions related to the value-added service; The optimizing the provision method of the value-added service according to the content generated by the AI ​​big model includes: Add explanatory information corresponding to the question to the information page of the value-added service.

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

9. An electronic device, characterized in that: include: one or more processors; as well as A memory associated with the one or more processors, the memory being used to store program instructions, wherein the program instructions, when read and executed by the one or more processors, execute the steps of the method described in any one of claims 1 to 7.

10. A computer program product comprising a computer program / computer executable instructions, characterized in that: When the computer program / computer executable instructions are executed by a processor in an electronic device, the steps of the method according to any one of claims 1 to 7 are implemented.

Images