WO2022267176A1

WO2022267176A1 - Product recommendation method and apparatus based on artificial intelligence, and device and storage medium

Info

Publication number: WO2022267176A1
Application number: PCT/CN2021/109363
Authority: WO
Inventors: 林卫旋
Original assignee: 未鲲(上海)科技服务有限公司
Priority date: 2021-06-23
Filing date: 2021-07-29
Publication date: 2022-12-29
Also published as: CN113393300B; CN113393300A

Abstract

A product recommendation method and apparatus based on artificial intelligence, and a device and a storage medium. The method comprises: by means of a target product recommendation path and product recommendation configuration data, sending a first pop-up window request to each terminal corresponding to a customer data set which is to be subjected to product recommendation and is determined according to customer screening condition data, and acquiring a pop-up window browsing result; performing speech outbound calling and intention recognition on a customer corresponding to a pop-up window browsing result indicating that a pop-up window has been browsed, so as obtain each intention recognition result corresponding to a browsed customer data set; sending a second pop-up window request to a terminal which corresponds to each piece of customer data, an intention recognition result of which is willing to accept recommendation, and acquiring a service operation result; and according to each service operation result, each intention recognition result and each pop-up window browsing result, determining a target product recommendation result corresponding to each piece of customer data. The automatic product recommendation for batches of customers is realized, and the effect of product recommendation is tracked in a timely manner.

Description

Product recommendation method, device, equipment and storage medium based on artificial intelligence

This application claims the priority of the Chinese patent application submitted to the China Patent Office on June 23, 2021, with the application number 202110700050.0, and the title of the invention is "artificial intelligence-based product recommendation method, device, equipment and storage medium", the entire content of which Incorporated in this application by reference.

technical field

The present application relates to the technical field of artificial intelligence, in particular to an artificial intelligence-based product recommendation method, device, equipment and storage medium.

Background technique

In order to discover the needs of consumers, the inventor realized that the traditional way of sending short messages by the system for product recommendation is difficult to track the progress of product recommendation when there are many customers, and it is easy to cause information harassment to customers and reduce customer experience.

technical problem

It aims to solve the technical problem that the existing technology adopts the system to send short messages to recommend products, it is difficult to track the progress of product recommendation when there are many customers, and it is easy to cause information harassment to customers and reduce customer experience.

technical solution

The main purpose of this application is to provide a product recommendation method, device, equipment and storage medium based on artificial intelligence, aiming to solve the problem that the existing technology uses the system to send short messages for product recommendation, and the progress of product recommendation is difficult when there are many customers. Tracking, and technical problems that easily cause information harassment to customers and reduce customer experience.

In order to achieve the above-mentioned purpose of the invention, the present application proposes an artificial intelligence-based product recommendation method, the method comprising:

Obtaining a product recommendation start request, the product recommendation start request carrying a target product recommendation path and product recommendation configuration data;

Through the customer screening node of the target product recommendation path, the customer data set to be recommended is obtained according to the customer screening condition data of the product recommendation configuration data;

Through the first pop-up window node of the target product recommendation path, according to the product recommendation configuration data, respectively send a first pop-up window request to the terminal corresponding to each customer data in the customer data set to be recommended by the product;

Obtain the pop-up browsing results sent by each terminal according to the first pop-up request through the first pop-up node;

When the browsing result of the pop-up window is browsed, each of the customer data whose browsing result of the pop-up window is browsed is used as the browsed customer data set;

Through the voice outbound node of the target product recommendation path, according to the product recommendation configuration data, respectively perform voice outbound calls and intention recognition for customers corresponding to each of the customer data in the browsed customer data set, Obtain the intention recognition results corresponding to each of the customer data in the browsed customer data set;

When the intention recognition result is willing to accept the recommendation, through the second pop-up window node of the target product recommendation path, according to the product recommendation basic information of the product recommendation configuration data, the intention recognition result is willing to accept the recommendation respectively. The terminal corresponding to each recommended customer data sends a second pop-up window request;

Obtain the service operation results corresponding to each of the second pop-up requests through the service operation node of the target product recommendation path;

Through the result confirmation node of the target product recommendation path, each of the customer data sets to be recommended is determined according to each of the service operation results, each of the intent recognition results, and each of the pop-up window browsing results. The target product recommendation results corresponding to the customer data.

The present application also proposes an artificial intelligence-based product recommendation device, which includes:

The request obtaining module is used to obtain a product recommendation start request, and the product recommendation start request carries a target product recommendation path and product recommendation configuration data;

The customer data set determination module to be recommended is used to obtain the customer data set to be recommended according to the customer filter condition data of the product recommendation configuration data through the customer screening node of the target product recommendation path;

The first pop-up window request sending module is used to correspond to each customer data in the customer data set to be recommended according to the product recommendation configuration data through the first pop-up window node of the target product recommendation path The terminal sends the first pop-up window request;

A pop-up browsing result acquisition module, configured to obtain, through the first pop-up node, the pop-up browsing results sent by each terminal according to the first pop-up request;

The browsed customer data set determination module is used to use each of the customer data whose pop-up browsing result is browsed as the browsed customer data set when there is a browsed result of the pop-up window;

The intention recognition result determination module is used to pass through the voice outbound node of the target product recommendation path, and according to the product recommendation configuration data, separately identify the customer corresponding to each of the customer data in the browsed customer data set performing voice outbound calls and intent recognition, and obtaining intent recognition results corresponding to each of the customer data in the browsed customer data set;

The second pop-up window request sending module is used to pass the second pop-up window node of the target product recommendation path, according to the product recommendation basic information of the product recommendation configuration data, when the intent recognition result is willing to accept the recommendation, sending a second pop-up window request to the terminals corresponding to each of the customer data for which the intention identification result is willing to accept the recommendation;

The service operation result determination module is used to obtain the service operation results corresponding to each of the second pop-up window requests through the service operation node of the target product recommendation path;

A target product recommendation result determination module, configured to determine the product to be recommended according to each of the service operation results, each of the intent recognition results, and each of the pop-up window browsing results through the result confirmation node of the target product recommendation path The target product recommendation results corresponding to each of the customer data in the customer data set.

The present application also proposes a computer device, including a memory and a processor, the memory stores a computer program, and the processor implements the following method steps when executing the computer program:

The present application also proposes a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the following method steps are implemented:

Beneficial effect

The artificial intelligence-based product recommendation method, device, equipment, and storage medium of the present application first pass through the customer screening node of the target product recommendation path, and obtain the customer data set to be recommended according to the customer screening condition data of the product recommendation configuration data, and then pass The first pop-up window node of the target product recommendation path, according to the product recommendation configuration data, sends the first pop-up window request to the terminal corresponding to each customer data in the customer data set to be recommended, and obtains each terminal according to the first pop-up window Request to send the pop-up window browsing result, when there is a pop-up window browsing result as browsed, use the pop-up window browsing result as the browsed customer data as the browsed customer data set, and then use the voice outbound node of the target product recommendation path , according to the product recommendation configuration data, perform voice outbound calls and intent recognition for each customer corresponding to each customer data in the browsed customer data set, and obtain the respective intent recognition results corresponding to each customer data in the browsed customer data set , when there is an intent recognition result that is willing to accept recommendations, through the second pop-up window node of the target product recommendation path, according to the product recommendation basic information of the product recommendation configuration data, each customer data corresponding to each customer whose intent recognition result is willing to accept recommendations The terminal of the terminal sends a second pop-up window request, obtains the corresponding service operation results of each second pop-up window request through the service operation node of the target product recommendation path, and finally confirms the node through the result confirmation node of the target product recommendation path, according to each service operation result, Each intent recognition result and each pop-up window browsing result determine the target product recommendation result corresponding to each customer data in the customer data set to be recommended, thereby realizing the automatic product recommendation of batch customers according to the target product recommendation path, and timely The effect of product recommendation is tracked, the efficiency of product recommendation is improved, and the existing technology uses the system to send short messages for product recommendation. When there are many customers, it is difficult to track the progress of product recommendation, and it is easy to cause information harassment to customers. Reduce customers Experienced technical issues.

Description of drawings

FIG. 1 is a schematic flow diagram of an artificial intelligence-based product recommendation method according to an embodiment of the present application;

FIG. 2 is a schematic block diagram of the structure of an artificial intelligence-based product recommendation device according to an embodiment of the present application;

FIG. 3 is a schematic block diagram of a computer device according to an embodiment of the present application.

The realization, functional features and advantages of the present application will be further described in conjunction with the embodiments and with reference to the accompanying drawings.

Embodiments of the present invention

Referring to Fig. 1, an artificial intelligence-based product recommendation method is provided in an embodiment of the present application, the method comprising:

S1: Obtain a product recommendation start request, the product recommendation start request carries a target product recommendation path and product recommendation configuration data;

S2: Obtain a customer data set to be recommended according to the customer screening condition data of the product recommendation configuration data through the customer screening node of the target product recommendation path;

S3: Through the first pop-up window node of the target product recommendation path, according to the product recommendation configuration data, respectively send a first pop-up window request to the terminal corresponding to each customer data in the customer data set to be recommended by the product ;

S4: Through the first pop-up node, obtain the pop-up browsing results sent by each terminal according to the first pop-up request;

S5: When the pop-up window browsing result is browsed, use each of the customer data whose pop-up window browsing result is browsed as the browsed customer data set;

S6: Through the voice outbound node of the target product recommendation path, according to the product recommendation configuration data, respectively perform voice outbound calls and intentions for customers corresponding to each of the customer data in the browsed customer data set identifying, obtaining the intention identification results corresponding to each of the customer data in the browsed customer data set;

S7: When the intention recognition result is willing to accept the recommendation, through the second pop-up window node of the target product recommendation path, according to the product recommendation basic information of the product recommendation configuration data, respectively send the intention recognition result to The terminal corresponding to each of the customer data that is willing to accept the recommendation sends a second pop-up window request;

S8: Obtain service operation results corresponding to each of the second pop-up requests through the service operation node of the target product recommendation path;

S9: Through the result confirmation node of the target product recommendation path, determine each of the customer data sets to be recommended according to each of the service operation results, each of the intent recognition results, and each of the pop-up window browsing results The target product recommendation results corresponding to each of the customer data.

In this embodiment, firstly, through the customer screening node of the target product recommendation path, the customer data set to be recommended is obtained according to the customer screening condition data of the product recommendation configuration data, and secondly, through the first pop-up window node of the target product recommendation path, according to the product recommendation configuration Data, respectively send the first pop-up request to the terminal corresponding to each customer data in the customer data set to be recommended by the product, and obtain the pop-up browsing results sent by each terminal according to the first pop-up request. When there is a pop-up browsing result of When it has been browsed, the pop-up window browsing result shows that the browsed customer data is the browsed customer data collection, and then through the voice outbound node of the target product recommendation path, according to the product recommendation configuration data, the browsed customer data collection is respectively The customer corresponding to each customer data in the system performs voice outbound calls and intent recognition, and obtains the respective intent recognition results corresponding to each customer data in the browsed customer data set. When there is an intent recognition result that is willing to accept recommendations, the target product The second pop-up window node of the recommendation path, according to the product recommendation basic information of the product recommendation configuration data, respectively sends a second pop-up window request to the terminal corresponding to each customer data whose intent recognition result is willing to accept the recommendation, through the target product recommendation path The service operation node obtains the corresponding service operation results of each second pop-up window request, and finally confirms the node through the result confirmation node of the target product recommendation path, and determines the product to be recommended according to each service operation result, each intent recognition result, and each pop-up window browsing result The target product recommendation results corresponding to each customer data in the customer data set, so as to realize the automatic product recommendation for batch customers according to the target product recommendation path, and track the effect of product recommendation in time, improve the efficiency of product recommendation, avoid The existing technology adopts the method of sending short messages by the system to recommend products. When there are many customers, it is difficult to track the progress of product recommendation, and it is easy to cause information harassment to customers and reduce the technical problems of customer experience.

Corresponding to S1, you can obtain the product recommendation start request input by the user, or the product recommendation start request sent by the third-party application system, or the product recommendation start that is actively triggered by the application implementing this application according to the product recommendation start time input by the user ask.

The product recommendation start request is an automated request for product recommendation to batch customers.

The target product recommendation path, that is, the product recommendation path that is initially requested for this product recommendation. The product recommendation path includes: a node component identifier and a node serial number, wherein the node component identifier corresponds to the node serial number one by one. The product recommendation path is used to describe the path nodes generated from customer screening, product recommendation for batch customers, progress tracking of product recommendation, and final result of product recommendation.

The node serial number is the serial number of the arrangement position of the node corresponding to the node component identifier.

Node components are pre-packaged components that can be configured by dragging or filling in the node component ID. The node component is a component developed in Java (object-oriented programming language) language.

A node (that is, a node component) is a packaged component of a processing step in product recommendation. Nodes include but are not limited to: customer screening node, first pop-up node, voice outbound node, second pop-up node, service operation node, and result confirmation node.

The product recommendation configuration data is the configuration data that needs to be used when the product recommendation starts requesting product recommendation. Product recommendation configuration data includes: node configuration data, path start time, customer filter data and basic product recommendation information. The node configuration data is the configuration data of the components of each node in the target product recommendation path. The path start time is the time when the first node of the product recommendation path starts to execute. The customer filter condition data is the filter condition for filtering customers from the customer database. The basic product recommendation information is the product information of the product recommended this time. Product information includes, but is not limited to: product identification, product name, and product category. The product identification may be a product ID.

The customer filtering condition data includes: one or more of gender, age group, customer type, customer area, customer label and customers meeting product requirements. For example, the customer filtering condition data is customers in city A who meet the requirements of product A, where "city A" is the customer area, and "customers who meet the requirements of product A" are customers who meet the requirements of the product. I will not give specific examples here. limited.

The customer label may be a customer classification label determined according to user data by using a classification model, or may be a customer classification label input by a user.

Corresponding to S2, pass through the customer screening node of the target product recommendation path, take the path start time of the product recommendation configuration data as the start time, and process customer data from the customer database according to the customer screening condition data of the product recommendation configuration data. Screening, using all the customer data obtained through the screening as the customer data collection for the product to be recommended. The set of customer data to be recommended includes a set of customer data for which product recommendation starts to request product recommendation. That is to say, the target product recommendation path takes the path start time of the product recommendation configuration data as the starting time.

Customer data is the basic information of customers. Customer data includes, but is not limited to: customer identification, contact information, and terminal identification.

The customer database includes a plurality of customer data.

Optionally, the customer screening node is also used to map the nodes of the target product recommendation path into a tree structure for each customer data in the customer data set of the product to be recommended to obtain the product to be recommended The tree structure product recommendation path data corresponding to each of the customer data in the customer data set; obtain the product recommendation path execution panel display request, and the product recommendation path execution panel display request carries the customer identification to be displayed; according to the The customer identification to be displayed is searched from each of the tree-structured product recommendation path data to obtain the tree-structure product recommendation path data to be displayed; the tree-structure product recommendation path data to be displayed is obtained by using the tree structure Display on the Web (Global Wide Area Network) page.

Corresponding to S3, through the first pop-up window node of the target product recommendation path, first pop-up window information is generated according to the product recommendation configuration data, and the first pop-up window information is sent to the customer data set of the product to be recommended according to the first pop-up window information The terminal corresponding to each customer data sends a first pop-up request.

The terminal may be a client of a mobile electronic device, may also be a client of a desktop computer, or may be a web client.

Corresponding to S4, the terminal performs a pop-up window according to the first pop-up window information requested by the first pop-up window, and when the terminal monitors that the pop-up window is displayed and/or the pop-up window is clicked, the terminal determines the pop-up window The browsing result is browsed. When the terminal is opened during the pop-up window, the terminal determines that the pop-up window browsing result is browsed. When within the preset pop-up window duration of the node configuration data of the product recommendation configuration data When no pop-up window is displayed, the pop-up window is clicked, or the terminal is opened, it is determined that the pop-up window browsing result is not browsed, and the pop-up window browsing result is sent to the first pop-up window node.

Wherein, through the first pop-up node, the pop-up window browsing results corresponding to the first pop-up request sent by each of the terminals through the communication connection with the first pop-up node can be understood to be obtained The number of pop-up window browsing results is less than or equal to the customer data in the customer data set to be recommended.

Optionally, when the terminal does not send the pop-up browsing result within the first preset feedback time length, the user data corresponding to the terminal that does not send the pop-up browsing result within the preset feedback time length The corresponding browsing result of the pop-up window is determined to be unbrowsed.

Corresponding to S5, when the browsing result of the pop-up window is browsed, it means that the corresponding client who has browsed the pop-up window is interested in the information of the first pop-up window. Interested customers make outbound voice calls to determine their intentions. Therefore, all the customer data that have been browsed as a result of the pop-up window browsing are taken as a set, and this set is regarded as a browsed customer data set. Thereby, the customer group that needs to make an outbound voice call to determine the intention is determined.

Corresponding to S6, through the voice outbound node of the target product recommendation path, first determine the speech technique according to the product recommendation configuration data, then call the voice outbound call robot to make a voice outbound call according to the determined speech technique, and convert the result of the voice outbound call Recognize the customer's intention after writing the text, and judge whether the customer is willing to accept the recommendation according to the data of the intention recognition, so as to obtain the respective intention recognition results corresponding to each of the customer data in the browsed customer data set.

The voice outbound robot is a voice robot that can automatically make voice outbound calls, and the specific implementation method will not be repeated here.

Corresponding to S7, when there is the intention identification result that is willing to accept the recommendation, it means that the intention identification result is the product corresponding to the customer who is willing to accept the recommendation and is willing to understand or accept the product recommendation basic information of the product recommendation configuration data (may be is a physical product, or it may be a service), through the second pop-up window node of the target product recommendation path, generate second pop-up window information according to the product recommendation configuration data, and identify As a result, the terminal corresponding to each customer data that is willing to accept the recommendation sends a second pop-up window request.

Corresponding to S8, through the service operation node of the target product recommendation path, obtain the pending browsing results sent by each terminal according to the second pop-up window request, obtain service application result data, and obtain service application result data according to the service application result data and Each of the browsing results to be processed determines the service operation result corresponding to the customer data corresponding to each of the second pop-up requests.

Wherein, the terminal performs a pop-up window according to the second pop-up window information requested by the second pop-up window, and when the terminal monitors that the pop-up window is displayed and/or the pop-up window is clicked, the terminal determines the browsing result to be processed As browsed, when the terminal is opened during the pop-up window, the terminal determines that the pop-up window browsing result is browsed, when the node configuration data of the product recommendation configuration data does not appear within the preset pop-up window duration When a pop-up window is displayed, a pop-up window is clicked, or the terminal is opened, it is determined that the browsing result to be processed is unbrowsed.

Corresponding to S9, through the result confirmation node of the target product recommendation path, according to each of the service operation results, each of the intention recognition results that are unwilling to accept recommendations, and each of the pop-up window browsing results that have not been browsed, determine the pending The target product recommendation results corresponding to each of the customer data in the product recommended customer data set.

In one embodiment, the above step of obtaining a product recommendation start request includes:

S11: Obtain a product recommendation path configuration start request;

S12: Start requesting to display a product recommendation path configuration interface according to the product recommendation path configuration;

S13: Obtain node component identifiers, node serial numbers, node configuration data, path start time, customer filter condition data, and the product recommendation basic information according to the product recommendation path configuration interface;

S14: Generate a product recommendation path according to the node component identifier and the node serial number, and obtain the target product recommendation path;

S15: Generate configuration data according to the node configuration data, the path start time, the customer filtering condition data and the product recommendation basic information, and obtain the product recommendation configuration data;

S16: Obtain a product recommendation path configuration submission request;

S17: In response to the product recommendation path configuration submission request, generate the product recommendation start request according to the target product recommendation path and the product recommendation configuration data.

This embodiment automatically realizes the generation of the target product recommendation path, the product recommendation configuration data, and the product recommendation start request according to the configuration, thereby providing support for product recommendation to batch customers.

Corresponding to S11, the product recommendation path configuration start request input by the user may be obtained, or the product recommendation path configuration start request input by the third-party application may be obtained.

The product recommendation path configuration start request is a request for configuring the target product recommendation path and product recommendation configuration data.

Corresponding to S12, when the product recommendation path configuration start request is received, the product recommendation path configuration interface is displayed. It can be understood that the product recommendation path configuration interface is a Web page.

The product recommendation path configuration interface includes: node component display area, product recommendation path configuration area, and node component attribute configuration area. The user only needs to drag the node component in the node component display area to the product recommendation path configuration area, and then click the node component in the product recommendation path configuration area. The property configuration area of the node component displays the properties of the clicked node component. Users can Configure the properties of the clicked node component in the property configuration area of the node component, and use the configured properties of the clicked node component as the node configuration data of the clicked node component.

Corresponding to S13, according to the product recommendation path configuration interface, the node component identifier, node serial number, node configuration data, path start time, customer filter condition data and the basic product recommendation information input by the user are obtained.

Corresponding to S14, according to each of the node serial numbers, map the nodes corresponding to all the node component identifiers into a tree structure, and use the mapped tree structure as the target product recommendation path.

Corresponding to S15, the preset configuration data generation rules are used to generate configuration data according to the node configuration data, the path start time, the customer filter condition data and the product recommendation basic information, and the generated configuration data is used as the Recommended configuration data for the above products.

Corresponding to S16, the product recommendation path configuration submission request input by the user may be acquired, or the product recommendation path configuration submission request input by the third-party application may be acquired.

The product recommendation path configuration submission request is a request to complete the configuration of the product recommendation path and product recommendation configuration data to generate a product recommendation start request.

Corresponding to S17, in response to the product recommendation path configuration submission request, generate the product recommendation start request according to the target product recommendation path and the product recommendation configuration data, and when generating the product recommendation start request, recommend the target product The path and the product recommendation configuration data are used as the data carried in the product recommendation start request.

In one embodiment, the above-mentioned first pop-up window node passing through the target product recommendation path, according to the product recommendation configuration data, respectively sends to the terminal corresponding to each customer data in the customer data set to be recommended The steps of the first pop-up request include:

S31: Through the first pop-up node, generate a node timer according to the node configuration data of the product recommendation configuration data and the path start time, and obtain the first pop-up node timer;

S32: Generate first pop-up information according to the product recommendation basic information through the first pop-up node;

S33: Through the first pop-up node, according to the first pop-up node timer and the first pop-up information, respectively correspond to each of the customer data in the customer data set to be recommended by the product The terminal sends the first pop-up window request.

In this embodiment, according to the first pop-up window node timer and the first pop-up window information, the terminal corresponding to each of the customer data in the customer data set to be recommended is respectively sent the The first pop-up window request is beneficial to control the execution time and execution progress of the first pop-up window node, and provides a basis for accurately tracking the progress of product recommendation.

Corresponding to S31, through the first pop-up window node, add the node preset opening time and path start time of the first pop-up window node in the node configuration data of the product recommendation configuration data to obtain the first pop-up window node An actual start time of a pop-up node, a node timer is generated according to the actual start time of the first pop-up node, and the generated node timer is used as the first pop-up node timer.

The preset start time of the node is the starting time of the node's recommended path for the target product. It can be understood that the node preset opening time of the first node of the target product recommendation path is 0.

The path start time is the time when a target product recommendation path starts to be executed. That is to say, the first node of the product recommendation path is executed at the path start time.

Corresponding to S32, through the first pop-up node, the preset pop-up information generation rule is used to generate pop-up information according to the product recommendation basic information, and the generated pop-up information is used as the first pop-up information.

Corresponding to S33, the first pop-up window node timer sends a first start execution signal to the first pop-up window node when the current time is equal to the actual start time of the first pop-up window node; the first pop-up window node When receiving the first start execution signal, according to the first pop-up window information, respectively send the first pop-up window to the terminal corresponding to each of the customer data in the customer data set to be recommended by the product ask.

In one embodiment, the above-mentioned voice outbound node passing through the target product recommendation path, according to the product recommendation configuration data, separately conducts the customer corresponding to each of the customer data in the browsed customer data set Voice outbound call and intent recognition, the step of obtaining the respective intent recognition results corresponding to each of the customer data in the browsed customer data set includes:

S61: Through the voice outbound node, determine the speech skills according to the product recommendation basic information, and obtain the target voice outbound call skills;

S62: Invoke a preset voice outbound robot through the voice outbound node, and according to the target voice outbound utterance, separately call the customer corresponding to each of the customer data in the browsed customer data set Make a voice outbound call, and obtain the voice outbound call results corresponding to each of the customer data in the browsed customer data set;

S63: Invoke a preset voice conversion model through the voice outbound node, and perform voice conversion on each of the voice outbound results to text, and obtain voice outbound text data corresponding to each of the voice outbound results;

S64: Invoke a preset intent recognition model through the voice outbound node, and perform intent recognition on each of the voice outbound text data according to the target voice outbound utterance, to obtain the browsed customers An intent set corresponding to each of the customer data in the data set;

S65: Through the voice outbound node, according to the intent configuration data of the product recommendation basic information of the product recommendation configuration data, each of the intent sets is judged, and each of the browsed customer data sets is obtained. The intention recognition results corresponding to the respective customer data.

This embodiment realizes the use of the preset voice outbound robot, the preset voice conversion model and the preset intent recognition model for voice outbound calls and intent recognition, thereby determining the attitude of customers to accept product recommendations, in order to further promote customers' willingness to Accepting recommendations provides support; because the preset voice outbound robot, preset voice conversion model and preset intent recognition model are used for voice outbound calls and intent recognition are automatically realized without manual processing, thus improving the process The efficiency of voice outbound calls and intent recognition.

Corresponding to S61, the product recommendation speech database is obtained through the voice outbound node, the product identification of the product recommendation basic information is searched in the product recommendation speech database, and the product recommendation speech database is searched. The utterance data corresponding to the found product identifier is used as the target voice outbound utterance.

Corresponding to S62, call a plurality of preset voice outbound robots through the voice outbound node, and correspond to each of the customer data in the browsed customer data set according to the target voice outbound call technique in batches Customers make voice outbound calls, and use the voice data of each customer as a voice outbound call result.

It can be understood that the preset voice outbound robot is a one-to-one voice outbound robot.

Corresponding to S63, through the voice outbound node, call the preset voice conversion model, respectively perform voice conversion on each of the voice outbound results, and use each piece of text data obtained by voice conversion as a voice outbound text data . That is to say, each said outbound voice call result corresponds to a text data of an outbound voice call. In the text data of the voice outbound call, only the recorded data of the customer who was called out by voice is included.

Corresponding to S64, obtain the query customer's willing speech technique from the target voice outbound call; according to the inquiry customer's willing speech technique, extract the answer text data from each of the voice outbound text data respectively; The answer text data is input into the preset intent recognition model for intent recognition, and all the intents corresponding to each answer text data are regarded as an intent set.

The preset intent recognition model is a model trained based on an NLP model (Natural Language Processing model).

Corresponding to S65, the target intent set is compared with the intent configuration data of the product recommendation basic information of the product recommendation configuration data through the voice outbound node, and when the target intent set contains the intent configuration data, determine the target intent set corresponding The intention identification result corresponding to the customer data is willing to accept the recommendation, otherwise it is determined that the intention identification result corresponding to the customer data corresponding to the target intention set is not willing to accept the recommendation; wherein the target intention set is the Each of the customer data in the browsed customer data set corresponds to any one of the intent sets in the intent sets.

In one embodiment, when the above-mentioned intent identification result is willing to accept the recommendation, through the second pop-up window node of the target product recommendation path, according to the product recommendation basic information of the product recommendation configuration data, respectively send to all The step of sending a second pop-up window request to the terminal corresponding to each of the customer data that is willing to accept the recommendation as a result of the intent recognition includes:

S71: When the intent identification result is willing to accept the recommendation, generate a node timer according to the node configuration data of the product recommendation configuration data and the path start time through the second pop-up node, and obtain the second pop-up node timer;

S72: Generate second pop-up window information according to the product recommendation basic information through the second pop-up window node;

S73: Through the second pop-up window node, according to the second pop-up window node timer and the second pop-up window information, respectively send to each of the customer data corresponding to the intention recognition result that is willing to accept the recommendation The terminal sends the second pop-up request.

This embodiment realizes that the terminal corresponding to the customer data corresponding to the willingness to accept the recommendation is re-initiated a pop-up frame request, which is convenient for the customer to quickly understand and/or apply for the product, and provides support for further promoting the customer's willingness to accept the recommendation. customer experience.

Corresponding to S71, through the second pop-up window node, add the node preset opening time and path start time of the node configuration data of the product recommendation configuration data of the second pop-up window node to obtain the second pop-up window node The actual start time of the pop-up node generates a node timer according to the actual start time of the second pop-up node, and uses the generated node timer as the second pop-up node timer.

Corresponding to S72, through the second pop-up window node, the preset pop-up window information generation rule is used to generate pop-up window information according to the product recommendation basic information, and the generated pop-up window information is used as the second pop-up window information.

Corresponding to S73, the second pop-up window node timer sends a second start execution signal to the second pop-up window node when the current time is equal to the actual start time of the second pop-up window node; the second pop-up window node When receiving the second execution start signal, according to the second pop-up window information, respectively send the second pop-up window request to the terminal corresponding to each of the customer data whose intention identification result is willing to accept the recommendation .

In one embodiment, the step of obtaining the service operation results corresponding to each of the second pop-up requests through the service operation node of the target product recommendation path includes:

S81: Obtain, through the service operation node, pending browsing results sent by each terminal according to the second pop-up window request;

S82: Obtain service application result data through the service operation node;

S83: Use the service operation node to search from the service application result data according to the target customer data to obtain the service application search result, wherein the target customer data is any one of the browsed customer data sets said customer data;

S84: When the search result of the service application is successful, through the service operation node, determine that the service operation result corresponding to the target customer data is an applied service;

S85: When the search result of the service application is a failure, and the pending browsing result corresponding to the target customer data is browsed, use the service operation node to determine the The service operation result is browsed but not applied;

S86: When the search result of the service application is a failure, and the pending browsing result corresponding to the target customer data is unbrowsed, use the service operation node to determine the The service operation result is not browsed and not applied.

In this embodiment, the service operation result is determined according to the service application result data and the pending browsing result corresponding to the second pop-up window request, which provides a basis for automatically determining the target product recommendation result and improves the efficiency of product recommendation.

Corresponding to S81, wherein, through the service operation node, obtain the browsing results to be processed corresponding to the second pop-up request sent by each terminal through the communication connection with the service operation node, it can be understood that obtaining The number of browsing results to be processed is less than or equal to the number of customer data corresponding to the intention identification result being willing to accept recommendations.

Optionally, when there is a terminal that does not send the browsing result to be processed within the second preset feedback time length, the terminal that does not send the browsing result to be processed within the second preset feedback time length The browsing result to be processed corresponding to the corresponding user data is determined to be unbrowsed.

Corresponding to S82, through the service operation node, the service application result data can be obtained from the database, or the service application result data can be obtained from a third-party application system.

The service application result data includes: customer identification, service application information.

Corresponding to S83, the customer identification of the target customer data is searched from the service application result data through the service operation node, and when the customer identification is found in the service application result data, the service corresponding to the target customer data is The application search result is determined as successful, otherwise, the service application search result corresponding to the target customer data is determined as failure.

Corresponding to S84, when the search result of the service application is successful, it means that the customer corresponding to the target customer data has completed the application for the product corresponding to the product recommendation basic information, so the target customer data can be determined through the service operation node The corresponding service operation result is applied for service.

Corresponding to S85, when the search result of the service application is failure, and the pending browsing result corresponding to the target customer data is browsed, it means that the target customer data has browsed the information of the second pop-up window However, there is no application for products corresponding to the recommended basic information of the product, so the service operation result corresponding to the target customer data can be determined to be browsed but not applied for through the service operation node.

Corresponding to S86, when the search result of the service application is failure, and the pending browsing result corresponding to the target customer data is not browsed, it means that the target customer data has not browsed the information of the second pop-up window And there is no application for the product corresponding to the product recommendation basic information, so the service operation result corresponding to the target customer data can be determined as unbrowsed and unapplied through the service operation node.

In one embodiment, the above-mentioned result confirmation node through the target product recommendation path determines the customer to be recommended according to each of the service operation results, each of the intent recognition results and each of the pop-up window browsing results The step of recommending the target product corresponding to each of the customer data in the data set includes:

S91: When the service operation result is browsed but not applied for, use the result confirmation node to recommend the target product corresponding to each of the customer data corresponding to the service operation result as browsed but not applied for The product recommendation result of the result is determined as failure, and the failure reason of the target product recommendation result corresponding to each of the customer data corresponding to the service operation result being browsed but not applied is determined as browsed but not accepted;

S92: When the service operation result is unbrowsed and unapplied, respectively recommend the target product corresponding to each of the customer data corresponding to the service operation result being unbrowsed and unapplied through the result confirmation node The result of the product recommendation result is determined as a failure, and the failure reason of the target product recommendation result corresponding to each of the customer data corresponding to the service operation result is not browsed and not applied is determined as not browsed and not applied. accept;

S93: When the service operation result is an applied service, through the result confirmation node, respectively set the service operation result as the target product recommendation result corresponding to each of the customer data corresponding to the applied service The product recommendation result is determined to be successful;

S94: through the result confirmation node, respectively determine the product recommendation result of the target product recommendation result corresponding to each of the customer data that is not willing to accept the recommendation as a failure, and confirm the intention The identification result is that the failure reason of the target product recommendation result corresponding to each of the customer data that is unwilling to accept the recommendation is determined to be an outbound call and unwilling to learn more;

S95: Through the result confirmation node, determine the product recommendation result of the target product recommendation result corresponding to each of the unbrowsed customer data as failed, and confirm the pop-up window The failure reason of the target product recommendation result corresponding to each customer data that has not been browsed is determined to be not interesting.

In this embodiment, according to each of the service operation results, each of the intent recognition results, and each of the pop-up window browsing results, determine the target product recommendation results corresponding to each of the customer data in the customer data set to be recommended. , so that the final tracking result of the product recommendation start request is obtained automatically, without manual operation, and the efficiency of product recommendation is improved.

Corresponding to S91, when the service operation result is browsed but not applied, it means that although the customer is interested in the product corresponding to the product recommendation basic information and has browsed the second pop-up window information, but is unwilling to apply for the product, you can pass The result confirmation node respectively determines that the product recommendation result of the service operation result that is the target product recommendation result corresponding to each of the customer data that has not been applied for has been browsed as a failure, and respectively confirms that the service operation result The failure reason of the target product recommendation result corresponding to each of the customer data corresponding to browsed but not applied is determined as browsed but not accepted.

That is to say, the target product recommendation results include: product recommendation results and failure reasons, and the product recommendation results and failure reasons correspond one-to-one.

Corresponding to S92, when the service operation result is not browsed and not applied, it means that although the customer is interested in the product corresponding to the product recommendation basic information, he has not browsed the second pop-up window information and is unwilling to apply for the product, which means that although the customer Interested in the product corresponding to the product recommendation basic information but unwilling to apply for the service, at this time, through the result confirmation node, the service operation result is the target corresponding to each of the customer data that has not been browsed or applied for The product recommendation result of the product recommendation result is determined as failure, and the failure reason of the target product recommendation result corresponding to each of the customer data corresponding to the service operation result is not browsed and not applied is determined as failure. Browse not accepted.

Corresponding to S93, when the service operation result is applied for service, it means that the customer has applied for the product corresponding to the product recommendation basic information. At this time, the service operation result can be confirmed as the applied service through the result confirmation node. The product recommendation result corresponding to the target product recommendation result corresponding to each of the customer data is determined to be successful.

Corresponding to S94, the result of the intention recognition is that the customer corresponding to the customer data who is not willing to accept the recommendation has browsed the information in the pop-up window for the first time, but is unwilling to further understand and accept the product corresponding to the basic information of the product recommendation. A result confirmation node, respectively determining the product recommendation result of the target product recommendation result corresponding to each of the customer data for which the intention identification result is unwilling to accept the recommendation as failure, and determining the intention identification result as unwilling The reason for the failure of the target product recommendation result corresponding to each customer data that accepts the recommendation is determined to be an outbound call and unwilling to learn more.

Corresponding to S95, the browsing result of the pop-up window is the information that the customer corresponding to the unbrowsed customer data has not browsed the first pop-up window, so through the result confirmation node, the browsing result of the pop-up window is respectively unbrowsed for each The product recommendation result of the target product recommendation result corresponding to the customer data is determined as a failure, and the pop-up browsing result is the unbrowsed item of the target product recommendation result corresponding to each of the customer data. The failure reason was determined to be not of interest.

Referring to Fig. 3, the present application also proposes a product recommendation device based on artificial intelligence, which includes:

The request obtaining module 100 is used to obtain a product recommendation start request, and the product recommendation start request carries a target product recommendation path and product recommendation configuration data;

The customer data set to be recommended product determination module 200 is used to obtain the customer data set to be recommended by the customer screening node of the target product recommendation path according to the customer filter condition data of the product recommendation configuration data;

The first pop-up window request sending module 300 is used to pass the first pop-up window node of the target product recommendation path, according to the product recommendation configuration data, respectively send each customer data in the customer data set to be recommended by the product The corresponding terminal sends a first pop-up request;

The pop-up window browsing result acquisition module 400 is used to obtain the pop-up window browsing results sent by each terminal according to the first pop-up window request through the first pop-up window node;

The browsed customer data set determination module 500 is used to use each of the customer data whose pop-up window browsing result is browsed as the browsed customer data set when there is the pop-up window browsing result as browsed;

The intent recognition result determination module 600 is configured to, according to the product recommendation configuration data, respectively perform an outbound call node for each of the customer data in the browsed customer data set through the voice outbound node of the target product recommendation path. The customer performs voice outbound calls and intent recognition, and obtains respective intent recognition results corresponding to each of the customer data in the browsed customer data set;

The second pop-up window request sending module 700 is used to pass the second pop-up window node of the target product recommendation path, according to the product recommendation basic information of the product recommendation configuration data, when the intent recognition result is willing to accept the recommendation. , respectively sending a second pop-up window request to the terminals corresponding to each of the customer data for which the intention identification result is willing to accept the recommendation;

The service operation result determination module 800 is configured to obtain the service operation results corresponding to each of the second pop-up window requests through the service operation node of the target product recommendation path;

The target product recommendation result determination module 900 is configured to determine the target product according to each of the service operation results, each of the intent recognition results, and each of the pop-up window browsing results through the result confirmation node of the target product recommendation path The target product recommendation results corresponding to each of the customer data in the recommended customer data set.

Referring to FIG. 3 , an embodiment of the present application also provides a computer device, which may be a server, and its internal structure may be as shown in FIG. 3 . The computer device includes a processor, memory, network interface and database connected by a system bus. Among them, the processor designed by the computer is used to provide calculation and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, computer programs and databases. The memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The database of the computer equipment is used to store data such as product recommendation methods based on artificial intelligence. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, an artificial intelligence-based product recommendation method is realized. The artificial intelligence-based product recommendation method includes: obtaining a product recommendation start request, the product recommendation start request carrying a target product recommendation path and product recommendation configuration data; passing through the customer screening node of the target product recommendation path, according to the According to the customer filter condition data of the product recommendation configuration data, the customer data set to be recommended is obtained; through the first pop-up window node of the target product recommendation path, according to the product recommendation configuration data, respectively to the customers who are to be recommended by the product The terminal corresponding to each customer data in the data set sends a first pop-up window request; through the first pop-up window node, obtain the pop-up window browsing results sent by each terminal according to the first pop-up window request; when there are all When the browsing result of the pop-up window is browsed, each of the customer data that has been browsed as the browsing result of the pop-up window is used as the browsed customer data collection; through the voice outbound node of the target product recommendation path, according to the Product recommendation configuration data, performing voice outbound calls and intent recognition for each customer corresponding to each of the customer data sets in the browsed customer data set, and obtaining each of the customer data in the browsed customer data set Respectively corresponding intent recognition results; when the intention recognition result is willing to accept the recommendation, through the second pop-up window node of the target product recommendation path, according to the product recommendation basic information of the product recommendation configuration data, send to all The result of the intent recognition is that the terminal corresponding to each of the customer data willing to accept the recommendation sends a second pop-up window request; through the service operation node of the target product recommendation path, obtain the respective corresponding The result of the service operation; through the result confirmation node of the target product recommendation path, according to each of the service operation results, each of the intent recognition results and each of the pop-up window browsing results, determine the customer data set to be recommended by the product The target product recommendation results corresponding to each of the customer data in .

An embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, an artificial intelligence-based product recommendation method is implemented, including the steps of: obtaining a product recommendation start request, and The product recommendation start request carries the target product recommendation path and product recommendation configuration data; through the customer screening node of the target product recommendation path, the customer data set to be recommended is obtained according to the customer screening condition data of the product recommendation configuration data; Through the first pop-up window node of the target product recommendation path, according to the product recommendation configuration data, send the first pop-up window request to the terminal corresponding to each customer data in the customer data set to be recommended by the product; The first pop-up node obtains the pop-up browsing result sent by each terminal according to the first pop-up request; when the pop-up browsing result is browsed, set the pop-up browsing result as already browsed Each of the browsed customer data is regarded as a browsed customer data set; through the voice outbound node of the target product recommendation path, according to the product recommendation configuration data, each of the browsed customer data sets is respectively The customer corresponding to the customer data performs voice outbound calls and intention identification, and obtains the intention identification results corresponding to each of the customer data in the browsed customer data set; when there is the intention identification result that is willing to accept the recommendation , through the second pop-up window node of the target product recommendation path, according to the product recommendation basic information of the product recommendation configuration data, respectively send the said intention identification result to the corresponding to each of the said customer data who are willing to accept the recommendation The terminal sends a second pop-up request; through the service operation node of the target product recommendation path, obtain the service operation results corresponding to each of the second pop-up requests; through the result confirmation node of the target product recommendation path, according to each The service operation results, each of the intent recognition results and each of the pop-up window browsing results determine the target product recommendation results corresponding to each of the customer data in the customer data set to be recommended.

In the aforementioned AI-based product recommendation method, first, through the customer screening node of the target product recommendation path, the customer data set to be recommended is obtained according to the customer screening condition data of the product recommendation configuration data, and secondly, through the first step of the target product recommendation path. The pop-up node, according to the product recommendation configuration data, sends the first pop-up request to the terminal corresponding to each customer data in the customer data set to be recommended, and obtains the pop-up browsing results sent by each terminal according to the first pop-up request , when there is a pop-up window browsing result that has been browsed, use the pop-up window browsing result as the browsed customer data as the browsed customer data set, and then use the voice outbound node of the target product recommendation path to configure the data according to the product recommendation, Carry out voice outbound calls and intent recognition for each customer corresponding to each customer data in the browsed customer data set, and obtain the respective intent recognition results corresponding to each customer data in the browsed customer data set. When there is an intent recognition result of When willing to accept the recommendation, through the second pop-up window node of the target product recommendation path, according to the product recommendation basic information of the product recommendation configuration data, send the second pop-up window to the terminal corresponding to each customer data whose intent recognition result is willing to accept the recommendation Request, through the service operation node of the target product recommendation path, obtain the corresponding service operation results of each second pop-up window request, and finally through the result confirmation node of the target product recommendation path, according to each service operation result, each intent recognition result and each pop-up window Browse the results in the window to determine the target product recommendation results corresponding to each customer data in the customer data set to be recommended, so as to realize the automatic product recommendation for batch customers according to the target product recommendation path, and track the effect of product recommendation in time. The efficiency of product recommendation is improved, and the prior art avoids the technical problem that the prior art adopts the system to send short messages for product recommendation, it is difficult to track the progress of product recommendation when there are many customers, and it is easy to cause information harassment to customers and reduce customer experience.

The computer-readable storage medium may be non-volatile or volatile.

Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be implemented through computer programs to instruct related hardware, and the computer programs can be stored in a non-volatile computer-readable memory In the medium, when the computer program is executed, it may include the processes of the embodiments of the above-mentioned methods. Wherein, any references to memory, storage, database or other media provided in the present application and used in the embodiments may include non-volatile and/or volatile memory. Nonvolatile memory can include read only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Double Data Rate SDRAM (SSRSDRAM), Enhanced SDRAM (ESDRAM), Synchronous Link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

Claims

A method for product recommendation based on artificial intelligence, wherein the method includes:

Obtaining a product recommendation start request, the product recommendation start request carrying a target product recommendation path and product recommendation configuration data;

Through the customer screening node of the target product recommendation path, the customer data set to be recommended is obtained according to the customer screening condition data of the product recommendation configuration data;

Through the first pop-up window node of the target product recommendation path, according to the product recommendation configuration data, respectively send a first pop-up window request to the terminal corresponding to each customer data in the customer data set to be recommended by the product;

Obtain the pop-up browsing results sent by each terminal according to the first pop-up request through the first pop-up node;

When the browsing result of the pop-up window is browsed, each of the customer data whose browsing result of the pop-up window is browsed is used as the browsed customer data set;

Through the voice outbound node of the target product recommendation path, according to the product recommendation configuration data, respectively perform voice outbound calls and intention recognition for customers corresponding to each of the customer data in the browsed customer data set, Obtain the intention recognition results corresponding to each of the customer data in the browsed customer data set;

When the intention recognition result is willing to accept the recommendation, through the second pop-up window node of the target product recommendation path, according to the product recommendation basic information of the product recommendation configuration data, the intention recognition result is willing to accept the recommendation respectively. The terminal corresponding to each recommended customer data sends a second pop-up window request;

Obtain the service operation results corresponding to each of the second pop-up requests through the service operation node of the target product recommendation path;

Through the result confirmation node of the target product recommendation path, each of the customer data sets to be recommended is determined according to each of the service operation results, each of the intent recognition results, and each of the pop-up window browsing results. The target product recommendation results corresponding to the customer data.
The artificial intelligence-based product recommendation method according to claim 1, wherein the step of obtaining a product recommendation start request includes:

Obtain product recommendation path configuration start request;

According to the product recommendation path configuration, start requesting to display the product recommendation path configuration interface;

Acquiring node component identification, node serial number, node configuration data, path start time, customer filter condition data and basic product recommendation information according to the product recommendation path configuration interface;

generating a product recommendation path according to the node component identifier and the node serial number, to obtain the target product recommendation path;

generating configuration data according to the node configuration data, the path start time, the customer filtering condition data and the product recommendation basic information, to obtain the product recommendation configuration data;

Obtain a product recommendation path configuration submission request;

In response to the product recommendation path configuration submission request, the product recommendation start request is generated according to the target product recommendation path and the product recommendation configuration data.
The artificial intelligence-based product recommendation method according to claim 1, wherein the first pop-up window node passing through the target product recommendation path, according to the product recommendation configuration data, respectively recommends the product to the customer to be recommended The step of sending the first pop-up window request by the terminal corresponding to each customer data in the data set includes:

Through the first pop-up window node, a node timer is generated according to the node configuration data and path start time of the product recommendation configuration data, and the first pop-up window node timer is obtained;

Generate first pop-up window information according to the product recommendation basic information through the first pop-up window node;

Through the first pop-up window node, according to the first pop-up window node timer and the first pop-up window information, respectively, to the customer data corresponding to each of the customer data sets to be recommended by the product The terminal sends the first pop-up request.
The artificial intelligence-based product recommendation method according to claim 1, wherein the voice outbound node passing through the target product recommendation path, according to the product recommendation configuration data, respectively collects the browsed customer data The customer corresponding to each of the customer data in the customer performs voice outbound calls and intent recognition, and obtains the steps of the intent recognition results corresponding to each of the customer data in the browsed customer data set, including:

Through the voice outgoing call node, the speech technique is determined according to the product recommendation basic information, and the target voice outgoing call speech technique is obtained;

Call the preset voice outbound robot through the voice outbound node, and make voice calls to the customers corresponding to each of the customer data in the browsed customer data set according to the target voice outbound call technique Outbound calls, obtaining voice outbound results corresponding to each of the customer data in the browsed customer data set;

Calling a preset voice conversion model through the voice outbound node, and performing voice conversion text on each of the voice outbound results respectively, to obtain voice outbound text data corresponding to each of the voice outbound results;

Through the voice outbound node, call the preset intention recognition model, according to the target voice outbound utterance, respectively perform intention recognition on each of the voice outbound text data, and obtain the browsed customer data set Intent sets corresponding to each of the customer data in ;

Through the voice outbound node, according to the intent configuration data of the product recommendation basic information of the product recommendation configuration data, each of the intent sets is judged, and each of the customers in the browsed customer data set is obtained. The intent recognition results corresponding to each of the data.
The artificial intelligence-based product recommendation method according to claim 1, wherein, when the intent recognition result is willing to accept the recommendation, pass through the second pop-up window node of the target product recommendation path, according to the product The step of recommending the basic product recommendation information of the configuration data, respectively sending a second pop-up window request to the terminal corresponding to each of the customer data whose intention identification result is willing to accept the recommendation, includes:

When the intent identification result is willing to accept the recommendation, through the second pop-up node, a node timer is generated according to the node configuration data and path start time of the product recommendation configuration data, and the second pop-up node timer is obtained ;

Generate second pop-up window information according to the product recommendation basic information through the second pop-up window node;

Through the second pop-up window node, according to the second pop-up window node timer and the second pop-up window information, respectively, to the said customer data corresponding to each of the customer data whose intention identification result is willing to accept the recommendation The terminal sends the second pop-up request.
The method for recommending products based on artificial intelligence according to claim 1, wherein the step of obtaining service operation results corresponding to each of the second pop-up window requests through the service operation node of the target product recommendation path includes: :

Obtain the browsing results to be processed sent by each terminal according to the second pop-up request through the service operation node;

Obtain service application result data through the service operation node;

Through the service operation node, search according to the target customer data from the service application result data to obtain the service application search result, wherein the target customer data is any one of the browsed customer data sets customer data;

When the search result of the service application is successful, through the service operation node, determine that the service operation result corresponding to the target customer data is an applied service;

When the search result of the service application is failed, and the pending browsing result corresponding to the target customer data is browsed, determine the service operation corresponding to the target customer data through the service operation node The result is browsed but not applied;

When the search result of the service application is a failure, and the pending browsing result corresponding to the target customer data is not browsed, determine the service operation corresponding to the target customer data through the service operation node The result is not browsed and not applied.
The product recommendation method based on artificial intelligence according to claim 1, wherein the result confirmation node passing through the target product recommendation path, according to each of the service operation results, each of the intention recognition results and each of the pop-up Browse the results in the window, and determine the steps of the target product recommendation results corresponding to each of the customer data in the customer data set to be recommended, including:

When the service operation result is browsed but not applied, through the result confirmation node, the service operation result is the target product recommendation result corresponding to each of the customer data corresponding to browsed but not applied The product recommendation result is determined as failure, and the reason for the failure of the target product recommendation result corresponding to each of the customer data corresponding to the service operation result being browsed but not applied is determined as browsed but not accepted;

When the service operation result is unbrowsed and unapplied, through the result confirmation node, the service operation result is the target product recommendation result corresponding to each of the customer data corresponding to unbrowsed and unapplied. The product recommendation result is determined as failure, and the failure reason of the target product recommendation result corresponding to each of the customer data corresponding to the service operation result being not browsed and not applied is determined as not browsed and not accepted;

When the service operation result is an applied service, through the result confirmation node, the service operation result is the product recommendation of the target product recommendation result corresponding to each of the customer data corresponding to the applied service The result is determined to be successful;

Through the result confirmation node, the product recommendation results of the target product recommendation results corresponding to each of the customer data that are not willing to accept the recommendation as the result of the intention identification are respectively determined as failed, and the result of the identification of the intention is determined as a failure. The reason for the failure of the target product recommendation result corresponding to each of the customer data that is unwilling to accept the recommendation is determined to be an outbound call and unwilling to learn more;

Through the result confirmation node, the product recommendation results of the target product recommendation results corresponding to the target product recommendation results corresponding to each of the unbrowsed customer data are respectively determined as failed, and the pop-up browsing results are determined as failures. The failure reason of the target product recommendation result corresponding to each of the unbrowsed customer data is determined to be not interesting.
A product recommendation device based on artificial intelligence, wherein the device includes:

The request obtaining module is used to obtain a product recommendation start request, and the product recommendation start request carries a target product recommendation path and product recommendation configuration data;

The customer data set determination module to be recommended is used to obtain the customer data set to be recommended according to the customer filter condition data of the product recommendation configuration data through the customer screening node of the target product recommendation path;

The first pop-up window request sending module is used to correspond to each customer data in the customer data set to be recommended according to the product recommendation configuration data through the first pop-up window node of the target product recommendation path The terminal sends the first pop-up window request;

A pop-up browsing result acquisition module, configured to obtain, through the first pop-up node, the pop-up browsing results sent by each terminal according to the first pop-up request;

The browsed customer data set determination module is used to use each of the customer data whose pop-up browsing result is browsed as the browsed customer data set when there is a browsed result of the pop-up window;

The intention recognition result determination module is used to pass through the voice outbound node of the target product recommendation path, and according to the product recommendation configuration data, separately identify the customer corresponding to each of the customer data in the browsed customer data set performing voice outbound calls and intent recognition, and obtaining intent recognition results corresponding to each of the customer data in the browsed customer data set;

The second pop-up window request sending module is used to pass the second pop-up window node of the target product recommendation path, according to the product recommendation basic information of the product recommendation configuration data, when the intent recognition result is willing to accept the recommendation, sending a second pop-up window request to the terminals corresponding to each of the customer data for which the intention identification result is willing to accept the recommendation;

The service operation result determination module is used to obtain the service operation results corresponding to each of the second pop-up window requests through the service operation node of the target product recommendation path;

A target product recommendation result determination module, configured to determine the product to be recommended according to each of the service operation results, each of the intent recognition results, and each of the pop-up window browsing results through the result confirmation node of the target product recommendation path The target product recommendation results corresponding to each of the customer data in the customer data set.
A computer device, comprising a memory and a processor, the memory stores a computer program, wherein the processor implements the following method steps when executing the computer program:

Obtaining a product recommendation start request, the product recommendation start request carrying a target product recommendation path and product recommendation configuration data;

Through the customer screening node of the target product recommendation path, the customer data set to be recommended is obtained according to the customer screening condition data of the product recommendation configuration data;

Through the first pop-up window node of the target product recommendation path, according to the product recommendation configuration data, respectively send a first pop-up window request to the terminal corresponding to each customer data in the customer data set to be recommended by the product;

Obtain the pop-up browsing results sent by each terminal according to the first pop-up request through the first pop-up node;

When the browsing result of the pop-up window is browsed, each of the customer data whose browsing result of the pop-up window is browsed is used as the browsed customer data set;

Through the voice outbound node of the target product recommendation path, according to the product recommendation configuration data, respectively perform voice outbound calls and intention recognition for customers corresponding to each of the customer data in the browsed customer data set, Obtain the intention recognition results corresponding to each of the customer data in the browsed customer data set;

When the intention recognition result is willing to accept the recommendation, through the second pop-up window node of the target product recommendation path, according to the product recommendation basic information of the product recommendation configuration data, the intention recognition result is willing to accept the recommendation respectively. The terminal corresponding to each recommended customer data sends a second pop-up window request;

Obtain the service operation results corresponding to each of the second pop-up requests through the service operation node of the target product recommendation path;

Through the result confirmation node of the target product recommendation path, each of the customer data sets to be recommended is determined according to each of the service operation results, each of the intent recognition results, and each of the pop-up window browsing results. The target product recommendation results corresponding to the customer data.
The computer device according to claim 9, wherein said step of obtaining a product recommendation start request comprises:

Obtain product recommendation path configuration start request;

According to the product recommendation path configuration, start requesting to display the product recommendation path configuration interface;

Acquiring node component identification, node serial number, node configuration data, path start time, customer filter condition data and basic product recommendation information according to the product recommendation path configuration interface;

generating a product recommendation path according to the node component identifier and the node serial number, to obtain the target product recommendation path;

generating configuration data according to the node configuration data, the path start time, the customer filtering condition data and the product recommendation basic information, to obtain the product recommendation configuration data;

Obtain a product recommendation path configuration submission request;

In response to the product recommendation path configuration submission request, the product recommendation start request is generated according to the target product recommendation path and the product recommendation configuration data.
The computer device according to claim 9, wherein the first pop-up window node passing through the target product recommendation path, according to the product recommendation configuration data, respectively sends to each of the customer data sets to be recommended products The steps of sending the first pop-up window request by the terminal corresponding to each customer data include:

Through the first pop-up window node, a node timer is generated according to the node configuration data and path start time of the product recommendation configuration data, and the first pop-up window node timer is obtained;

Generate first pop-up window information according to the product recommendation basic information through the first pop-up window node;

Through the first pop-up window node, according to the first pop-up window node timer and the first pop-up window information, respectively, to the customer data corresponding to each of the customer data sets to be recommended by the product The terminal sends the first pop-up request.
The computer device according to claim 9, wherein the voice outbound node that passes through the target product recommendation path, according to the product recommendation configuration data, separately for each of the browsed customer data sets The customer corresponding to the above customer data performs voice outbound calls and intent recognition, and obtains the step of intent recognition results corresponding to each of the customer data in the browsed customer data set, including:

Through the voice outgoing call node, the speech technique is determined according to the product recommendation basic information, and the target voice outgoing call speech technique is obtained;

Call the preset voice outbound robot through the voice outbound node, and make voice calls to the customers corresponding to each of the customer data in the browsed customer data set according to the target voice outbound call technique Outbound calls, obtaining voice outbound results corresponding to each of the customer data in the browsed customer data set;

Calling a preset voice conversion model through the voice outbound node, and performing voice conversion text on each of the voice outbound results respectively, to obtain voice outbound text data corresponding to each of the voice outbound results;

Through the voice outbound node, call the preset intention recognition model, according to the target voice outbound utterance, respectively perform intention recognition on each of the voice outbound text data, and obtain the browsed customer data set Intent sets corresponding to each of the customer data in ;

Through the voice outbound node, according to the intent configuration data of the product recommendation basic information of the product recommendation configuration data, each of the intent sets is judged, and each of the customers in the browsed customer data set is obtained. The intent recognition results corresponding to each of the data.
The computer device according to claim 9, wherein when the intention recognition result is willing to accept the recommendation, the product according to the product recommendation configuration data is passed through the second pop-up window node of the target product recommendation path The step of recommending basic information and sending a second pop-up window request to the terminal corresponding to each of the customer data for which the intention identification result is willing to accept the recommendation includes:

When the intent identification result is willing to accept the recommendation, through the second pop-up node, a node timer is generated according to the node configuration data and path start time of the product recommendation configuration data, and the second pop-up node timer is obtained ;

Generate second pop-up window information according to the product recommendation basic information through the second pop-up window node;

Through the second pop-up window node, according to the second pop-up window node timer and the second pop-up window information, respectively, to the said customer data corresponding to each of the customer data whose intention identification result is willing to accept the recommendation The terminal sends the second pop-up request.
The computer device according to claim 9, wherein the step of obtaining the service operation results corresponding to each of the second pop-up window requests through the service operation node of the target product recommendation path includes:

Obtain the browsing results to be processed sent by each terminal according to the second pop-up request through the service operation node;

Obtain service application result data through the service operation node;

Through the service operation node, search according to the target customer data from the service application result data to obtain the service application search result, wherein the target customer data is any one of the browsed customer data sets customer data;

When the search result of the service application is successful, through the service operation node, determine that the service operation result corresponding to the target customer data is an applied service;

When the search result of the service application is failed, and the pending browsing result corresponding to the target customer data is browsed, determine the service operation corresponding to the target customer data through the service operation node The result is browsed but not applied;

When the search result of the service application is a failure, and the pending browsing result corresponding to the target customer data is not browsed, determine the service operation corresponding to the target customer data through the service operation node The result is not browsed and not applied.
A computer-readable storage medium, on which a computer program is stored, wherein, when the computer program is executed by a processor, the following method steps are implemented:

Obtaining a product recommendation start request, the product recommendation start request carrying a target product recommendation path and product recommendation configuration data;

Through the customer screening node of the target product recommendation path, the customer data set to be recommended is obtained according to the customer screening condition data of the product recommendation configuration data;

Through the first pop-up window node of the target product recommendation path, according to the product recommendation configuration data, respectively send a first pop-up window request to the terminal corresponding to each customer data in the customer data set to be recommended by the product;

Obtain the pop-up browsing results sent by each terminal according to the first pop-up request through the first pop-up node;

When the browsing result of the pop-up window is browsed, each of the customer data whose browsing result of the pop-up window is browsed is used as the browsed customer data set;

Through the voice outbound node of the target product recommendation path, according to the product recommendation configuration data, respectively perform voice outbound calls and intention recognition for customers corresponding to each of the customer data in the browsed customer data set, Obtain the intention recognition results corresponding to each of the customer data in the browsed customer data set;

When the intention recognition result is willing to accept the recommendation, through the second pop-up window node of the target product recommendation path, according to the product recommendation basic information of the product recommendation configuration data, the intention recognition result is willing to accept the recommendation respectively. The terminal corresponding to each recommended customer data sends a second pop-up window request;

Obtain the service operation results corresponding to each of the second pop-up requests through the service operation node of the target product recommendation path;

Through the result confirmation node of the target product recommendation path, each of the customer data sets to be recommended is determined according to each of the service operation results, each of the intent recognition results, and each of the pop-up window browsing results. The target product recommendation results corresponding to the customer data.
The computer-readable storage medium according to claim 15, wherein the step of obtaining a product recommendation start request comprises:

Obtain product recommendation path configuration start request;

According to the product recommendation path configuration, start requesting to display the product recommendation path configuration interface;

Acquiring node component identification, node serial number, node configuration data, path start time, customer filter condition data and basic product recommendation information according to the product recommendation path configuration interface;

generating a product recommendation path according to the node component identifier and the node serial number, to obtain the target product recommendation path;

generating configuration data according to the node configuration data, the path start time, the customer filtering condition data and the product recommendation basic information, to obtain the product recommendation configuration data;

Obtain a product recommendation path configuration submission request;

In response to the product recommendation path configuration submission request, the product recommendation start request is generated according to the target product recommendation path and the product recommendation configuration data.
The computer-readable storage medium according to claim 15, wherein the first pop-up window node passing through the target product recommendation path, according to the product recommendation configuration data, respectively sets customer data sets for the product to be recommended The steps of sending the first pop-up window request by the terminal corresponding to each customer data in , including:

Through the first pop-up window node, a node timer is generated according to the node configuration data and path start time of the product recommendation configuration data, and the first pop-up window node timer is obtained;

Generate first pop-up window information according to the product recommendation basic information through the first pop-up window node;

Through the first pop-up window node, according to the first pop-up window node timer and the first pop-up window information, respectively, to the customer data corresponding to each of the customer data sets to be recommended by the product The terminal sends the first pop-up request.
The computer-readable storage medium according to claim 15, wherein, the voice outbound node passing through the target product recommendation path, according to the product recommendation configuration data, separately selects each of the browsed customer data sets The customer corresponding to each of the customer data performs voice outbound calls and intent recognition, and the step of obtaining the respective intent recognition results corresponding to each of the customer data in the browsed customer data set includes:

Through the voice outgoing call node, the speech technique is determined according to the product recommendation basic information, and the target voice outgoing call speech technique is obtained;

Call the preset voice outbound robot through the voice outbound node, and make voice calls to the customers corresponding to each of the customer data in the browsed customer data set according to the target voice outbound call technique Outbound calls, obtaining voice outbound results corresponding to each of the customer data in the browsed customer data set;

Through the voice outbound node, call the preset voice conversion model, respectively perform voice conversion text on each of the voice outbound results, and obtain the voice outbound text data corresponding to each of the voice outbound results;

Through the voice outbound node, call the preset intention recognition model, according to the target voice outbound utterance, respectively perform intention recognition on each of the voice outbound text data, and obtain the browsed customer data set Intent sets corresponding to each of the customer data in ;

Through the voice outbound node, according to the intent configuration data of the product recommendation basic information of the product recommendation configuration data, each of the intent sets is judged, and each of the customers in the browsed customer data set is obtained. The intent recognition results corresponding to each of the data.
The computer-readable storage medium according to claim 15, wherein, when the intention recognition result is willing to accept the recommendation, pass through the second pop-up window node of the target product recommendation path, according to the product recommendation configuration The basic product recommendation information of the data, the step of sending a second pop-up window request to the terminal corresponding to each of the customer data for which the intention identification result is willing to accept the recommendation, includes:

When the intent identification result is willing to accept the recommendation, through the second pop-up node, a node timer is generated according to the node configuration data and path start time of the product recommendation configuration data, and the second pop-up node timer is obtained ;

Generate second pop-up window information according to the product recommendation basic information through the second pop-up window node;

Through the second pop-up window node, according to the second pop-up window node timer and the second pop-up window information, respectively, to the said customer data corresponding to each of the customer data whose intention identification result is willing to accept the recommendation The terminal sends the second pop-up request.
The computer-readable storage medium according to claim 15, wherein the step of obtaining the service operation results corresponding to each of the second pop-up window requests through the service operation node of the target product recommendation path includes:

Obtain the browsing results to be processed sent by each terminal according to the second pop-up request through the service operation node;

Obtain service application result data through the service operation node;

Through the service operation node, search according to the target customer data from the service application result data to obtain the service application search result, wherein the target customer data is any one of the browsed customer data sets customer data;

When the search result of the service application is successful, through the service operation node, determine that the service operation result corresponding to the target customer data is an applied service;

When the search result of the service application is failed, and the pending browsing result corresponding to the target customer data is browsed, determine the service operation corresponding to the target customer data through the service operation node The result is browsed but not applied;

When the search result of the service application is a failure, and the pending browsing result corresponding to the target customer data is not browsed, determine the service operation corresponding to the target customer data through the service operation node The result is not browsed and not applied.