Intelligent interaction method and system based on digital employees

Abstract

The invention discloses an intelligent interaction method and system based on digital employees, and relates to the technical field of data analysis, and the method comprises the steps: carrying out the current use stage analysis of a product based on product state information data and historical product scrapping condition data; performing question correlation analysis based on the analysis result of the current use stage of the product and the question information data of the current user; and replying questions of the user on the basis of a product question correlation analysis result, and on the basis of understanding the user questions, closely combining with the identified current-stage information comprehensive analysis of the product, and finally generating response contents which are highly adaptive to the current stage of the product, are balanced in risk and value, and are most targeted and effective. The user experience is obviously improved, and the decision-making efficiency and the resource utilization rationality are supported.

Description

TECHNICAL FIELD

[0001] The application belongs to the field of data analysis, and specifically relates to an intelligent interaction method and system based on digital employees. BACKGROUND

[0002] In traditional product support services (such as customer service systems, online help centers, and device maintenance platforms), the answers to user questions about product use, failure, maintenance, or scrapping are usually provided in the following ways: Systems mainly rely on predefined question and answer pairs (FAQs) or keyword matching. When a user's question is analyzed, the system searches the most relevant answer in the static knowledge base and returns it. This approach relies heavily on the coverage and update frequency of the knowledge base and cannot perceive the current actual state and historical evolution of the product. For example, for the same question of "device running slowly", a new device in the running-in period and an old device about to be scrapped may have completely different root causes and optimization suggestions, but the static knowledge base cannot make such a distinction. Some advanced systems consider the user's historical behavior or past ticket records to provide personalized services. However, this approach focuses on user-side information and lacks deep integration of real-time product state and full-life-cycle data. It cannot accurately determine the specific stage of the product (such as stable operation period, performance degradation period, and near-scrapping period), so the suggestions provided may not be suitable for the current actual situation of the product. Some systems attempt to set simple rules, such as providing general suggestions based on product purchase age. However, such rules are too general and static and cannot consider the actual use intensity of the product (such as running hours, load conditions, etc.), environmental factors, maintenance records, and pattern analysis based on historical scrapping data of similar products. It is difficult to accurately quantify the current health status and remaining life stage of the product. To solve the problems raised in the background, the application designs an intelligent interaction method and system based on digital employees. SUMMARY

[0003] To address the above technical deficiencies, the application provides an intelligent interaction method and system based on digital employees.

[0004] To solve the above technical problems, the application adopts the following technical solutions: The application provides an intelligent interaction method based on digital employees, which includes the following specific steps: S1, obtaining product state information data, historical product scrapping data, and current user question information data; S2, analyzing the current use stage of the product based on the product state information data and the historical product scrapping data; S3, performing question relevance analysis based on the product current use stage analysis result and the current user question information data; S4, replying to the user's question based on the product problem correlation analysis result.

[0005] It should be noted that, as a preferred technical solution of the intelligent interaction method based on digital employees, the specific steps of S1 are as follows: S11, obtaining product state information data through product operation logs, the product state information data including product maintenance frequency data and operation duration data of each period; S12, obtaining historical product scrap data through a database, wherein the historical product scrap data includes reliability analysis result data of each part at the time of historical product scrap and use condition data of the historical user using each vulnerability analysis result product; S13, collecting user question information data in real time through an online customer service system; S14, storing the obtained data in a storage component for use in the analysis process.

[0006] It should be noted that, as a preferred technical solution of the intelligent interaction method based on digital employees, S2 includes the following specific steps: S21, performing product part reliability analysis based on product state information data; S22, performing current product vulnerability analysis based on product part reliability analysis result and historical product scrap data; S23, performing product use stage prediction analysis based on current product vulnerability analysis result and use condition data of the historical user using each vulnerability analysis result product; S24, obtaining the current use stage of the product based on the product use stage prediction analysis result.

[0007] It should be noted that, as a preferred technical solution of the intelligent interaction method based on digital employees, the specific steps of S21 are as follows: obtaining product part maintenance frequency data and product operation duration data, and performing product part reliability analysis based on the product part maintenance frequency data and the product operation duration data, wherein the product part reliability analysis process is as follows: dividing the product part maintenance frequency data by the reference maintenance frequency to quantify the current health degree of the product part; dividing the product operation duration by the product theoretical use duration to quantify the current use degree of the product, and adding the current health degree of the product part and the current use degree after weighting to obtain the product part reliability analysis result. It should be noted that the current health degree of the product part is added to the current use degree for analysis, i.e., the physical maintenance data and the user use behavior data are superimposed, the reliability degree of the product in the technical aspect and the user experience is quantified, and the real reliability degree of the product part is reflected through reasonable dynamic weight distribution.

[0008] It should be noted that, as a preferred technical solution of the intelligent interaction method based on digital employees, the specific steps of S22 are: obtaining the reliability analysis results of each part of the product and the reliability analysis results of each part of the historical product at the time of scrapping, and performing product vulnerability analysis based on the reliability analysis results of each part of the product and the reliability analysis results of each part of the historical product at the time of scrapping, wherein the current product vulnerability analysis process is: dividing the reliability analysis results of each part of the product by the reliability analysis results of each part of the historical product at the time of scrapping, quantifying the vulnerability degree of each part of the product, and obtaining the current product vulnerability by summing and averaging the vulnerability degree of each part of the product. It should be noted that by summing and averaging the vulnerability degree of each part of the product, a single index representing the current vulnerability state of the whole product is obtained, which is helpful for comparing the robustness of different products or batches, and combined with historical scrapping data, the trend of product vulnerability over time can be observed to predict when it may reach the scrapping state.

[0009] It should be noted that, as a preferred technical solution of the intelligent interaction method based on digital employees, the specific steps of S23 are: obtaining the use case data consistent with the current product vulnerability analysis result from the historical user use of each vulnerability analysis result product use case data, taking the obtained use case data as the use case data corresponding to the current product, and performing product use stage prediction analysis based on the use case data corresponding to the current product. The product use stage prediction analysis process is: dividing the use frequency data corresponding to the current product by the reference use frequency, quantifying the current use ability of the product, dividing the use time data corresponding to the current product by the average use time of the product, quantifying the current use load degree of the product, and multiplying the current use ability of the product and the current use load degree to obtain the product use stage prediction analysis result. It should be noted that instead of considering frequency or time length in isolation, the two dimensions of "ability" and "load" are integrated through multiplication. Multiplication means that the two dimensions interact with each other, improving the accuracy of the product use stage prediction analysis result.

[0010] It should be noted that, as a preferred technical solution of the intelligent interaction method based on digital employees, the specific steps of S24 are: obtaining the product use stage prediction analysis result range corresponding to each use stage of the product, comparing the product use stage prediction analysis result with the product use stage prediction analysis result range corresponding to each use stage of the product, and if the product use stage prediction analysis result is in the product use stage prediction analysis result range corresponding to a certain use stage of the product, taking the use stage corresponding to the product use stage prediction analysis result range as the current use stage of the product. It should be noted that the abstract product use stage is converted into a quantifiable and objectively judged product use stage prediction analysis range, realizing accurate identification of the current state of the product.

[0011] It should be noted that as a preferred technical solution of the intelligent interaction method based on digital employees, the specific steps of S3 are: obtaining each keyword in the current user's question and each question keyword set in the current use stage of the product, and performing question relevance analysis based on each keyword in the current user's question and each question keyword set in the current use stage of the product, wherein the question relevance analysis process is: comparing each keyword in the current user's question with each question keyword set in the current use stage of the product to obtain the keyword coincidence degree corresponding to each question in the current user's question and the current use stage of the product, and taking the keyword coincidence degree corresponding to each question in the current user's question and the current use stage of the product as the question relevance analysis result; it should be noted that in the keyword comparison, in order to eliminate the expression differences of different users, the original question of the user should be replaced by synonyms and other technologies to avoid ambiguity and reduce user loyalty.

[0012] It should be noted that as a preferred technical solution of the intelligent interaction method based on digital employees, the specific steps of S4 are: arranging the keyword coincidence degree corresponding to each question in the current user's question and the current use stage of the product in descending order, taking the question arranged in the first place as the question that needs to be replied to the user, obtaining the reply information corresponding to the question, and replying to the user's question according to the reply information.

[0013] An intelligent interaction system based on digital employees, which is implemented based on the intelligent interaction method based on digital employees, and specifically includes a basic data acquisition module, a product use stage analysis module, a question relevance analysis module, and a question information reply module. The basic data acquisition module is used to acquire product state information data, historical product scrap data, and current user question information data. The product use stage analysis module is used to perform product current use stage analysis based on the product state information data and the historical product scrap data. The question relevance analysis module is used to perform question relevance analysis based on the product current use stage analysis result and the current user question information data. The question information reply module is used to reply to the user's question based on the product question relevance analysis result.

[0014] Compared with the prior art, the application has the beneficial effects that: the product state information data, the historical product scrapping situation data and the current user query information data are acquired; the product current use stage analysis is carried out based on the product state information data and the historical product scrapping situation data; the query correlation analysis is carried out based on the product current use stage analysis result and the current user query information data; the user query is replied based on the product query correlation analysis result, on the basis of understanding the user question, the product current stage information recognized is closely combined for comprehensive analysis, finally the response content which is highly adapted to the product current stage, balanced in risk and value, most targeted and effective is generated, the user experience, the decision-making efficiency and the resource utilization rationality are significantly improved. BRIEF DESCRIPTION OF DRAWINGS

[0015] Figure 1 For the overall flowchart of the intelligent interaction method based on digital employees of the application.

[0016] Figure 2 For the current use stage acquisition flowchart of the intelligent interaction method based on digital employees of the application.

[0017] Figure 3 For the overall framework diagram of the intelligent interaction system based on digital employees of the application. DETAILED DESCRIPTION

[0018] In order to better understand the application, various aspects of the application will be described in more detail with reference to the accompanying drawings.

[0019] To solve the technical problems proposed in the background art, the application provides a preferred embodiment: The specific content of the embodiment is: As shown in Figure 1 A kind of intelligent interaction method based on digital employees, it includes the following specific steps: S1, product state information data, historical product scrapping situation data and current user query information data are acquired; In the embodiment, the specific steps of S1 are as follows: the product state information data includes product maintenance frequency data and running time data in each period, and the acquisition process is as follows: obtaining product maintenance data and running time data from product operation logs, and obtaining product maintenance frequency data in each period (for example, 1 maintenance in the first year and 3 maintenances in the second year) from the product maintenance data; the historical product scrap data includes reliability analysis result data of each part of the historical product at the time of scrap and use condition data of the historical user using the product with each vulnerability analysis result, and the acquisition process is as follows: obtaining the reliability analysis result data of each part of the historical product at the time of scrap and the use condition data of the historical user using the product with each vulnerability analysis result from the database, and the use condition data of the historical user using the product with each vulnerability analysis result includes use frequency data and use time data; the current user question information data acquisition process is as follows: collecting user consultation questions (such as voice or text) in real time through an online customer service system. As shown in Figure 2 S2, product current use stage analysis is performed based on the product state information data and the historical product scrap data; S21, product part reliability analysis is performed based on the product state information data; In the embodiment, S21 includes the following specific steps: obtaining product part maintenance frequency data and product running time data, and performing product part reliability analysis based on the product part maintenance frequency data and the product running time data, wherein the product part reliability analysis process is as follows: dividing the product part maintenance frequency data by the reference maintenance frequency to quantify the current health degree of the product part; dividing the product running time by the product theoretical use time to quantify the current use degree of the product, and adding the current health degree of the product part and the current use degree after weighting to obtain the product part reliability analysis result; it should be noted that when analyzing the reliability of the product part, the implicit failure of the product part is identified by analyzing the product part maintenance frequency, and the product part defects are revealed by horizontal comparison; the product current use degree is analyzed to identify the product resource utilization rate; the current health degree of the product part is added to the current use degree for analysis, that is, the physical maintenance data and the user use behavior data are superimposed, the reliability degree of the product in the technical level and the user experience is quantified, and the real reliability degree of the product part is reflected through reasonable dynamic weight distribution.

[0020] S22, current product vulnerability analysis is performed based on the product part reliability analysis result and the historical product scrap data; In the embodiment, S22 includes the following specific steps: obtaining the reliability analysis results of each part of the product and the reliability analysis results of each part of the product at the time of historical product scrapping, and performing product vulnerability analysis based on the reliability analysis results of each part of the product and the reliability analysis results of each part of the product at the time of historical product scrapping, wherein the current product vulnerability analysis process is: dividing the reliability analysis results of each part of the product by the reliability analysis results of each part of the product at the time of historical product scrapping, quantifying the vulnerability degree of each part of the product, and performing summing and averaging operations on the vulnerability degree of each part of the product to obtain the current product vulnerability; it should be noted that the "reliability of each part of the product at the time of historical product scrapping" is used as a reference, which is equivalent to setting a "vulnerability" reference point (i.e. the scrapping state) for each part. By comparing with the "reference point", the state of each part of the current product relative to its end of life is truly reflected, and the vulnerability of each part is quantified. This quantification method avoids subjective judgment (for example, if the reliability of a part is low at the time of historical product scrapping, and the reliability analysis result of the current product at the part has no significant improvement, the vulnerability of the part may be high); by going deep into each part (component, subsystem) of the product, it can be accurately located which parts are the most fragile and closest to the scrapping state. By focusing the analysis on the parts with high product vulnerability, the occurrence of failures can be reduced and the service life of the product can be prolonged by identifying these key parts in advance; by summing and averaging the vulnerability degrees of each part of the product, a single index (product vulnerability) representing the current vulnerability state of the whole product is obtained, which is helpful for comparing the robustness of different products or batches. The lower the vulnerability value, the farther the product as a whole is from the scrapping state, and the more "robust" it is. At the same time, combined with historical scrapping data, the trend of product vulnerability over time can be observed, and when the product may reach the scrapping state can be predicted.

[0021] S23, product use stage prediction analysis is performed by the current product vulnerability analysis result and the use condition data of each vulnerability analysis result product used by the historical user; In the embodiment, the specific steps of S23 are as follows: obtaining the usage data consistent with the current product vulnerability analysis result from the historical user usage of each vulnerability analysis result product, taking the obtained usage data as the usage data corresponding to the current product, and performing product usage stage prediction analysis based on the usage data corresponding to the current product. The product usage stage prediction analysis process is as follows: dividing the usage frequency data corresponding to the current product by the reference usage frequency, quantifying the current available capacity of the product, dividing the usage time data corresponding to the current product by the average usage time of the product, quantifying the current usage load degree of the product, and multiplying the current available capacity of the product by the current usage load degree to obtain the product usage stage prediction analysis result. It should be noted that the specific indicators such as "usage frequency data corresponding to the current product", "usage time data", "reference usage frequency" and "average usage time of the product" are used as inputs for product usage stage prediction analysis, which is completely based on actual quantifiable usage data, avoiding subjective speculation or deviation caused by experience. By dividing the current data by the reference value (reference usage frequency, average usage time), the data is standardized, so that the data of different products and different user groups can be compared, and the deviation of the current vulnerability product state from a benchmark value (reference value) can be measured, eliminating the interference caused by the absolute numerical size. The focus is on the relative performance of each vulnerability product, and the current available capacity of each vulnerability product is quantified, that is, the remaining potential of each vulnerability product is activated or touched. The current usage load degree of each vulnerability product is quantified, that is, the depth of user investment or dependence on the product (product usage time relative to the benchmark) is quantified. The higher the value, the more time the user invests each time, and the higher the product dependence or the heavier the task. Instead of looking at the usage frequency or time of each vulnerability product in isolation, the "available capacity" (i.e. breadth) and "load degree" (i.e. depth) are combined through multiplication. Multiplication means that the two dimensions interact with each other (for example, the product usage frequency is high but the usage time is very short each time, that is, high capacity multiplied by low load; the product usage frequency is low but the usage time is long each time, that is, low capacity multiplied by high load, and the product usage stage prediction analysis result may not be high, which may indicate that the product is in different stages (such as the exploration period or the decline period)). The product usage stage prediction analysis result is analyzed comprehensively to improve the accuracy of the product usage stage prediction analysis result.

[0022] S24, obtaining the current product usage stage from the product usage stage prediction analysis result; In the embodiment, the specific steps of S24 are: obtaining the product use stage prediction analysis result range corresponding to each use stage of the product, comparing the product use stage prediction analysis result with the product use stage prediction analysis result range corresponding to each use stage of the product, if the product use stage prediction analysis result is in the product use stage prediction analysis result range corresponding to a use stage of the product, taking the use stage corresponding to the product use stage prediction analysis result range as the current use stage of the product; it should be noted that the abstract product use stage is converted into the quantifiable and objective product use stage prediction analysis range, the accurate identification of the current state of the product is realized, and the product use stage (such as “running-in period”, “stable period”, “decline period” and “scrap period”) is a fuzzy concept in the traditional sense. By establishing the product use stage prediction analysis range corresponding to each stage, an objective and unified numerical standard is provided for product life cycle division, the subjectivity of human judgment is eliminated, and the stage judgment of different products and different times is comparable.

[0023] S3, performing question relevance analysis based on the product current use stage analysis result and the current user question information data; In the embodiment, the specific steps of S3 are: obtaining each keyword in the current user question and each problem keyword set in the product current use stage, and performing question relevance analysis according to each keyword in the current user question and each problem keyword set in the product current use stage, wherein the question relevance analysis process is: comparing each keyword in the current user question with each problem keyword set in the product current use stage, obtaining the keyword coincidence degree corresponding to each problem in the current user question and the product current use stage, and taking the keyword coincidence degree corresponding to each problem in the current user question and the product current use stage as the question relevance analysis result; it should be noted that in the keyword comparison, in order to eliminate the expression differences of different users (such as the user may use different words to describe the same demand), the original question of the user question should be replaced by synonym replacement technology (such as the words “login failure”, “unable to log in” and “account cannot enter” can be uniformly mapped to the standard keyword “login failure”) at the same time, the core demand of the user question (such as function consultation, fault repair, purchase decision, etc.) is quickly identified, ambiguity is avoided, and user loyalty is reduced.

[0024] S4, replying to the user question based on the product question relevance analysis result; In the embodiment, the specific steps of S4 are: arranging the keywords coincidence degrees of the current user's question and the questions in the current use stage of the product in descending order, taking the question ranked first as the question to be answered by the user, obtaining the reply information corresponding to the question, and replying to the user's question according to the reply information; it should be noted that comparing the keywords coincidence degrees of the user's question and the questions in the current use stage of the product obtained by analysis saves computing resources and improves the accuracy of the reply to the user's question.

[0025] According to the above implementation, the embodiment has the following advantages over the prior art: the embodiment obtains product state information data, historical product scrap situation data, and current user question information data; performs product current use stage analysis based on the product state information data and the historical product scrap situation data; performs question relevance analysis based on the product current use stage analysis result and the current user question information data; and replies to the user's question based on the product question relevance analysis result. On the basis of understanding the user's question, the product current stage information recognized is closely combined for comprehensive analysis, and finally the response content that is highly adapted to the current stage of the product, balanced in risk and value, most targeted and effective is generated, which significantly improves the user experience, supports the decision-making efficiency, and rationalizes the resource utilization.

[0026] As shown in Figure 3 The embodiment also provides an intelligent interaction system based on digital employees, which is implemented based on the above-mentioned intelligent interaction method based on digital employees, and specifically includes a basic data acquisition module, a product use stage analysis module, a question relevance analysis module, and a question information reply module. The basic data acquisition module is used to acquire product state information data, historical product scrap situation data, and current user question information data. The product use stage analysis module is used to perform product current use stage analysis based on the product state information data and the historical product scrap situation data. The question relevance analysis module is used to perform question relevance analysis based on the product current use stage analysis result and the current user question information data. The question information reply module is used to reply to the user's question based on the product question relevance analysis result.

[0027] The specific steps of each unit module in the above-mentioned intelligent interaction system based on digital employees for implementing the corresponding functions can refer to the steps in the above-mentioned embodiment of the intelligent interaction method based on digital employees, which will not be repeated here.

[0028] The above description is only the preferred embodiment of the present application and the explanation of the technical principles. It should be understood by those skilled in the art that the application scope of the present application is not limited to the technical solutions with the specific combination of the above technical features, and should also cover other technical solutions formed by combining the above technical features or their equivalent features without departing from the concept of the application. For example, the technical solutions formed by replacing the above features with the technical features with similar functions applied in the present application (but not limited to) with each other.

Claims

1. A smart interaction method based on digital employees, characterized in that, include: S1. Obtain product status information data, historical product scrapping data, and current user question information data; S2. Analyze the current usage stage of the product based on product status information data and historical product scrapping data; S3. Conduct question correlation analysis based on the analysis results of the current product usage stage and the current user question information data; S4. Respond to user questions based on the product issue correlation analysis results.

2. The intelligent interaction method based on digital employees as described in claim 1, characterized in that, S2 includes the following specific steps: S21. Perform reliability analysis on various parts of the product based on product status information data; S22. Conduct a vulnerability analysis of the current product based on the reliability analysis results of each part of the product and historical product scrapping data; S23. Based on the current product fragility analysis results and historical user data on the usage of each product according to the fragility analysis results, predict and analyze the product usage stage. S24. The current usage stage of the product is obtained from the product usage stage prediction analysis results.

3. The intelligent interaction method based on digital employees as described in claim 2, characterized in that, The specific steps of S21 are as follows: obtain the maintenance frequency data and product runtime data for each part of the product, and perform reliability analysis on each part of the product based on the maintenance frequency data and product runtime data. The reliability analysis process for each part of the product is as follows: divide the maintenance frequency data of each part of the product by the reference maintenance frequency to quantify the current health status of each part of the product; divide the product runtime by the theoretical usage time of the product to quantify the current usage level of the product; and add the current health status and current usage level of each part of the product together to obtain the reliability analysis result of each part of the product.

4. The intelligent interaction method based on digital employees as described in claim 3, characterized in that, The specific steps of S22 are as follows: obtain the reliability analysis results of each part of the product and the reliability analysis results of each part when the product was scrapped in the past; perform product vulnerability analysis based on the reliability analysis results of each part of the product and the reliability analysis results of each part when the product was scrapped in the past; wherein, the current product vulnerability analysis process is as follows: divide the reliability analysis results of each part of the product by the reliability analysis results of each part when the product was scrapped in the past, quantify the vulnerability of each part of the product at present, and sum and average the vulnerability of each part of the product at present to obtain the current product vulnerability.

5. The intelligent interaction method based on digital employees as described in claim 4, characterized in that, The specific steps of S23 are as follows: obtain usage data consistent with the current product's vulnerability analysis results from the historical user usage data of each product with vulnerability analysis results, use the obtained usage data as the usage data corresponding to the current product, and perform product usage stage prediction analysis based on the usage data corresponding to the current product. The product usage stage prediction analysis process is as follows: divide the current product's usage frequency data by the reference usage frequency to quantify the product's current usable capacity, divide the current product's usage duration data by the product's average usage duration to quantify the product's current usage load level, and multiply the product's current usable capacity by the current usage load level to obtain the product usage stage prediction analysis result.

6. The intelligent interaction method based on digital employees as described in claim 5, characterized in that, The specific steps of S24 are as follows: obtain the product usage stage prediction analysis result range corresponding to each usage stage of the product, compare the product usage stage prediction analysis result with the product usage stage prediction analysis result range corresponding to each usage stage of the product, and if the product usage stage prediction analysis result is within the product usage stage prediction analysis result range corresponding to a certain usage stage of the product, then take the usage stage corresponding to the product usage stage prediction analysis result range as the current usage stage of the product.

7. The intelligent interaction method based on digital employees as described in claim 6, characterized in that, The specific steps of S3 are as follows: obtain the keyword set of each question in the current user's question and the keyword set of each question in the current stage of product use; perform question correlation analysis based on the keyword set of each question in the current user's question and the keyword set of each question in the current stage of product use; wherein, the question correlation analysis process is as follows: compare the keyword set of each question in the current user's question with the keyword set of each question in the current stage of product use to obtain the keyword overlap degree corresponding to each question in the current stage of product use; and take the keyword overlap degree corresponding to each question in the current stage of product use as the question correlation analysis result.

8. The intelligent interaction method based on digital employees as described in claim 7, characterized in that, The specific steps of S4 are as follows: sort the current user's question in descending order of keyword overlap with the questions in the current stage of product use, take the question ranked first as the question that needs to be answered, obtain the answer information corresponding to the question, and answer the user's question based on the answer information.

9. A digital employee-based intelligent interaction system, implemented based on the digital employee-based intelligent interaction method according to any one of claims 1-8, characterized in that, Specifically, it includes a basic data acquisition module, a product usage stage analysis module, a question correlation analysis module, and a question information response module. The basic data acquisition module is used to acquire product status information data, historical product scrapping data, and current user question information data. The product usage stage analysis module is used to analyze the current usage stage of the product based on product status information data and historical product scrapping data. The question correlation analysis module is used to perform question correlation analysis based on the analysis results of the current product usage stage and the current user question information data; The question response module is used to answer user questions based on the product question correlation analysis results.

Images