An ai-driven customer sentiment dynamic portrait and intelligent touch system and method

Abstract

The application discloses an AI driving-based customer emotion dynamic portrait and intelligent touch system and method, relates to the technical field of customer emotion management, stores customer service historical data, analyzes transaction rates and quantities of various products, counts average transaction rates and quantities, collects customer historical data of various products, sets an interest transfer period, analyzes an interest coefficient of customers to products, counts product transaction probabilities within a customer period, analyzes a chain prediction coefficient of customers, analyzes a comprehensive intention coefficient in combination with the interest coefficient and the chain prediction coefficient, selects a priority recommended product, collects a transaction probability of the customer to the priority recommended product, analyzes matching degrees of customer service and the customer to products, calculates an average value of the customer matching degrees, informs customer service reception from high to low, and until customer service is accepted, when the customer reselects customer service, the demand for transaction rates in the analysis of the matching degrees is adjusted, and the application enables the whole service process to be more adaptive by establishing a flexible service adjustment mechanism.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of customer relationship management, in particular to a customer relationship dynamic portrait and intelligent touch system and method based on AI driving. BACKGROUND

[0002] In the traditional customer relationship management and customer service allocation technology, there are many pain points that restrict service efficiency and customer experience. The past customer service allocation mechanism often relies on manual experience or random assignment, lacks systematic analysis of customer historical service data, and is difficult to accurately grasp the service ability and transaction level of each customer in different product categories. This blind allocation method often leads to a mismatch between customer expertise and customer demand, resulting in waste of service resources and difficulty in improving transaction conversion rate, which cannot meet the needs of fine operation. Traditional customer demand analysis focuses on current interest preferences and lacks dynamic monitoring of customer interest shift cycles and deep mining of historical transaction patterns. Due to the failure to capture the demand migration trend of customers among different product categories, recommended services often remain at the static level and cannot predict potential demand in advance. This limitation makes the recommended products less consistent with the real needs of customers, and customers are not receptive to the recommended content, which further affects the cultivation and improvement of customer loyalty. The traditional customer service system lacks a flexible response mechanism when facing dynamic demand adjustment in the customer service process. When customers propose personalized needs such as changing customer service, the system is difficult to quickly optimize the matching scheme, often still following the inherent allocation logic, resulting in persistent service mismatch problems. This rigid handling method not only easily causes customer dissatisfaction, but also may cause customer loss, which is difficult to adapt to the individualization and instantaneity requirements of modern customers for service experience, and a more adaptive dynamic adjustment mechanism is needed to optimize the service process. SUMMARY

[0003] The purpose of the present application is to provide a customer relationship dynamic portrait and intelligent touch system and method based on AI driving to solve the problems raised in the background art.

[0004] To solve the above technical problems, the present application provides the following technical solution: a customer relationship dynamic portrait and intelligent touch method based on AI driving, comprising the following steps:

[0005] S1, store customer historical data, analyze transaction rates and quantities of various products, and count average transaction rates and quantities;

[0006] S2, collect customer historical data of various products, set an interest shift cycle, and analyze customer interest coefficients for products;

[0007] S3, count product transaction probabilities within a customer cycle, and analyze customer chain prediction coefficients;

[0008] S4, analyze the comprehensive intention coefficient by combining the interest coefficient and the linkage prediction coefficient, and select the priority recommended product;

[0009] S5, collect the transaction probability of the customer to the priority recommended product, and analyze the matching degree of the customer service and the customer to the product;

[0010] S6, calculate the average value of the customer matching degree, and notify the customer service reception from high to low, until a customer service accepts, and adjust the demand for analyzing the matching degree and the transaction rate when the customer reselects the customer service.

[0011] Further, in step S1, after authorization, the historical data of the customer service personnel is stored in the database, the state of the customer service personnel is queried, and the idle customer service personnel includes {A1, A2, …, AX}, where X represents the number of idle customer service personnel, and Ax represents the xth idle customer service personnel. The customer service personnel A is analyzed, the current time point is taken as the time end point, the historical data of the customer service personnel in T monitoring time periods is queried, the monitoring time period with the current time point as the end point is the Tth monitoring time period, T is the number of monitoring time periods, the tth monitoring time period is analyzed, t=1,2,…,T, the product category is Y categories, in the tth monitoring time period, the transaction rate of the customer service personnel A to the yth product is Cy, the transaction quantity of the customer service personnel A to the yth product is Dy, the average transaction rate of the customer service personnel A to Y products is Cy, and the average transaction quantity of the customer service personnel A to Y products is Dy. x ,…,A X} wherein X represents the number of idle customer service personnel, and Ax represents the xth idle customer service personnel. The customer service personnel A is analyzed, the current time point is taken as the time end point, the historical data of the customer service personnel in T monitoring time periods is queried, the monitoring time period with the current time point as the end point is the Tth monitoring time period, T is the number of monitoring time periods, the tth monitoring time period is analyzed, t=1,2,…,T, the product category is Y categories, in the tth monitoring time period, the transaction rate of the customer service personnel A to the yth product is Cy, the transaction quantity of the customer service personnel A to the yth product is Dy, the average transaction rate of the customer service personnel A to Y products is Cy, and the average transaction quantity of the customer service personnel A to Y products is Dy. x x x x_y x_y x x x 1_y 2_y x_y X_y 1_y 2_y x_y X_y x X x X ​​​​​​​​​​​​​​​​​​​};By authorized storage customer service history data and query state, can fully grasp the idle customer service resources. The different time period of each customer service for each type of product transaction in-depth analysis, can accurately grasp each customer in different product service ability characteristics and performance level. Based on these analysis results, can be based on product category and customer expertise for scientific matching, improve the service of the pertinence and effectiveness, and then improve the product conversion rate. At the same time, through the calculation of the average conversion rate and the number of customer service evaluation provides objective basis, help management optimization personnel allocation and resource scheduling, enhance the overall efficiency of customer service team, for business development to provide data support and decision reference, promote service quality improvement.

[0012] Further, in step S2, after authorization, the history data of the nth customer for Y type product is collected, the interest transfer period is set, the interest transfer period with the current time as the end point is set as the first interest transfer period, I is the number of set interest transfer period, the i-th interest transfer period is analyzed, i=1, 2,…,I, in the i-th interest transfer period, the transaction volume of the customer for Y type product is {E1,E2,…,E y ,…,E α ,…,E Y}, wherein E y represents the transaction volume of the customer for the y type product, E α represents the transaction volume of the customer for the alpha type product, and then the interest coefficient w y of the customer for the y type product when the customer applies for customer service is obtained:

[0013] ;

[0014] By authorized collection of customer history data for each type of product and analysis of interest transfer period, the interest preference and trend of customers at different stages for different products can be accurately captured. This in-depth analysis of customer interest dynamics can help enterprises clearly understand customer demand trends, accurately match corresponding product services based on the interest coefficient when customers apply for customer service, and improve the pertinence and fit of services. At the same time, based on the analysis of interest period, enterprises can develop product recommendation strategies and provide scientific basis for optimizing resource allocation, making services more in line with customer real needs, enhancing customer recognition and satisfaction of services, and then promoting the improvement of customer transaction intention, helping enterprises to achieve more accurate customer service and more efficient business conversion.

[0015] Further, in step S3, when the customer transacts the alpha type product in an interest transfer period, the probability of transacting the y type product in the next interest transfer period is F α_y, and the transaction probability of the customer to the y-th product in the next interest transfer period is {F 1_y ,F 2_y ,…,F α_y ,…,F Y_y} when the customer transacts the Y-th product in an interest transfer period. y :

[0016] ;

[0017] wherein f α_y is a transaction prediction coefficient of the y-th product in the next interest transfer period according to the transaction of the customer to the a-th product in an interest transfer period, and the transaction prediction coefficient f α_y =F α_y when the customer transacts the a-th product in an interest transfer period, and f α_y =0 when the customer does not transact the a-th product in an interest transfer period. By statistically obtaining the transaction correlation probability of the customer to different products in an interest transfer period, the continuity and potential trend of the consumption behavior of the customer can be deeply mined. The chain analysis of the transaction probability can accurately predict the purchase possibility of the customer to various products in the subsequent period, so that the enterprise can clearly master the dynamic evolution path of the demand of the customer. Therefore, the enterprise can develop a response strategy in advance according to the interest transfer of the customer, provide more forward-looking product recommendations in the customer service process, and change the service from passive response to active prediction. Meanwhile, the analysis provides data support for the enterprise to optimize resource allocation and plan product service direction, helps the enterprise to more accurately meet the potential demand of the customer, improves the predictability and effectiveness of customer service, enhances the dependence and loyalty of the customer to the service, and further creates more transaction opportunities and promotes the sustainable growth of the business.

[0018] Further, in step S4, a comprehensive intention coefficient M y of the customer to the y-th product is calculated according to the interest coefficient w y of the customer to the y-th product and the chain prediction coefficient W y of the customer to the y-th product, and the comprehensive intention coefficient M y is obtained by weighting the interest coefficient w y and the comprehensive intention coefficient M y , wherein y=1, 2, …, Y, and the comprehensive intention coefficients of the customer to the Y-th product are {M1, M2, …, M y ,…,M Y} and selecting the beta product with the highest comprehensive intention coefficient as the priority recommended product for the customer; the comprehensive intention coefficient is calculated by combining the customer's interest coefficient for the product and the chain prediction coefficient, which can deeply analyze the customer's real intention from the static interest preference and dynamic demand evolution. This evaluation method that integrates multi-dimensional data can accurately capture the customer demand change trajectory, enabling the enterprise to break free from the limitations of single-dimensional analysis. Based on the comprehensive intention coefficient, the priority recommended product is screened, which can realize the full-chain service upgrade from customer interest insight to demand prediction, making the recommended service not only fit the current preference but also anticipate future demand. This not only improves the accuracy and predictability of customer service, enhances the customer's acceptance and satisfaction of the recommended product, but also helps the enterprise optimize resource allocation, improve product recommendation efficiency, create more transaction opportunities, and promote the intelligent and precise upgrade of the service mode, achieving efficient matching of customer demand and enterprise service.

[0019] Further, in step S5, when the customer applies for customer service, the authorized monitoring of the customer's beta priority recommended product includes the yth product, and the transaction probability H n_y , n = 1, 2, …, N, and then the transaction probability of the nth customer for the yth product is obtained {H 1_y ,H 2_y ,…,H n_y ,…,H N_y}, wherein the maximum transaction rate is H max_y , and then the matching degree J x_y of the xth customer service for the nth customer about product y is calculated:

[0020] ;

[0021] wherein k1 is the influence weight of transaction rate on matching degree, k1= , k2 is the influence weight of transaction volume on matching degree, k2= ; by obtaining the customer's priority recommended product and collecting the transaction probability, the customer's inclination for specific products can be accurately grasped. Based on the transaction probability data of different customers, combined with the influence weights of transaction rate and transaction volume on matching degree, the adaptation degree of customer service and customer in specific product service can be scientifically calculated. This matching mechanism can realize the precise docking of customer service resources and customer demand, and let the customer service with corresponding product service advantage provide service for customers with high transaction intention, improving the pertinence and effectiveness of service. At the same time, this method provides a scientific basis for the enterprise to optimize the allocation of customer service resources, helps to improve customer service efficiency, enhances customer recognition of service, and thus improves product transaction probability and promotes the improvement of enterprise service quality.

[0022] Further, in step S6, the average matching degree of the Xth customer service to the nth customer about the beta product is calculated, and the customer service is notified of the reception of the customer in descending order of the average matching degree until a customer service accepts the notification;

[0023] If the customer applies for a change of customer service during the customer service reception, the influence weight of the transaction rate on the matching degree is increased, the average matching degree of the Xth customer service to the nth customer about the beta product is reanalyzed, and the customer service is notified of the reception of the customer in descending order of the average matching degree until a customer service accepts the notification; by calculating the average matching degree of the customer service and the customer and notifying in descending order, the precise connection of customer service resources and customer demand can be realized, ensuring that high-matching-degree customer services are preferentially accepted, and the efficiency and quality of the first service matching are improved. When the customer applies for a change of customer service, the system automatically increases the influence weight of the transaction rate on the matching degree and reanalyzes the order. This dynamic adjustment mechanism can optimize the matching logic according to actual service feedback, better meet the personalized needs and real service expectations of customers. This process not only improves the relevance of customer service reception and customer satisfaction, but also reduces service interruptions through continuous optimization of matching strategies, enhances customer trust and reliance on services, and helps enterprises efficiently dispatch customer service resources, reduce resource waste, improve overall service efficiency and product transaction probability, and promote the formation of a more intelligent and flexible customer service system.

[0024] An AI-driven customer emotion dynamic portrait and intelligent touch system, the system comprising: a customer service data storage module, a customer interest coefficient analysis module, a transaction probability chain analysis module, a comprehensive intention coefficient analysis module, a matching degree analysis module, and a customer service order management module;

[0025] The customer service data storage module is used to store customer service historical data, analyze the transaction rate and quantity of various products, and calculate the average transaction rate and quantity;

[0026] The customer interest coefficient analysis module is used to collect customer historical data of various products, set an interest transfer period, and analyze the customer's interest coefficient for the product;

[0027] The transaction probability chain analysis module is used to calculate the product transaction probability within the customer cycle and analyze the customer's chain prediction coefficient;

[0028] The comprehensive intention coefficient analysis module is used to analyze the comprehensive intention coefficient by combining the interest coefficient and the chain prediction coefficient, and to select the preferred recommended product;

[0029] The matching degree analysis module is used to collect the transaction probability of the customer for the preferred recommended product and analyze the matching degree of the customer service and the customer for the product;

[0030] The customer service order management module is used for calculating the average value of customer matching degree, and informing customer service reception from high to low until a customer service accepts, and adjusting the demand for analyzing the matching degree when the customer reselects the customer service.

[0031] Compared with the prior art, the beneficial effects achieved by the present application are: on the one hand, by analyzing the historical transaction data of customer service personnel, the service ability and transaction performance of each customer service in different product categories are accurately mastered, and a scientific matching degree evaluation model is constructed by combining the interest preference of customers for products and the transaction probability, so that the accurate docking of customer service and customer demand is realized, and the service efficiency and transaction conversion rate are greatly improved. This data-driven allocation mechanism avoids the blindness of traditional random allocation and makes the service resources more reasonably utilized.

[0032] On the one hand, with the help of dynamic monitoring and chain prediction analysis of customer interest transfer period, the demand migration law of customers among different product categories can be captured, not only according to the current interest coefficient, but also combining the chain influence of historical transactions to predict potential demand, forming a comprehensive intention coefficient to recommend priority products. This forward-looking demand prediction makes the recommended service more in line with the real needs of customers, enhances the acceptance of customers to the recommended products, and further improves customer stickiness.

[0033] On the other hand, a flexible service adjustment mechanism is established, when the customer proposes to change the customer service in the reception, the system automatically increases the influence weight of transaction rate on matching degree, and reoptimizes the customer matching scheme. This dynamic response mechanism can timely solve the individualized needs of customer service, avoid customer loss caused by service mismatch, and further optimize the matching logic through weight adjustment, so that the entire service process is more adaptive and humanized, and the customer satisfaction and service quality are continuously improved. BRIEF DESCRIPTION OF DRAWINGS

[0034] The accompanying drawings are used to provide a further understanding of the present application, and constitute a part of the specification, together with the embodiments of the present application, to explain the present application, and do not constitute a limitation of the present application. In the drawings:

[0035] Figure 1 is a structural diagram of an AI-driven customer emotion dynamic portrait and intelligent touch system of the present application;

[0036] Figure 2 is a flowchart of an AI-driven customer emotion dynamic portrait and intelligent touch method of the present application. DETAILED DESCRIPTION

[0037] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative effort belong to the scope of the present application.

[0038] Please refer to Figure 1 and Figure 2 The present application provides a technical solution: an AI-driven customer emotion dynamic portrait and intelligent touch method, comprising the following steps:

[0039] S1, store customer service historical data, analyze the transaction rate and quantity of various products, and count the average transaction rate and quantity;

[0040] S2, collect customer historical data of various products, set an interest transfer period, and analyze the customer interest coefficient of the product;

[0041] S3, count the product transaction probability in the customer cycle, and analyze the customer chain prediction coefficient;

[0042] S4, analyze the comprehensive intention coefficient by combining the interest coefficient and the chain prediction coefficient, and select the priority recommended product;

[0043] S5, collect the transaction probability of the customer to the priority recommended product, and analyze the matching degree of the customer service and the customer to the product;

[0044] S6, calculate the average value of the customer matching degree, notify the customer service reception from high to low, and adjust the demand for transaction rate when analyzing the matching degree when the customer reselects the customer service. In step S1, after authorization, the historical data of the customer service personnel is stored in the database, the state of the customer service personnel is queried, the idle customer service personnel includes {A1, A2, …, Ax}, where X represents the number of idle customer service personnel, Ax represents the xth idle customer service personnel, the customer service personnel Ax is analyzed, the current time point is taken as the time end point, the historical data of the customer service personnel in T monitoring time periods is queried, the monitoring time period with the current time point as the end point is the Tth monitoring time period, T is the number of monitoring time periods set, the tth monitoring time period is analyzed, t=1, 2, …, T, there are Y types of products, in the tth monitoring time period, the transaction rate of the customer service personnel Ax to the yth product is Cy, the transaction quantity of the customer service personnel Ax to the yth product is Dy, the average transaction rate of the customer service personnel Ax to Y products is Cy, and the average transaction quantity of the customer service personnel Ax to Y products is Dy. x ,…,A X} where X represents the number of idle customer service personnel, Ax represents the xth idle customer service personnel, the customer service personnel Ax is analyzed, the current time point is taken as the time end point, the historical data of the customer service personnel in T monitoring time periods is queried, the monitoring time period with the current time point as the end point is the Tth monitoring time period, T is the number of monitoring time periods set, the tth monitoring time period is analyzed, t=1, 2, …, T, there are Y types of products, in the tth monitoring time period, the transaction rate of the customer service personnel Ax to the yth product is Cy, the transaction quantity of the customer service personnel Ax to the yth product is Dy, the average transaction rate of the customer service personnel Ax to Y products is Cy, and the average transaction quantity of the customer service personnel Ax to Y products is Dy. x x x x_y x_y x ​​​​​​x , the average transaction quantity of the Y-type product is D x , and then x = 1, 2, …, X is substituted into the above equation to obtain the conversion rates of the X customer service personnel for the y-type product, which are {C 1_y , C 2_y , …, C x_y , …, C X_y}, the transaction quantities of the X customer service personnel for the y-type product are {D 1_y , D 2_y , …, D x_y , …, D X_y}, the average conversion rates of the X customer service personnel for the Y-type product are {C1, C2, …, C x , …, C X}, and the average transaction quantities of the X customer service personnel for the Y-type product are {D1, D2, …, D x , …, D X} ; the authorized storage of customer service historical data and the query of the state can comprehensively grasp the idle customer service resources. The in-depth analysis of the conversion of various products by various customer services in different time periods can accurately grasp the ability characteristics and performance level of each customer service in different product services. Based on these analysis results, the products can be matched with the customer service expertise according to the product category, the service pertinence and effectiveness can be improved, and the product conversion rate can be improved. At the same time, the average conversion rate and quantity are calculated to provide an objective basis for customer service evaluation, help the management party to optimize personnel allocation and resource scheduling, enhance the overall efficiency of the customer service team, provide data support and decision reference for business development, and promote the improvement of service quality.

[0045] In step S2, after authorization, the historical data of the nth customer for the Y-type product is collected, the interest transfer period is set, the interest transfer period with the current time as the end point is set as the i th interest transfer period, I is the number of set interest transfer periods, the i th interest transfer period is analyzed, i = 1, 2, …, I, and in the i th interest transfer period, the transaction quantity of the customer for the Y-type product is {E1, E2, …, E y , …, E α , …, E Y}, wherein E y represents the transaction quantity of the customer for the y-type product, E α represents the transaction quantity of the customer for the a-type product, and then the interest coefficient w y of the customer for the y-type product when the customer applies for customer service is obtained:

[0046] ;

[0047] By authorized collection of customers' historical data on various products and analysis of interest shift cycle, the interest preferences and change trends of customers on different products at different stages can be accurately captured. This deep analysis of customers' interest dynamics can help enterprises clearly understand the demand trends of customers, accurately match the corresponding product services according to the interest coefficient when customers apply for customer service, and improve the pertinence and fitness of services. At the same time, the analysis based on the interest cycle can provide scientific basis for enterprises to formulate product recommendation strategies and optimize resource allocation, so that the services are more in line with the real needs of customers, and the recognition and satisfaction of customers to the services are enhanced, thereby promoting the improvement of customers' transaction intention and helping enterprises to achieve more accurate customer service and more efficient business conversion.

[0048] In step S3, the probability of existing transaction of the y-th product in the next interest shift cycle when the customer transacts the a-th product in an interest shift cycle is F α_y , and a = 1, 2, …, Y is substituted to obtain the transaction probability of the y-th product in the next interest shift cycle when the customer transacts Y products in an interest shift cycle as {F 1_y , F 2_y , …, F α_y , …, F Y_y}, the transaction of the customer on the Y products in the T-th interest shift cycle is counted, and the chain prediction coefficient W y of the customer on the y-th product is calculated.

[0049] ;

[0050] wherein f α_y is the transaction prediction coefficient of the y-th product in the next interest shift cycle according to the transaction of the customer on the a-th product in an interest shift cycle, and the transaction prediction coefficient f α_y of the y-th product in the next interest shift cycle when the customer transacts the a-th product in an interest shift cycle is F α_y ; otherwise f α_y= 0; By statistically analyzing the transaction association probability of customers for different products in the interest shift cycle, the continuity and potential trend of customer consumption behavior can be deeply mined. This kind of chain analysis of transaction probability can accurately predict the purchase probability of customers for various products in the subsequent cycle, so that enterprises can clearly master the dynamic evolution path of customer demand. Therefore, enterprises can develop countermeasures in advance in response to the interest shift of customers, provide more forward-looking product recommendations in the customer service process, and change the service from passive response to active prediction. At the same time, this analysis provides data support for enterprises to optimize resource allocation and plan product service direction, helps enterprises to better meet the potential needs of customers, improves the predictability and effectiveness of customer service, enhances the dependence and loyalty of customers to the service, and thus creates more transaction opportunities and promotes the sustainable growth of business.

[0051] In step S4, the interest coefficient w y of the customer for the yth product is calculated according to the customer's interest coefficient w y and the chain prediction coefficient W y of the yth product, and the comprehensive intention coefficient M y of the customer for the yth product is calculated by weighting the interest coefficient w y and the comprehensive intention coefficient M y , that is, y = 1, 2, …, Y, to obtain the comprehensive intention coefficient {M y , M Y} of the customer for Y products, and the βth product with the highest comprehensive intention coefficient is selected as the priority recommended product for the customer; By calculating the comprehensive intention coefficient by combining the customer's interest coefficient and the chain prediction coefficient, the real intention of the customer can be deeply analyzed from the static interest preference and the dynamic demand evolution. This evaluation method that integrates multi-dimensional data can accurately capture the demand change trajectory of the customer, so that the enterprise can break free from the limitations of single-dimensional analysis. Based on the comprehensive intention coefficient, the priority recommended product can be selected to realize the whole-chain service upgrade from customer interest insight to demand prediction, so that the recommended service not only meets the current preference but also anticipates future demand. This not only improves the accuracy and predictability of customer service, enhances the acceptance and satisfaction of customers to the recommended product, but also helps enterprises to optimize resource allocation, improve product recommendation efficiency, create more transaction opportunities, promote the upgrading of service mode to intelligent and accurate direction, and realize the efficient matching of customer demand and enterprise service.

[0052] In step S5, when the customer applies for customer service, the priority recommended product of the customer is monitored after authorization, including the yth product, the transaction probability H n_y of the nth customer for the yth product is collected, that is, n = 1, 2, …, N, and then the transaction probability {H 1_y , H 2_y , …, Hn_y ..., H N_y}, wherein the maximum transaction rate is H max_y , and then calculating the matching degree J of the xth customer service to the nth customer about product y x_y :

[0053] ;

[0054] wherein k1 is the influence weight of transaction rate on matching degree, k1= , k2 is the influence weight of transaction volume on matching degree, k2= ; By obtaining the customer's preferred recommended product and collecting the transaction probability, the customer's inclination to a specific product can be accurately grasped. Based on the transaction probability data of different customers, combined with the influence weights of transaction rate and transaction volume on matching degree, the adaptation degree of customer service and customer in specific product service can be scientifically calculated. This matching mechanism can realize the precise docking of customer service resources and customer demand, and let the customer service with corresponding product service advantage provide service for high transaction intention customers, improve the pertinence and effectiveness of service. At the same time, this method provides a scientific basis for enterprises to optimize customer service resource allocation, helps to improve customer service efficiency, enhances customer's recognition of service, and then improves product transaction probability, and promotes the improvement of enterprise service quality.

[0055] In step S6, the average matching degree of the xth customer service to the nth customer about product β is calculated, and the customer service is notified to receive the customer according to the average matching degree from high to low, until there is a customer service accepting the customer service notification;

[0056] If the customer proposes to change the customer service in the customer service reception, the influence weight of transaction rate on matching degree is increased, the average matching degree of the xth customer service to the nth customer about product β is reanalyzed, and the customer service is notified to receive the customer according to the average matching degree from high to low, until there is a customer service accepting the customer service notification; By calculating the average matching degree of customer service and customer and notifying according to high and low, the precise docking of customer service resources and customer demand can be realized, and the high adaptation degree customer service is ensured to accept service first, which improves the matching efficiency and quality of the first service. When the customer proposes to change the customer service, the system automatically increases the influence weight of transaction rate on matching degree and reanalyzes and sorts, which can optimize the matching logic according to the actual service feedback, and better meet the customer's personalized demand and real service expectation. This process not only improves the pertinence of customer service reception and customer satisfaction, but also reduces service interruption by continuously optimizing the matching strategy, enhances the customer's trust and dependence on service, and helps enterprises to efficiently dispatch customer service resources, reduce resource waste, improve overall service efficiency and product transaction probability, and promote the formation of a more intelligent and flexible customer service system.

[0057] An AI-driven customer emotion dynamic portrait and intelligent contact system, the system comprising: a customer service data storage module, a customer interest coefficient analysis module, a transaction probability chain analysis module, a comprehensive intention coefficient analysis module, a matching degree analysis module and a customer order management module;

[0058] The customer service data storage module is used to store customer service historical data, analyze the transaction rate and quantity of various products, and calculate the average transaction rate and quantity;

[0059] The customer interest coefficient analysis module is used to collect customer historical data of various products, set an interest transfer period, and analyze the customer's interest coefficient for the product;

[0060] The transaction probability chain analysis module is used to calculate the product transaction probability within the customer cycle and analyze the customer's chain prediction coefficient;

[0061] The comprehensive intention coefficient analysis module is used to analyze the comprehensive intention coefficient by combining the interest coefficient and the chain prediction coefficient, and to select the priority recommended product;

[0062] The matching degree analysis module is used to collect the transaction probability of the customer for the priority recommended product, and analyze the matching degree of the customer service and the customer for the product;

[0063] The customer order management module is used to calculate the average value of the customer matching degree, notify the customer service reception from high to low, until a customer service accepts, and adjust the demand for transaction rate when analyzing the matching degree when the customer reselects the customer service.

[0064] The sorting stability analysis module, the task urgency evaluation module, the running state determination module, the task reassignment module, the adjustment effect monitoring module and the abnormal overrun processing module are connected with the cloud disk through the wireless network, the cloud disk receives and stores the log, when the stability degree of any sorting line is abnormal, an alarm signal is transmitted to the downstream of the sorting line through the blockchain technology, the alarm signal is used to warn that the stability degree of the sorting line is abnormal, and the number of any sorting line stability degree abnormality is updated in the log; through the blockchain technology, the alarm data cannot be maliciously modified after being chained, ensuring that the information received by the downstream system is real and reliable.

[0065] Embodiment 1: After being authorized, the customer service system of a certain e-commerce platform accesses the database of all historical service data of customer service, including the product categories received by each customer service, response speed, transaction conditions, etc. When the system detects that the customer service is in an idle state, it will automatically extract the idle customer service list. Taking home furnishing customer service as an example, the system retrieves the service records of a certain idle customer service in the past half year, analyzes its performance in different product categories: the customer service's transaction rate and transaction volume data in each cycle are sorted out one by one to form the customer service's service capability portrait in various product categories, and the average transaction level of all idle customer services in the same product category is compared to provide a data basis for subsequent matching.

[0066] When the customer consults sofa products, the system synchronously collects the customer's historical purchase and browsing data in the past year, sets an interest transfer period of a quarter. Analysis found that the customer frequently browsed high-end environmentally friendly material sofas in the past two periods, and in the recent period, he generated collection and purchase behaviors on several high-priced sofas. The system calculates the customer's interest coefficient in the current consulted sofa product category according to the customer's actual transaction volume in each period, and quantifies the customer's demand urgency and preference depth by judging whether the customer's recent behavior is concentrated in this category.

[0067] The system further analyzes the correlation between different product categories in the customer's historical consumption: for example, the customer purchased an environmentally friendly sofa and then purchased an environmentally friendly carpet of the same brand in the next period. By analyzing a large amount of similar behavior data, the system concludes the probability of the customer purchasing another product after purchasing a certain product. For the current consultation scenario, the system calculates the probability of the customer possibly purchasing carpets, cushions, and other accessories after browsing sofas, forming a chain prediction coefficient for judging the customer's potential extended demand.

[0068] Combining the customer's interest coefficient in sofa products and the chain prediction coefficient of possibly associated purchases of accessories, the system generates a comprehensive intention coefficient through weighted calculation. The results show that the customer has the highest direct purchase intention for high-end environmentally friendly sofas, and also has a strong potential intention for environmentally friendly accessories. The system accordingly lists sofa products as the preferred recommendation category and accessories as the associated recommendation option, forming a hierarchical recommendation strategy.

[0069] The system screens all idle customer service groups skilled in the product category according to the customer's priority recommendation. For example, the historical data of an idle customer service shows that the transaction rate in the high-end environmental protection sofa category is higher than the average level, and it has handled similar customer's deep consultation demand. The system combines the influence weight of transaction rate and transaction volume on matching degree to calculate the matching degree between the customer service and the current customer, focusing on whether the customer service experience in the same product matches the customer's demand depth, such as whether it can answer professional questions about environmental material certification, long-term maintenance, etc.

[0070] The system sends a reception notice to the customer service from high to low according to the matching degree. If the first customer service does not respond in time or the customer requests to change in the reception (such as the first customer service cannot accurately answer the material question), the system will automatically increase the weight ratio of "transaction rate related factors" and recalculate the matching degree of the remaining customer service. For example, the second customer service performs better in the transaction rate and deep consultation satisfaction of the same product, and the system preferentially recommends it to access. Based on the customer's historical interest data, the customer service actively provides material detection report and customized maintenance scheme, successfully promotes the transaction, and verifies the effect of the dynamic weight adjustment mechanism on the improvement of service accuracy.

[0071] It is obvious to those skilled in the art that the present application is not limited to the details of the above exemplary embodiments, and can be implemented in other specific forms without departing from the spirit or essential characteristics of the present application. Therefore, from any point of view, the embodiments should be regarded as exemplary and non-limiting, and the scope of the present application is defined by the appended claims rather than the above description, and therefore all changes falling within the meaning and scope of the equivalent elements of the claims are intended to be included in the present application. Any reference signs in the claims should not be regarded as limiting the claims involved.

Claims

1. A method for dynamic customer profiling and intelligent outreach based on AI, characterized in that: The method includes the following steps: S1. Store historical customer service data, analyze the conversion rate and quantity of various products, and calculate the average conversion rate and quantity. S2. Collect historical data on various customer products, set interest transfer cycles, and analyze customer interest coefficients for products; S3. Statistically analyze the probability of product sales within a customer's cycle and analyze the customer's chain prediction coefficient. S4. Combine the interest coefficient and the chain prediction coefficient to analyze the comprehensive intention coefficient and select the priority recommended products; S5. Collect data on the probability of customers making a purchase of the recommended products and analyze the matching degree between customer service and customers for the products. S6. Calculate the average customer matching score and notify customer service to handle the customer from high to low until a customer service representative accepts the request. When the customer chooses a customer service representative again, adjust the conversion rate requirement when analyzing the matching score. In step S3, the probability F is calculated that if a customer makes a purchase of product type α in one interest shift cycle, the customer will make a purchase of product type y in the next interest shift cycle. α_y Substituting α = 1, 2, ..., Y, we statistically obtain the probability that when a customer makes a purchase of product type Y in one interest shift cycle, they will make a purchase of product type y in the next interest shift cycle: {F} 1_y ,F 2_y ,…,F α_y ,…,F Y_y }, statistically analyze customer transactions related to product type Y during the T-th interest shift cycle, and calculate the chain prediction coefficient W for customer behavior towards product type y. y : ； Where f α_y To calculate the prediction coefficient for customer transactions of product category α in the next interest shift cycle, based on customer transactions of product category α in one interest shift cycle, the prediction coefficient f for customer transactions of product category y in the next interest shift cycle is calculated when a customer makes a transaction of product category α in one interest shift cycle. α_y =F α_y Otherwise f α_y =0.

2. The AI-driven dynamic customer profile and intelligent outreach method according to claim 1, characterized in that: In step S1, after authorization, the historical data of customer service personnel is stored in the database, and the status of customer service personnel is queried. Idle customer service personnel include {A1, A2, ..., A...} x ,…,A X }, where X represents the number of available customer service personnel, A x This indicates that the xth available customer service representative is customer service representative A. x The analysis is performed, using the current time point as the end point, querying the historical data of customer service personnel for T monitoring time periods. The monitoring time period ending at the current time point is defined as the Tth monitoring time period, where T is the predetermined number of monitoring time periods. The analysis is performed on the tth monitoring time period, where t = 1, 2, ..., T, and there are Y product categories. In the tth monitoring time period, customer service personnel A... x The transaction rate for product category y is C x_y The number of transactions for product category y is D. x_y Customer Service Representative A x The average transaction rate for product type Y is C x The average number of transactions for product type Y is D. x Then, substituting x=1,2,…,X one by one, we obtain the conversion rate of X customer service personnel for the y-th type of product {C}. 1_y C 2_y ,…,C x_y ,…,C X_y }, the number of transactions for product category y by X customer service personnel is {D}. 1_y D 2_y ,…,D x_y ,…,D X_y The average conversion rate of X customer service representatives for product type Y is {C1, C2, ..., C}. x ,…,C X The average number of transactions for product type Y by X customer service representatives is {D1, D2, ..., D}. x ,…,D X } 3. The AI-driven dynamic customer profiling and intelligent outreach method according to claim 2, characterized in that: In step S2, after authorization, historical data on the nth customer's interest in product type Y is collected. An interest transfer cycle is set, with the current time as the endpoint, designated as the i-th interest transfer cycle, where I is the number of interest transfer cycles. The i-th interest transfer cycle is analyzed, where i = 1, 2, ..., I. In the i-th interest transfer cycle, the customer's transaction volume for product type Y is {E1, E2, ..., E...}. y ,…,E α ,…,E Y }, where E y E represents the customer's transaction volume for product category y. α This represents the transaction volume of customers for product category α, and further, it gives the interest coefficient w for product category y when a customer applies for customer service. y : 。 4. The AI-driven dynamic customer profile and intelligent outreach method according to claim 3, characterized in that: In step S4, based on the customer's interest coefficient w for product category y y And the chain prediction coefficient W for product class y y Calculate the overall intention coefficient M of customers for product type y. y The comprehensive intention coefficient M y By analyzing the interest coefficient w y and comprehensive intention coefficient M y We obtain the weighted average by substituting each value into y=1,2,…,Y to get the overall customer intention coefficient {M1,M2,…,M...} for product type Y. y ,…,M Y The product with the highest overall intention coefficient (β-class) is selected as the preferred product to recommend to the customer.

5. The AI-driven dynamic customer profiling and intelligent outreach method according to claim 4, characterized in that: In step S5, when a customer requests customer service, after authorization, it is detected that the customer's β-class preferred recommended products include the y-th product, and the probability H of the nth customer's purchase of the y-th product is collected. n_y Substitute each value into the sequence n=1,2,…,N to obtain the probability {H} of N customers purchasing the y-th product. 1_y H 2_y ,…,H n_y ,…,H N_y }, where the maximum transaction rate is H max_y Then, calculate the matching degree J between the x-th customer service representative and the n-th customer regarding product y. x_y : ； Where k1 is the weight of the impact of the conversion rate on the matching degree, k1= k2 represents the weight of the impact of trading volume on the matching degree, k2= .

6. The AI-driven dynamic customer profiling and intelligent outreach method according to claim 5, characterized in that: In step S6, the average matching degree of X customer service representatives with the nth customer regarding the β-type product is calculated, and customer service representatives are notified to receive the customer in descending order of the average matching degree, until a customer service representative accepts the customer service notification.

7. The AI-driven dynamic customer profile and intelligent outreach method according to claim 6, characterized in that: If a customer requests a change of customer service representative during the customer service process, the weight of the conversion rate on the matching degree will be increased. The average matching degree of X customer service representatives for the nth customer regarding the β-type product will be re-analyzed. Customer service representatives will be notified to receive the customer in descending order of the average matching degree until a customer service representative accepts the customer service notification.

8. An AI-driven dynamic customer relationship profiling and intelligent outreach system, wherein the system is applied to the AI-driven dynamic customer relationship profiling and intelligent outreach method described in any one of claims 1-7, characterized in that: The system includes: a customer service data storage module, a customer interest coefficient analysis module, a transaction probability chain analysis module, a comprehensive intention coefficient analysis module, a matching degree analysis module, and a customer service order management module. The customer service data storage module is used to store historical customer service data, analyze the conversion rate and quantity of various products, and calculate the average conversion rate and quantity. The customer interest coefficient analysis module is used to collect historical data of various products from customers, set interest transfer cycles, and analyze customers' interest coefficients on products. The transaction probability chain analysis module is used to statistically analyze the probability of product transactions within a customer's cycle and to analyze the customer's chain prediction coefficient. The comprehensive intention coefficient analysis module is used to combine interest coefficient and chain prediction coefficient to analyze the comprehensive intention coefficient and select priority recommended products; The matching degree analysis module is used to collect the probability of customers making a purchase of the preferred recommended products and to analyze the matching degree between customer service and customers for the products. The customer service scheduling management module is used to calculate the average customer matching score and notify customer service representatives to handle the task from high to low until a customer service representative accepts the task. When a customer selects a customer service representative again, the conversion rate requirement is adjusted based on the matching score analysis.

Images