Intelligent outbound method and device based on prediction information, equipment and medium

Abstract

The invention relates to the technical field of intelligent outbound calling, and discloses an intelligent outbound calling method and device based on prediction information, equipment and a medium, and the method comprises the steps: obtaining a historical data set which comprises historical behavior information corresponding to a plurality of clients; based on the historical behavior information corresponding to each customer, behavior prediction information of each customer is obtained through a preset machine learning model; the outbound rule of each customer is determined according to the behavior prediction information, and outbound is carried out on the customer based on the outbound rule, the historical data of the customer is analyzed, so that the outbound rule of each customer is determined in a targeted mode, the outbound quality of each time is guaranteed, and the outbound efficiency and the outbound effect are greatly improved.

Description

Technical Field

[0001] The present invention relates to the field of intelligent outbound calling technology, and in particular to an intelligent outbound calling method, device, computer equipment and storage medium based on prediction information. Background Art

[0002] In the existing customer service system, intelligent outbound call systems are widely used to send reminders to customers, promote information or conduct market research. However, the existing intelligent outbound call systems have some shortcomings in the call process, such as lack of personalized customization, low efficiency, poor user experience, etc. Enterprises usually use manual dialing to make outbound calls, which is inefficient and prone to manual misjudgment.

[0003] In recent years, intelligent outbound call systems have gradually emerged. The system can automatically dial, determine the customer's connection status, and record the outbound call results. Intelligent outbound call systems improve outbound call efficiency, but current intelligent outbound call systems often use fixed rules for outbound calls and cannot dynamically adjust outbound call strategies based on customer data, resulting in low outbound call efficiency and poor outbound call results. Summary of the invention

[0004] In view of this, the present invention provides an intelligent outbound calling method, device, computer equipment and storage medium based on prediction information to solve the problem that the outbound calling efficiency of the current intelligent outbound calling system is low and the outbound calling effect is poor.

[0005] In a first aspect, the present invention provides an intelligent outbound calling method based on prediction information, the method comprising:

[0006] Acquire a historical data set, the historical data set comprising: historical behavior information corresponding to a plurality of customers;

[0007] Based on the historical behavior information of each customer, the behavior prediction information of each customer is obtained through the preset machine learning model;

[0008] An outbound calling rule for each customer is determined according to the behavior prediction information, and an outbound calling is made to the customer based on the outbound calling rule.

[0009] This method uses a preset machine learning model to analyze and predict the historical behavior information of each customer, obtains the predicted behavior information of each customer in the future, and formulates targeted outbound calling rules based on the predicted behavior information of each customer to make outbound calls to the customer, thereby ensuring that the outbound call content matches the customer as much as possible, which can greatly improve the outbound call efficiency and outbound call effect and ensure the outbound call success rate.

[0010] In an optional implementation, the method of obtaining behavior prediction information of each customer based on the historical behavior information corresponding to each customer through a preset machine learning model includes:

[0011] Extracting the customer's behavior characteristics based on the historical behavior information, the behavior characteristics including: purchased products, purchase frequency, purchase amount and purchase time, and outbound call answering time;

[0012] The behavioral characteristics are processed through a preset machine learning model to determine the optimal outbound calling time period for each customer, as well as the predicted interest and focus angle for each type of product.

[0013] This embodiment extracts the customer's historical purchasing and answering behavior characteristics from the historical behavior information, and uses a preset machine learning model to predict the customer's optimal outbound calling time period and interest and focus on various products based on these characteristics, so that more targeted outbound calling rules can be set later to ensure the outbound calling success rate.

[0014] In an optional implementation, determining the outbound calling rules for each customer according to the behavior prediction information includes:

[0015] Get the recommended products corresponding to the current outbound call task;

[0016] Determine the target interest and target attention angle of each customer for the recommended product based on the predicted interest and predicted attention angle of each customer for each product;

[0017] Determine the outbound calling frequency of each customer based on the target interest level, and determine the outbound calling content of each customer based on the target focus angle;

[0018] The outbound calling rules for each customer are determined according to the outbound calling frequency, outbound calling content and optimal outbound calling time period.

[0019] This embodiment combines the product that currently needs to be recommended by outbound calls with the interest and focus of each customer in the product to determine the specific outbound call frequency and content, and sets targeted outbound call rules based on the outbound call time period to improve the outbound call effect.

[0020] In an optional embodiment, the method further includes:

[0021] Obtain key indicators during outbound calls, including call success rate, single call duration, and customer feedback content;

[0022] The outbound call frequency, outbound call content and optimal outbound call time period are updated based on the key indicators.

[0023] In this implementation mode, by acquiring key indicators in the outbound calling process to update the outbound calling rules, it is possible to further ensure that the outbound calling rules match the actual outbound calling situation and improve the outbound calling effect.

[0024] In an optional implementation manner, updating the outbound call frequency, outbound call content, and optimal outbound call time period based on the key indicator includes:

[0025] When the call success rate is lower than a preset value, reducing the outbound call frequency;

[0026] Count the call success rates of each time period, and determine the time period with the highest call success rate as the updated optimal outbound call time period;

[0027] The actual interest of the customer in the recommended product is determined according to the monitored customer feedback content, and the call content is updated based on the actual interest.

[0028] This implementation method reduces the outbound call frequency when the outbound call success rate is too low to avoid invalid outbound calls. At the same time, the outbound call time period is updated according to the outbound call success rate in each time period to ensure that outbound calls are made during time periods with higher customer answering rates, thereby improving call success rates. At the same time, the outbound call content is updated according to the customer's actual interest in the recommended products at the time of the outbound call, thereby ensuring as much as possible that the content of subsequent outbound calls is of interest to the customer, thereby improving the outbound call effect.

[0029] In an optional implementation, when making an outbound call to a customer based on the outbound call rule, the method further includes:

[0030] Outbound calls are made through encrypted communications and / or identity authentication and / or access control.

[0031] This implementation method can ensure the privacy of customers during outbound calls and avoid information leakage by taking measures such as encrypted communication, identity authentication and access control.

[0032] In an optional implementation, the preset machine learning model is: a machine learning model constructed based on a random forest algorithm;

[0033] The method of obtaining behavior prediction information of each customer based on the historical behavior information corresponding to each customer by using a preset machine learning model includes:

[0034] The historical behavior information of each customer is analyzed by the machine learning model constructed based on the random forest algorithm to obtain the behavior prediction information of each customer.

[0035] In this implementation, customer behavior is analyzed by a machine learning model constructed based on a random forest algorithm, so that behavior prediction information of each customer can be obtained more accurately.

[0036] In a second aspect, the present invention provides an intelligent outbound calling device based on prediction information, the device comprising:

[0037] A historical data acquisition module is used to acquire a historical data set, wherein the historical data set includes: historical behavior information corresponding to a plurality of customers;

[0038] A customer behavior prediction module is used to obtain behavior prediction information of each customer through a preset machine learning model based on the historical behavior information corresponding to each customer;

[0039] The outbound call rule determination module is used to determine the outbound call rule of each customer according to the behavior prediction information, and make outbound calls to the customer based on the outbound call rule.

[0040] In a third aspect, the present invention provides a computer device, comprising: a memory and a processor, the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the intelligent outbound calling method based on predictive information of the above-mentioned first aspect or any corresponding embodiment thereof by executing the computer instructions.

[0041] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the intelligent outbound calling method based on predictive information of the above-mentioned first aspect or any corresponding embodiment thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0043] Figure 1 is a flow chart of an intelligent outbound calling method based on prediction information according to an embodiment of the present invention;

[0044] Figure 2 is a flow chart of another intelligent outbound calling method based on prediction information according to an embodiment of the present invention;

[0045] Figure 3 This is an example diagram of an outbound call process according to an embodiment of the present invention;

[0046] Figure 4 is a structural block diagram of an intelligent outbound calling device based on prediction information according to an embodiment of the present invention;

[0047] Figure 5 It is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0049] In the existing customer service system, intelligent outbound call systems are widely used to send reminders to customers, promote information or conduct market research. However, the existing intelligent outbound call systems have some shortcomings in the call process, such as lack of personalized customization, low efficiency, poor user experience, etc. Enterprises usually use manual dialing to make outbound calls, which is inefficient and prone to manual misjudgment.

[0050] In recent years, intelligent outbound call systems have gradually emerged. The system can automatically dial, determine the customer's connection status, and record the outbound call results. Intelligent outbound call systems improve outbound call efficiency, but current intelligent outbound call systems often use fixed rules for outbound calls and cannot dynamically adjust outbound call strategies based on customer data, resulting in low outbound call efficiency and poor outbound call results.

[0051] To this end, an embodiment of the present invention provides an intelligent outbound calling method based on predictive information, which extracts the customer's historical purchasing and answering behavior characteristics from the historical behavior information, and uses a preset machine learning model to predict the customer's optimal outbound calling time period and interest and focus on various products based on these characteristics, so that more targeted outbound calling rules can be set later to ensure the success rate of outbound calling.

[0052] According to an embodiment of the present invention, an embodiment of an intelligent outbound calling method based on predictive information is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0053] In this embodiment, a smart outbound calling method based on prediction information is provided, which can be used in the above-mentioned smart outbound calling system. Figure 1 is a flow chart of an intelligent outbound calling method based on prediction information according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0054] Step S101, obtaining a historical data set, where the historical data set includes: historical behavior information corresponding to a plurality of customers.

[0055] Outbound call systems are usually used in related industries such as banks, insurance companies, and telecommunications operators. They can use the outbound call system to perform related outbound call tasks, such as information promotion, product recommendations, market research, etc.

[0056] For the above-mentioned outbound call initiators, namely banks, insurance companies, communication operators, etc., the customer's historical behavior information is recorded in their internal relevant databases and can be directly obtained from the database. Taking banks as an example, the internal database will record the historical behavior information of each customer, such as financial product purchase records, service consultation records, historical outbound call response status and other information. These historical data can reflect the corresponding outbound call objects, that is, what types of products or services the customer was interested in in the past, and the historical outbound call answering situation. By analogy, the historical behavior information for other industries is also the type of information that can reflect the customer's historical situation. Here, only the banking industry is used as an example, and other industries will not be repeated.

[0057] Step S102, based on the historical behavior information corresponding to each customer, the behavior prediction information of each customer is obtained through a preset machine learning model.

[0058] After obtaining the historical data set, which contains a large amount of historical data corresponding to customers, the historical data of each customer, that is, historical behavior information, can be analyzed and processed through the trained machine learning model to predict the customer's behavior information in the future. The preset machine learning model can be a trained algorithm model or a related behavior analysis tool, which is not limited here.

[0059] After processing the relevant historical data corresponding to each customer, the predicted behavior information may include: which products or services the customer may be interested in in the next period of time, which angles he or she will pay more attention to, during which time period the call answering rate is higher, and other related behavioral information, so as to subsequently specify personalized outbound call rules corresponding to each customer based on this information.

[0060] Step S103, determining the outbound calling rules for each customer according to the behavior prediction information, and making outbound calls to the customer based on the outbound calling rules.

[0061] Customer behavior prediction information often includes the customer's current interest in various types of products, as well as information about which aspects the customer is more concerned about. Based on this information, targeted outbound calling rules can be specified for each customer according to preset rules.

[0062] Take the banking industry as an example. There are many types of financial products in the banking industry. Behavior prediction information can include customer interest in various types of financial products, as well as information about the aspects that customers focus on when purchasing financial products, such as yield, stability, etc. After determining these contents, we can recommend the most interesting financial products to different customers through outbound calls, and mainly introduce the contents that customers are interested in. At the same time, behavior prediction information can also include the time periods when customers have a high answering rate, so as to determine the outbound call time.

[0063] In a specific example, customer A is more interested in funds and is mainly concerned about the rate of return. His answer rate is higher in the evening every day. Therefore, the corresponding outbound call rule can be to recommend fund products to the customer in the evening at preset intervals and mainly introduce the rate of return content through the intelligent outbound call system.

[0064] The intelligent outbound calling method based on predictive information provided in this embodiment extracts the customer's historical purchasing and answering behavior characteristics from the historical behavior information, and uses a preset machine learning model to predict the customer's optimal outbound calling time period and interest and focus on various products based on these characteristics, so that more targeted outbound calling rules can be set later to ensure the outbound calling success rate.

[0065] According to an embodiment of the present invention, another embodiment of an intelligent outbound calling method based on prediction information is provided, which can be used in the above-mentioned intelligent outbound calling system. Figure 2 is a flow chart of another intelligent outbound calling method based on prediction information according to an embodiment of the present invention. Figure 2 As shown, the process includes the following steps:

[0066] Step S201, obtain a historical data set, the historical data set includes: historical behavior information corresponding to multiple customers. Figure 1 Step S101 of the illustrated embodiment will not be described in detail here.

[0067] Step S202, based on the historical behavior information corresponding to each customer, obtain the behavior prediction information of each customer through a preset machine learning model.

[0068] Specifically, the preset machine learning model is: a machine learning model built based on a random forest algorithm;

[0069] In the above step S202, based on the historical behavior information corresponding to each customer, the behavior prediction information of each customer is obtained through a preset machine learning model, including:

[0070] The historical behavior information of each customer is analyzed through a machine learning model built based on the random forest algorithm to obtain the behavior prediction information of each customer.

[0071] The random forest algorithm is an integrated learning algorithm that performs training and prediction based on multiple decision tree models. It is particularly suitable for processing complex nonlinear relationships and can effectively deal with high-dimensional data and unbalanced data problems. In the field of marketing, it is often used to determine customer preferences, so it can be used to analyze customer data to determine customer behavior prediction information. The machine learning model built based on this algorithm can analyze and process key data in historical behavior information, such as product purchase time, information consultation records, outbound call answering records, etc., to obtain customer interest in a certain product or service, that is, to obtain behavior prediction information corresponding to each customer.

[0072] For example, for a machine learning model built based on the random forest algorithm, the prediction process can be as follows:

[0073] 1. Data preparation: Collect and organize customer data, including historical transaction records, behavior patterns, preferences and other characteristic variables. Divide the data into training sets and test sets.

[0074] 2. Feature Engineering: Extract and process the features of the original data. For example, extract features such as the customer's purchase frequency, purchase amount, and purchase time. Perform one-hot encoding on categorical features and standardize numerical features.

[0075] 3. Model training: Use the training set data to train the random forest model. Random forest generates model clusters with different biases by building multiple decision trees and randomly selecting data subsets and feature subsets on each tree.

[0076] 4. Model prediction: Apply the trained random forest model on the test set to predict the customer’s future behavior. For example, predict whether a customer will be interested in a certain product, or predict the customer’s possible churn risk.

[0077] 5. Model evaluation: Use accuracy, precision, recall, F1 score and other indicators to evaluate the performance of the model. Cross-validation is used to further verify the robustness of the model.

[0078] 6. Model application: Apply the prediction results to actual business scenarios, such as recommending products that customers may be interested in through a recommendation system, or contacting customers at the best time.

[0079] Specifically, step S202 includes:

[0080] Step S202-1, extracting the customer's behavior characteristics based on historical behavior information, the behavior characteristics include: purchased products, purchase frequency, purchase amount and purchase time, and outbound call answering time.

[0081] It can be understood that each customer's historical data includes specific purchase records of various products, consultation records, outbound call answering time, answering duration, and other information that can be used to analyze customer behavior. Among these historical behavior information, purchase information can best represent the customer's interest. Therefore, based on the above information, it is possible to clarify the customer's specific purchase products, purchase frequency, purchase amount, etc. in history, and normalize this information to obtain the corresponding behavior characteristics. In addition, the historical data can also include the customer's historical outbound call answering time, that is, the time when the customer answers the outbound call and the time of each outbound call communication, and the corresponding normalization processing is also performed to obtain the corresponding behavior characteristics.

[0082] Step S202-2, the behavioral characteristics are processed by a preset machine learning model to determine the best outbound calling time period for each customer and the predicted interest and attention angle for each type of product.

[0083] After obtaining the above-extracted product purchase features and outbound call communication features, namely, purchased products, purchase frequency, purchase amount, and outbound call answering time. These feature vectors can be analyzed through a preset machine learning model to clarify the customer's preferences for different financial products, such as the customer's interest in different types of financial products, that is, predicting the interest, and the main focus when purchasing products, such as interest rates, risks, etc., that is, predicting the focus angle. In addition, for the outbound call answering time, specifically including when the customer answers the outbound call, and the duration of each outbound call, by analyzing the communication features, it can be predicted in which time period the customer is more willing to answer the outbound call, that is, the best outbound call time period.

[0084] Furthermore, in order to more accurately determine the customer's behavior information, the features extracted in the above steps may further include the customer's product inquiry information, the attention time of different products in the relevant app, and other information to further accurately analyze the customer's preferences.

[0085] Step S203, determining the outbound calling rules for each customer according to the behavior prediction information, and making outbound calls to the customer based on the outbound calling rules.

[0086] Specifically, step S203 includes:

[0087] Step S203-1, obtaining the recommended product corresponding to the current outbound call task.

[0088] In some actual application scenarios, when a bank launches a new product, it needs to make outbound sales calls to customers. Therefore, it is necessary to combine the specific information of the current recommended product with the predicted behavior information of each customer to specify targeted outbound call rules. Therefore, it is necessary to obtain the product that currently needs outbound sales, that is, the recommended product corresponding to the current outbound call task.

[0089] Step S203-2, determining the target interest and target attention angle of each customer for the recommended product based on the predicted interest and predicted attention angle of each customer for each product.

[0090] For each customer, the predicted interest and predicted focus of the customer for different types of products have been obtained in the above steps. Therefore, it is necessary to determine the customer's target interest and target focus for the product that needs to be recommended for the current outbound call task on this basis.

[0091] Step S203-3, determining the outbound call frequency of each customer based on the target interest level, and determining the outbound call content of each customer based on the target focus angle.

[0092] The higher the customer's target interest in the recommended product, the more frequently the product can be recommended to the customer. Therefore, different outbound call frequencies can be matched to different interest intervals according to preset rules.

[0093] At the same time, the target focus angle indicates which aspects the customer is more concerned about, such as interest rates, risks, return cycles, etc. The intelligent outbound call system can generate corresponding outbound call content based on the customer's target focus angle. For example, if the customer's target focus angle is interest rates, then when recommending the above products to the customer through outbound calls, the content related to interest rates will be mainly introduced.

[0094] Step S203-4, determining the outbound calling rules for each customer according to the outbound calling frequency, outbound calling content and the best outbound calling time period.

[0095] After obtaining the outbound call frequency, outbound call content, and optimal outbound call time period corresponding to the customer in the above steps, the corresponding outbound call rules can be formulated. For example, for the recommended product a and customer a, the outbound call frequency obtained in the above steps is once a week; the outbound call content is about interest rates; and the optimal time period is Friday night. Therefore, the corresponding outbound call rule is to initiate an outbound call communication for the recommended product a to customer a every Friday night, mainly introducing the interest rate content of product a, with a frequency of once a week.

[0096] Furthermore, in the above step S203, when making an outbound call to the customer based on the outbound call rule, the method further includes:

[0097] Outbound calls are made through encrypted communications and / or identity authentication and / or access control.

[0098] When making outbound calls, measures such as encrypted communication, identity authentication, and access control can be adopted to ensure the privacy of customers during the outbound call process and avoid information leakage.

[0099] Step S204, obtaining key indicators in the outbound calling process, the key indicators include: call success rate, single call duration, customer feedback content; updating the outbound calling frequency, outbound calling content and optimal outbound calling time period based on the key indicators.

[0100] It can be understood that when the outbound call rules are determined above, the outbound call rules are formulated based on the relevant information obtained by prediction based on historical data. In actual outbound calls, the outbound call rules may not be perfectly adapted to the corresponding users. Therefore, the key indicators in the answering process can be obtained to update the outbound call rules to further improve the outbound call effect. Specific key indicators can be: call success rate, that is, the user's connection status and recommendation success status within the preset time period; the call duration corresponding to each outbound call communication connected; the content of customer feedback during the outbound call process, such as other information consulted by the customer, etc. After obtaining these key indicators, the outbound call rules for the customer can be adjusted according to the actual outbound call situation.

[0101] Specifically, in step S204, the outbound call frequency, outbound call content and optimal outbound call time period are updated based on the key indicators, including:

[0102] Step S204-1, when the call success rate is lower than a preset value, reduce the outbound call frequency.

[0103] When the call success rate is too low, it means that the customer is not interested in the recommended product. At this time, you can reduce the frequency of outbound calls to the customer to avoid invalid outbound calls and avoid repeated disturbances to the customer.

[0104] Step S204-2, counting the call success rates of each time period, and determining the time period with the highest call success rate as the updated optimal outbound call time period.

[0105] The call success rate of each time period is counted to update the best outbound call time period. On this basis, the best outbound call time period can be further selected in combination with the communication duration. Because the longer the communication duration, the more free time the customer has in the corresponding time period to learn relevant information, so the best outbound call time period can be selected in combination with the call success rate and the communication duration. Specifically, an outbound call with a communication duration greater than the preset duration can be defined as a valid call success, and the best outbound call time period can be updated based on the valid call success rate of each time period.

[0106] Step S204-3: determining the actual interest of the customer in the recommended product according to the monitored customer feedback content, and updating the call content based on the actual interest.

[0107] When making outbound calls to customers, they may ask for specific information about certain aspects of the recommended products. Based on this, in subsequent outbound calls, recommended products can be introduced based on this specific information, thereby increasing customers' interest in the outbound calls and further ensuring the effectiveness of the outbound calls.

[0108] The intelligent outbound calling method based on prediction information provided by the embodiment of the present invention extracts the customer's historical behavior characteristics in terms of purchase and answering calls from the historical behavior information, and uses a preset machine learning model to predict the customer's best outbound calling time period and interest and attention angle for various products based on these characteristics, and combines the product that currently needs to be recommended for outbound calls with the interest and attention angle of each customer for the product, thereby determining the specific outbound calling frequency and outbound calling content, and setting targeted outbound calling rules in combination with the outbound calling time period, thereby improving the outbound calling effect. On this basis, key indicators in the outbound calling process are also obtained to update the outbound calling rules, which can further ensure that the outbound calling rules match the actual outbound calling situation and improve the outbound calling effect.

[0109] For example, Figure 3 As shown, it is an example diagram of an outbound call process according to an embodiment of the present invention. First, before performing an outbound call task, a predictive outbound call task corresponding to each number needs to be created through a lead number pool, where each number corresponds to a customer who needs to be recommended a product or service. The specific outbound call rule in each outbound call task is the outbound call rule determined according to the above method embodiment, and predictive outbound calls are performed based on the outbound call task.

[0110] When making an outbound call, if the call is connected, the intelligent outbound call system can make an intelligent introduction to the customer based on the relevant outbound call content in the outbound call task. If the customer is interested in the relevant content and wants to know more, he or she can be transferred to a human operator for further introduction.

[0111] If the call cannot be connected, it can detect whether the corresponding number is empty, and automatically match the re-call strategy if it is not empty, such as calling again after 30 minutes, or changing the outbound call number, etc. Then a new round of calls is made based on the re-call strategy.

[0112] The formulation of personalized call rules in the above method embodiment is mainly based on the prediction results of customer data, and the system formulates personalized call rules. These rules can include personalized settings in terms of call frequency, call time period, call content, etc. to ensure that the call content and time are attractive and effective to customers. In the formulation of personalized call rules, the association between the prediction results and the call rules is mainly reflected in how to customize the call strategy based on the customer behavior prediction. The following is an illustrative example to explain this process.

[0113] Suppose a telecommunications company wants to contact customers by phone to promote a new package service. They use machine learning models (such as the random forest mentioned above) to predict customer behavior so that they can reach customers more effectively.

[0114] First, we analyze and predict customer data. By analyzing the customer's historical data, the company uses the model to predict each customer's interest in the new package. For example:

[0115] Customer A: There is an 80% probability of being interested in the new package.

[0116] Customer B: There is a 30% probability of being interested in the new package.

[0117] Customer C: There is a 60% probability that he is interested in the new package.

[0118] Personalized call rules are then formulated. Based on these prediction results, the system can formulate different call strategies for different customers.

[0119] For Customer A:

[0120] Call frequency: Because customer A is very interested in the new package (high predicted probability), the call frequency can be increased, such as once a week.

[0121] Call time period: Based on customer A's historical behavior data, the system analyzes and concludes that customer A is most likely to answer the phone between 6pm and 8pm, so the call time is set during this period.

[0122] Content of the call: Focus on introducing the preferential strength of the new package and emphasize its advantages compared with customer A's current package.

[0123] Customer B:

[0124] Call frequency: Customer B has low interest in the new package (low predicted probability), so calls are reduced to perhaps once a month.

[0125] Call time period: If customer B rarely answers calls during working hours (such as 9:00 a.m. to 11:00 a.m.), the call time can be scheduled during a time period when the customer may be free, such as noon or afternoon.

[0126] Content of the call: Focus on understanding customer B’s needs. You may first ask about customer B’s satisfaction with existing services, and then adjust marketing strategies based on the feedback.

[0127] Customer C:

[0128] Call frequency: Customer C has some interest in the new package (medium prediction probability) and can choose a call frequency of once every two weeks.

[0129] Call time period: If customer C's historical data shows that he is more likely to answer calls on weekends, the call can be scheduled during the weekend midday time period.

[0130] Content of call: Based on customer C's historical usage habits, recommend the parts of the new package that are particularly suitable for his needs, perhaps mentioning additional data or international calling discounts.

[0131] When implementing outbound call rules, design user-friendly call content and interaction methods to ensure that the interruption to users during the call process is minimized. The call content should be personalized according to the interests and needs of the customer, and the call interaction method should be simple and intuitive, easy for users to understand and operate.

[0132] Finally, feedback adjustment is performed during the implementation process: After the system implements these personalized call rules, it will also monitor customer feedback, such as whether the customer answers the call, whether they are interested in learning more about the product, etc. Based on the feedback, the system can further optimize the call rules, such as adjusting the call frequency or time, or modifying the call content.

[0133] Furthermore, when making outbound calls, a call scheduling algorithm can be introduced to dynamically adjust the call strategy based on the real-time update of customer data and system load. This algorithm can realize intelligent scheduling of call tasks, optimize resource utilization and improve call efficiency based on factors such as customer priority and call history.

[0134] In the above method embodiment, the prediction results and the call rules are closely related. The prediction results provide data support for the customer's potential behavior, and the call rules are tailored according to these prediction results to ensure that the call is more efficient and attractive. Such a personalized call strategy helps to improve customer engagement and satisfaction, and also improves the success rate of marketing activities.

[0135] On the basis of the above solution, a real-time monitoring and feedback mechanism is further introduced to monitor and analyze key indicators in the call process in real time, including call success rate, customer feedback, etc. Call strategies are adjusted in a timely manner according to the monitoring results to ensure the stability of system operation and continuous optimization of call effects.

[0136] In the real-time monitoring and feedback mechanism, call strategy refers to the various strategies and decisions adopted by the company in the process of contacting customers. These strategies usually include the frequency, time, content, and contact method of calls. The introduction of real-time monitoring and feedback mechanism can help companies dynamically adjust these strategies during operation to achieve better results.

[0137] The specific contents of the calling strategy may include:

[0138] Call frequency: refers to the frequency of contacting customers. Too frequent may annoy customers, too infrequent may miss marketing opportunities.

[0139] Call time: Choose when to make calls or send messages. For example, during the weekdays, evenings, or weekends.

[0140] Call content: includes information delivered when communicating with customers, such as promotions, service recommendations, etc. Different customers may require different marketing messages.

[0141] Contact method: It can be phone, text message, email or interaction on social media. Choose the appropriate method to communicate with customers.

[0142] The key indicators for real-time monitoring are as follows:

[0143] Call success rate: the ratio of the number of successful calls to the total number of calls.

[0144] Customer feedback: Customer feedback during or after the call, such as whether they are interested, satisfied with the service or product, and willing to learn more.

[0145] Customer answering rate: the proportion of calls that are actually answered by customers.

[0146] Call length: The duration of the call, which can indirectly reflect the customer's interest level.

[0147] Customer behavior: For example, whether they visited the company website after answering the phone, whether they placed an order, etc.

[0148] Exemplarily, the call strategy adjustment based on monitoring information may be specifically as follows:

[0149] 1. Adjust the call frequency:

[0150] Situation: If the system monitors that the customer's answering rate is low, it may be because the call frequency is too high and the customer is unwilling to answer the call.

[0151] Adjustment: The system can automatically reduce the customer's call frequency and extend the interval to the next contact.

[0152] 2. Optimize call time period:

[0153] Situation: Monitoring finds that the probability of a customer answering the phone is lower during a specific time period, or the answering rate is significantly higher during certain time periods.

[0154] Adjustment: The system adjusts the call time to a time period when the customer is more likely to answer. For example, if it is found that the customer prefers to answer the phone on weekend afternoons, the system will schedule the follow-up call during this time period.

[0155] 3. Modify the call content:

[0156] Situation: Through customer feedback or behavior monitoring, it is found that customers are not interested in current promotional content or recommended products.

[0157] Adjust: Modify the call content based on the customer's historical preferences and current feedback, such as recommending other products or services or changing the marketing language.

[0158] 4. Change of contact method:

[0159] Situation: If the customer expresses during the call that he or she is uncomfortable or does not like telephone communication, or the answering rate remains low.

[0160] Adjustment: The system may choose to contact you via other means such as text message or email.

[0161] 5. Dynamically adjust strategies:

[0162] Situation: The overall call success rate has decreased or customer satisfaction has decreased, which may be due to changes in the market environment or the strategy is no longer applicable.

[0163] Adjustment: The system can automatically conduct A / B testing based on real-time data, try different calling strategies and adjust them in real time.

[0164] Specific example, initial call strategy: The system sets an initial call strategy based on the customer prediction model and historical data, such as calling customers at 7 pm every Wednesday to recommend new packages.

[0165] Real-time monitoring: After the first call, the system monitors the customer's response and feedback in real time.

[0166] If the customer does not answer the call, the system records this information.

[0167] If the customer answers but expresses disinterest, the system will also record this feedback.

[0168] Adjust strategies based on real-time monitoring results:

[0169] No answer call: The system automatically schedules another attempt to contact you on Saturday afternoon.

[0170] Feedback of disinterest: The system adjusts the content of the next call and may recommend products that are more in line with the customer's interests.

[0171] Continuous optimization: As multiple calls accumulate, the system continuously adjusts strategies to ensure that each call is more likely to attract customer interest and response.

[0172] Through real-time monitoring and feedback mechanisms, companies can dynamically adjust call strategies to ensure that the call process is more targeted and effective. The system can not only respond quickly after detecting customer behavior and feedback, but also improve the overall marketing effect through continuous strategy optimization.

[0173] Based on the above method implementation example, it can be expanded and customized: the design system has good scalability and customization capabilities, and can be customized and developed according to different customer needs and business scenarios to meet the needs of different industries and enterprises of different sizes.

[0174] Through the comprehensive application of the above technical solutions, the present invention can effectively solve the problems existing in the existing intelligent outbound call system, provide more intelligent, efficient and user-friendly outbound call services, and bring new solutions to the fields of customer service and marketing promotion.

[0175] The present invention uses machine learning algorithms to analyze and predict customer data and formulate personalized call rules, so that the call content, time and frequency can better meet the needs and preferences of customers, thereby improving the effectiveness and hit rate of calls. In addition, a real-time monitoring and feedback mechanism is introduced to monitor and analyze key indicators in the call process in real time, and adjust the call strategy in time to ensure the stability of system operation and continuous optimization of call effects. This innovation enables the system to discover and solve problems in a timely manner, continuously improve call effects and user satisfaction. It brings more intelligent, efficient and user-friendly outbound call solutions to areas such as customer service and marketing promotion.

[0176] In this embodiment, an intelligent outbound calling device based on prediction information is also provided, which is used to implement the above-mentioned embodiments and preferred implementation modes, and the descriptions that have been made will not be repeated. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.

[0177] This embodiment provides an intelligent outbound calling device based on prediction information, such as Figure 4 As shown, including:

[0178] The historical data acquisition module 401 is used to acquire a historical data set, where the historical data set includes: historical behavior information corresponding to a plurality of customers;

[0179] A customer behavior prediction module 402 is used to obtain behavior prediction information of each customer through a preset machine learning model based on the historical behavior information corresponding to each customer;

[0180] The outbound call rule determination module 403 is used to determine the outbound call rule for each customer according to the behavior prediction information, and make an outbound call to the customer based on the outbound call rule.

[0181] In some optional implementations, the customer behavior prediction module 402, when obtaining the behavior prediction information of each customer through a preset machine learning model based on the historical behavior information corresponding to each customer, includes:

[0182] Extract customer behavior characteristics based on historical behavior information, including: purchased products, purchase frequency, purchase amount and purchase time, and outbound call answering time;

[0183] By processing behavioral characteristics through preset machine learning models, we can determine the best time period for outbound calls for each customer, as well as the predicted interest and focus of each product.

[0184] In some optional implementations, the outbound call rule determination module 403, when determining the outbound call rule for each customer according to the behavior prediction information, includes:

[0185] Get the recommended products corresponding to the current outbound call task;

[0186] Determine the target interest and target focus of each customer for the recommended product based on the predicted interest and predicted focus of each customer for each product;

[0187] Determine the outbound call frequency for each customer based on the target interest level, and determine the outbound call content for each customer based on the target focus angle;

[0188] Determine the outbound calling rules for each customer based on the outbound calling frequency, outbound calling content and the best outbound calling time period.

[0189] In some optional implementations, the outbound call rule determination module 403 is also used to obtain key indicators in the outbound call process, including: call success rate, single call duration, customer feedback content;

[0190] Update outbound call frequency, outbound call content and optimal outbound call time period based on key indicators.

[0191] In some optional implementations, the outbound call rule determination module 403, when updating the outbound call frequency, outbound call content and optimal outbound call time period based on the key indicators, includes:

[0192] When the call success rate is lower than the preset value, reduce the outbound call frequency;

[0193] Count the call success rates of each time period, and determine the time period with the highest call success rate as the updated optimal outbound call time period;

[0194] The actual interest of the customer in the recommended product is determined based on the monitored customer feedback content, and the call content is updated based on the actual interest.

[0195] In some optional implementations, the outbound call rule determination module 403, when making an outbound call to a customer based on the outbound call rule, includes:

[0196] Outbound calls are made through encrypted communications and / or identity authentication and / or access control.

[0197] In some optional implementations, the preset machine learning model is: a machine learning model constructed based on a random forest algorithm;

[0198] The customer behavior prediction module 402, when obtaining the behavior prediction information of each customer through a preset machine learning model based on the historical behavior information corresponding to each customer, includes:

[0199] The historical behavior information of each customer is analyzed through a machine learning model built based on the random forest algorithm to obtain the behavior prediction information of each customer.

[0200] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0201] The intelligent outbound call device based on predictive information in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0202] The embodiment of the present invention also provides a computer device having the above Figure 4 The intelligent outbound calling device based on predictive information is shown.

[0203] See also Figure 5 , Figure 5 is a schematic diagram of the structure of a computer device provided by an optional embodiment of the present invention, such as Figure 5As shown, the computer device includes: one or more processors 10, a memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components are connected to each other using different buses for communication, and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Similarly, multiple computer devices can be connected, and each device provides some necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 5 A processor 10 is taken as an example.

[0204] The processor 10 may be a central processing unit, a network processor or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be a dedicated integrated circuit, a programmable logic device or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic or any combination thereof.

[0205] The memory 20 stores instructions executable by at least one processor 10, so that at least one processor 10 executes the method shown in the above embodiment.

[0206] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function; the data storage area may store data created according to the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely arranged relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0207] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid state drive; the memory 20 may also include a combination of the above types of memory.

[0208] The computer device also includes an input device 30 and an output device 40. The processor 10, the memory 20, the input device 30 and the output device 40 may be connected via a bus or other means. Figure 5 The example of connecting through bus is taken in the following.

[0209] The input device 30 can receive input digital or character information, and generate key signal input related to the user settings and function control of the computer device, such as a touch screen, a keypad, a mouse, a track pad, a touch pad, an indicator bar, one or more mouse buttons, a trackball, a joystick, etc. The output device 40 may include a display device, an auxiliary lighting device (e.g., an LED) and a tactile feedback device (e.g., a vibration motor), etc. The above-mentioned display device includes but is not limited to a liquid crystal display, a light emitting diode, a display and a plasma display. In some optional embodiments, the display device can be a touch screen.

[0210] The embodiment of the present invention also provides a computer-readable storage medium. The method according to the embodiment of the present invention can be implemented in hardware, firmware, or can be implemented as a computer code that can be recorded in a storage medium, or can be implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and will be stored in a local storage medium through a network download, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state hard disk, etc.; further, the storage medium can also include a combination of the above types of memories. It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor, or hardware, the method shown in the above embodiment is implemented.

[0211] Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention, and such modifications and variations are all within the scope defined by the appended claims.

Claims

1. An intelligent outbound calling method based on prediction information, characterized in that: The method comprises: Acquire a historical data set, the historical data set comprising: historical behavior information corresponding to a plurality of customers; Based on the historical behavior information of each customer, the behavior prediction information of each customer is obtained through the preset machine learning model; An outbound calling rule for each customer is determined according to the behavior prediction information, and an outbound calling is made to the customer based on the outbound calling rule.

2. The method according to claim 1, characterized in that: The behavior prediction information of each customer is obtained by using a preset machine learning model based on the historical behavior information corresponding to each customer, including: Extracting the customer's behavior characteristics based on the historical behavior information, the behavior characteristics including: purchased products, purchase frequency, purchase amount and purchase time, and outbound call answering time; The behavioral characteristics are processed through a preset machine learning model to determine the optimal outbound calling time period for each customer, as well as the predicted interest and focus angle for each type of product.

3. The method according to claim 2, characterized in that Determining the outbound calling rules for each customer according to the behavior prediction information includes: Get the recommended products corresponding to the current outbound call task; Determine the target interest and target attention angle of each customer for the recommended product based on the predicted interest and predicted attention angle of each customer for each product; Determine the outbound calling frequency of each customer based on the target interest level, and determine the outbound calling content of each customer based on the target focus angle; The outbound calling rules for each customer are determined according to the outbound calling frequency, outbound calling content and optimal outbound calling time period.

4. The method according to claim 3, characterized in that The method further comprises: Obtain key indicators during outbound calls, including call success rate, single call duration, and customer feedback content; The outbound call frequency, outbound call content and optimal outbound call time period are updated based on the key indicators.

5. The method according to claim 4, characterized in that The updating of the outbound call frequency, outbound call content and optimal outbound call time period based on the key indicators includes: When the call success rate is lower than a preset value, reducing the outbound call frequency; Count the call success rates of each time period, and determine the time period with the highest call success rate as the updated optimal outbound call time period; The actual interest of the customer in the recommended product is determined according to the monitored customer feedback content, and the call content is updated based on the actual interest.

6. The method according to claim 1, characterized in that When making an outbound call to a customer based on the outbound call rule, the method further includes: Outbound calls are made through encrypted communications and / or identity authentication and / or access control.

7. The method according to claim 1, characterized in that The preset machine learning model is: a machine learning model constructed based on a random forest algorithm; The method of obtaining behavior prediction information of each customer based on the historical behavior information corresponding to each customer by using a preset machine learning model includes: The historical behavior information of each customer is analyzed by the machine learning model constructed based on the random forest algorithm to obtain the behavior prediction information of each customer.

8. An intelligent outbound calling device based on prediction information, characterized in that: The device comprises: A historical data acquisition module is used to acquire a historical data set, wherein the historical data set includes: historical behavior information corresponding to a plurality of customers; A customer behavior prediction module is used to obtain behavior prediction information of each customer through a preset machine learning model based on the historical behavior information corresponding to each customer; The outbound call rule determination module is used to determine the outbound call rule of each customer according to the behavior prediction information, and make outbound calls to the customer based on the outbound call rule.

9. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the intelligent outbound calling method based on predictive information as described in any one of claims 1 to 7 by executing the computer instructions.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the intelligent outbound calling method based on predictive information according to any one of claims 1 to 7.

Images