Outbound prediction method and device, computer equipment and storage medium

Abstract

The invention relates to the technical field of network communication, in particular to an outbound prediction method and device, computer equipment and a storage medium, and the method comprises the steps: obtaining user information data and service type data, and carrying out the analysis according to the user information data and the service type data, obtaining user portrait data correspondingly associated with the service type data; obtaining a batch call-out instruction, and generating corresponding interactive verbal skill data according to the user portrait data through a preset verbal skill large model; obtaining customer demand data corresponding to the interactive verbal skill data, and updating the user portrait data according to the customer demand data to obtain portrait update data; and according to a preset time period, generating a user demand progress according to the portrait updating data, and generating a corresponding user follow-up prompt according to the user demand progress. The method and the device have the effect of improving the call content arrangement efficiency.

Description

Technical Field

[0001] This application relates to the technical field of network communication, and in particular to a method, apparatus, computer device, and storage medium for predicting outbound calls. Background Technology

[0002] Currently, outbound calling is widely used in sales and other fields for customer outreach and marketing. This model typically relies on customer service agents manually dialing out to potential customers to promote products, gather information, or facilitate transactions.

[0003] The existing outbound calling process is as follows: Customer service agents must strictly follow the established script to conduct the call; during the call, they must manually record key information (such as customer intentions, needs, objections, agreed matters, etc.); after the call, the agent must spend a lot of time organizing and entering the handwritten or temporary records into the system, and make subjective judgments and feedback on the result of the call (such as reaching an agreement, rejection, need for follow-up, etc.).

[0004] The existing technical solutions described above have the following drawbacks: Because of the large number of calls made every day, a lot of time is spent processing the call content, which is labor-intensive and inefficient. Summary of the Invention

[0005] To improve the efficiency of call content organization, this application provides a method, apparatus, computer device, and storage medium for predicting outbound calls.

[0006] The above-mentioned objective of this application is achieved through the following technical solution: A method for predicting outbound calls, the method comprising: Obtain user information data and business type data, analyze the user information data and business type data, and obtain user profile data associated with the business type data; Obtain batch call instructions, and generate corresponding interactive script data based on the user profile data using a preset script model; Obtain customer demand data corresponding to the interactive dialogue data, and update the user profile data based on the customer demand data to obtain profile update data; According to the predetermined time cycle, a user demand progress is generated based on the profile update data, and a corresponding user follow-up prompt is generated based on the user demand progress.

[0007] By adopting the above technical solutions, user profile data is obtained by analyzing user information data and business type data. This allows for the preliminary creation of customer profiles, enabling targeted inquiries about potential customer interests and improving the customer call experience. Furthermore, by generating interactive scripts based on user profile data using a pre-set script model, batch calls can be made, improving call efficiency and reducing the overhead of manual calls. Acquiring customer demand data and updating user profile data helps sales personnel understand customer situations. Finally, generating user demand progress and follow-up prompts based on updated profile data reminds sales personnel to follow up on customer situations.

[0008] In a preferred embodiment, this application can be further configured as follows: the step of acquiring user information data and business type data, and analyzing the user information data and business type data to obtain user profile data associated with the business type data, specifically includes: User age data, spending power data, and interest tag data are obtained from the user information data; The user profile data is obtained by analyzing the user age data, the spending power data, the interest tag data, and the business type data.

[0009] By adopting the above technical solution, user age, spending power, and interest tag data are extracted from user information data, and combined with business type data for analysis to obtain user profile data. This allows for more accurate user profiling, which helps to target customers' potential areas of interest and improve their call experience. Furthermore, the profile data can be used to generate more customer-specific interactive scripts and can be updated based on customer needs data to help sales staff understand customer situations. Additionally, updated profile data can be used to generate user follow-up prompts.

[0010] In a preferred embodiment, this application can be further configured such that, before obtaining the batch call-out instructions and generating corresponding interactive dialogue data based on the user profile data using a preset dialogue model, the method for training the dialogue model specifically includes: Obtain the historical script library, and extract excellent script data with order conversion rates exceeding a preset value and corresponding product type data from the historical script library to obtain the script to be trained; Obtain the original user profile data, and use the original user profile data and the training script to train the initial model to obtain the large model of the script.

[0011] By adopting the above technical solution, excellent script data with order conversion rates exceeding preset values ​​and corresponding product type data from the historical script library are obtained as scripts to be trained. Combined with the original user profile data, the initial model is trained to obtain a large script model, which makes the generated interactive scripts more targeted, improves call efficiency, reduces the consumption of manual calls, and helps to improve the customer call experience by accurately portraying customer profiles in the early stage and asking about the needs of customers in areas that they may be interested in.

[0012] In a preferred embodiment, this application can be further configured as follows: the step of obtaining batch outgoing call instructions, through a preset large-scale dialogue model, generates corresponding interactive dialogue data based on the user profile data, specifically including: Obtain the list of users to be connected from the batch call instructions, and obtain the matching profile data corresponding to the list of users to be connected based on the user profile data; The profile data to be matched is input into the large dialogue model, and the corresponding potential business type is obtained from each profile data to be matched; The interactive script information is generated based on the potential business type and the profile data to be matched.

[0013] By adopting the above technical solutions, the list of users to be connected can be obtained from the batch call instructions, which can accurately determine the users that need to be called; the matching profile data corresponding to the list of users to be connected can be obtained based on user profile data, which can make the subsequent scripts more targeted; the matching profile data can be input into the script model to obtain potential business types, and the interactive script information can be generated by combining the potential business types and the matching profile data, which can help improve call efficiency, reduce the consumption of manual calls, and also allow for targeted inquiry into the areas that customers may be interested in, thereby improving the customer's call experience.

[0014] In a preferred embodiment, this application can be further configured as follows: obtaining customer demand data corresponding to the interactive dialogue data, and updating the user profile data based on the customer demand data to obtain profile update data, specifically includes: Obtain user interaction data from the customer demand data, and obtain the intention matching degree from the user interaction data; Based on the intention matching degree and the historical dialogue database, corresponding follow-up strategy data is generated, and the user profile data is updated based on the follow-up strategy data.

[0015] By adopting the above technical solutions, user information data and business type data are acquired and analyzed to obtain user profile data, which helps to analyze users in a targeted manner in conjunction with business needs. Training the large-scale script model can generate more suitable interactive scripts. Obtaining the matching profile data corresponding to the list of users to be contacted and inputting it into the large-scale script model can uncover potential business types to generate interactive scripts. Obtaining user interaction data and intention matching degree from customer demand data, combined with the historical script library, generates follow-up strategy data and updates user profiles, which helps to accurately grasp changes in user needs and adjust follow-up strategies in a timely manner. According to the profile update data, user demand progress and follow-up prompts are generated at predetermined time cycles, which can promptly remind customers to follow up. Overall, it improves call quality and work efficiency, reduces the consumption of manual outbound calls, enhances the customer call experience, and helps sales personnel understand customer situations.

[0016] The second objective of this invention is achieved through the following technical solution: A predictive outbound calling device, the predictive outbound calling device comprising: The number control module is used to acquire user information data and business type data, analyze the user information data and business type data, and obtain user profile data associated with the business type data. The outbound module is used to acquire batch outbound commands and generate corresponding interactive dialogue data based on the user profile data through a preset dialogue model. The profile update module is used to obtain customer demand data corresponding to the interactive dialogue data, and update the user profile data according to the customer demand data to obtain profile update data. The supervision module is used to generate a user demand progress based on the profile update data according to a predetermined time period, and to generate corresponding user follow-up prompts based on the user demand progress.

[0017] By adopting the above technical solutions, user profile data is obtained by analyzing user information data and business type data. This allows for the preliminary creation of customer profiles, enabling targeted inquiries about potential customer interests and improving the customer call experience. Furthermore, by generating interactive scripts based on user profile data using a pre-set script model, batch calls can be made, improving call efficiency and reducing the overhead of manual calls. Acquiring customer demand data and updating user profile data helps sales personnel understand customer situations. Finally, generating user demand progress and follow-up prompts based on updated profile data reminds sales personnel to follow up on customer situations.

[0018] The above-mentioned objective three of this application is achieved through the following technical solution: A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the above-described predictive outbound call method.

[0019] The fourth objective of this application is achieved through the following technical solution: A computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the above-described predictive outbound call method.

[0020] In summary, this application includes at least one of the following beneficial technical effects: 1. By analyzing user information data and business type data, user profile data is obtained, and targeted inquiries are made about the needs of customers in areas that they may be interested in, thereby improving the customer call experience; 2. By generating corresponding interactive script data based on user profile data through a pre-set script model, batch calls can be made, which can improve the efficiency of outbound calls and reduce the consumption of manual outbound calls; 3. Update user profile data based on customer needs data, and generate user follow-up prompts based on the updated profile data. This helps sales staff understand customer situations and follow up with customers in a timely manner. Attached Figure Description

[0021] Figure 1 This is a flowchart of a predictive outbound call method according to an embodiment of this application; Figure 2 This is a block diagram of a predictive outbound call system according to one embodiment of this application; Figure 3 This is a schematic diagram of a device according to one embodiment of this application. Detailed Implementation

[0022] The present application will be further described in detail below with reference to the accompanying drawings.

[0023] In one embodiment, such as Figure 1 As shown, this application discloses a method for predicting outbound calls, which specifically includes the following steps: S10: Obtain user information data and business type data, analyze the user information data and business type data, and obtain user profile data that corresponds to the business type data.

[0024] Specifically, when acquiring user information data, the system collects various aspects of customer information through online channels, including user age data, spending power data, and interest tag data. These data sources are diverse; age data can be obtained through user registration information, publicly available data, etc.; spending power data can be comprehensively assessed by combining user spending records, credit card statements, etc.; and interest tag data can be analyzed through user browsing history, social interaction, and other behaviors. Business type data comes from the company's own business scope and classification.

[0025] Furthermore, the system employs data analysis algorithms when analyzing user information and business type data. For age data, users in different age groups exhibit significant differences in their needs and preferences for products or services. For example, young people may prefer fashion and technology products, while middle-aged and elderly individuals may focus more on health and elderly care services. Through analysis of age and business type data, the system can initially identify age groups that may match different business types. Spending power data is also an important analytical factor. Users with high spending power may be more willing to purchase high-end, high-quality products or services, while users with low spending power are more focused on cost-effectiveness. The system will further refine the matching degree between business types and users based on spending power data. Interest tag data can more accurately reflect users' personalized needs. The system will perform deep matching between users' interest tags and business types. For example, if a user has a strong interest in outdoor sports, then products or services related to outdoor sports will be highlighted. Through such comprehensive and in-depth analysis, the system can obtain user profile data corresponding to business type data, providing precise targeting for subsequent outbound calls.

[0026] S20: Obtain batch call instructions, and generate corresponding interactive script data based on user profile data through a preset script model.

[0027] Specifically, upon receiving batch outbound call commands, a pre-defined large-scale script model is used to generate corresponding interactive script data based on user profile data. Prior to this, the large-scale script model needs to be trained. During training, a historical script library is acquired, containing a large number of past outbound script records. From this library, high-performing scripts with order conversion rates exceeding a preset value, along with corresponding product type data, are selected and used as training scripts. Simultaneously, raw user profile data is acquired, containing basic user characteristics and potential needs. The raw user profile data and the training scripts are used to train the initial model. During training, the model continuously learns the correlation between high-performing scripts and user profiles, as well as the optimal communication methods for different business types. After multiple iterations of training, a large-scale script model capable of generating effective interactive scripts based on user profile data is obtained.

[0028] When generating interactive dialogue data based on batch call instructions, the system retrieves the list of users to be connected from the batch call instructions, and then obtains the corresponding matching profile data based on the user profile data. For each user to be connected, their matching profile data is input into the dialogue model. Based on the association between user profiles and business types learned during previous training, the dialogue model retrieves the corresponding potential business type from each matching profile. Then, based on the potential business type and the matching profile data, personalized interactive dialogue information is generated. For example, if the user's profile data shows that they are young, have high spending power, and are interested in technology products, then the interactive dialogue generated by the dialogue model might emphasize the product's technological content, innovation, and high-end quality.

[0029] S30: Obtain customer demand data corresponding to the interactive dialogue data, update the user profile data based on the customer demand data, and obtain profile update data.

[0030] Specifically, the system acquires customer demand data corresponding to interactive dialogue data and updates user profile data based on this data to obtain updated profile data. During outbound calls, customer feedback is recorded in real time, and user interaction data is extracted from the customer demand data. User interaction data includes various aspects such as the customer's language expression, tone, and questions. The system then extracts the intent matching degree from this data, reflecting the customer's level of interest in the current service and the degree to which their needs are met. For example, if a customer actively inquires about product details, price, and purchase methods, the intent matching degree will be high; conversely, if a customer displays a lukewarm or uninterested attitude, the intent matching degree will be low.

[0031] Furthermore, follow-up strategy data is generated based on the degree of intent matching and the historical dialogue database. If the degree of intent matching is high, the follow-up strategy data may suggest providing more detailed product information or arranging a product demonstration; if the degree of intent matching is low, the follow-up strategy data may suggest adjusting the communication method or recommending other related services. The user profile data is then updated based on the follow-up strategy data, including updates to the user's latest needs, changes in interests, and degree of intent. For example, if a customer expresses interest in another product during the outbound call, this interest tag will be added to the user profile data.

[0032] S40: Generate user demand progress based on profile update data according to the predetermined time cycle, and generate corresponding user follow-up prompts based on the user demand progress.

[0033] Specifically, the system generates a user demand progress report based on user profile update data according to a predetermined time cycle, and generates corresponding user follow-up prompts based on this progress. The system sets a reasonable time cycle, such as daily, weekly, or monthly. At the end of each time cycle, a comprehensive analysis of the user profile update data is performed. Based on the user's latest needs and level of intent, the system assesses the progress of the user demand and generates a user demand progress report. For example, for users with a high degree of intent matching and who have already had multiple communications, the progress report may show as nearing the closing stage; for users with a lower degree of intent matching but still some potential, the progress report may show as the initial contact stage.

[0034] Furthermore, based on the user's progress, corresponding follow-up prompts are generated. For users nearing the closing stage, follow-up prompts may remind sales staff to communicate with the customer promptly, address any questions, and facilitate the transaction. For users in the initial contact stage, follow-up prompts may suggest that sales staff contact the customer again to further understand their needs and adjust communication strategies. This approach encourages relevant personnel to follow up with customers promptly and effectively, thereby improving customer conversion rates and business completion rates.

[0035] In one embodiment, step S10 involves acquiring user information data and business type data, analyzing the user information data and business type data to obtain user profile data associated with the business type data, specifically including: S11: Obtain user age data, spending power data, and interest tag data from user information data.

[0036] Specifically, the foundation for building accurate user profiles begins with acquiring user information and business type data. User information can be acquired in various ways. On one hand, it can be obtained through online channels, such as the company's official website and social media platforms, where users leave relevant information when registering accounts or browsing product or service information; this information is then collected and integrated. On the other hand, it can be obtained through offline channels, such as collecting user information through direct interaction and registration forms at various events and exhibitions. Business type data primarily comes from the company's internal analysis and classification of its own business, clearly defining the various products and services the company provides.

[0037] Next, user age, spending power, and interest tag data are extracted from the user information data. User age can be calculated from the birth date entered during registration, or indirectly inferred by analyzing the content browsed on the website and the activities participated in. Obtaining spending power data is more complex, requiring a comprehensive assessment combining factors such as the user's purchase history, spending amount, and purchase frequency. For example, analyzing the number of times a user purchases high-value products within a certain period and the average amount spent per purchase. Interest tag data can be tagged based on user browsing behavior, search keywords, and saved products or services. For instance, if a user frequently browses content related to electronic products, they can be tagged with the "electronic products" interest tag.

[0038] S12: Analyze user age data, spending power data, interest tag data, and business type data to obtain user profile data.

[0039] Specifically, extracted user age data, spending power data, and interest tag data are analyzed in conjunction with business type data to obtain user profile data. This analysis process is a complex matching and integration process. Regarding user age data, users of different age groups often have different needs and preferences for products and services. For example, young people may prefer fashionable and novel products, while middle-aged and elderly people may place more emphasis on the practicality and stability of products. By matching user age data with business type data, we can understand the potential needs of users of different age groups for the company's various business operations.

[0040] Consumer spending power data also plays a crucial role. Users with high spending power are more likely to purchase high-end, high-quality products or services, while those with low spending power are more focused on cost-effectiveness. Combining spending power data with business type data during the analysis process can determine which businesses are more suitable for users at different spending levels. For example, for users with high spending power, the company's high-end product lines can be highlighted; for users with low spending power, discounted packages or affordable products can be offered.

[0041] Interest tag data further refines user needs. Users with different interest tags show varying degrees of interest in different business types. By matching interest tag data with business type data, it's possible to accurately pinpoint a user's specific business interests. For example, if a user's interest tag is "outdoor sports," then the company can recommend related products or services, such as outdoor equipment or outdoor sports courses.

[0042] During the analysis process, various data analysis methods and techniques can be employed. For example, data mining algorithms can be used to deeply mine large amounts of user information and business type data to identify potential patterns and correlations. Machine learning algorithms can also be used to train models to predict users' interests and needs for different business types. Simultaneously, human analysis and judgment can be combined to further verify and refine the results of the data analysis.

[0043] In one embodiment, prior to step S20, the method for training the large-scale speech model specifically includes: S201: Obtain the historical script library. Extract excellent script data with order conversion rates exceeding preset values ​​and corresponding product type data from the historical script library to obtain the scripts to be trained.

[0044] Specifically, this involves acquiring a historical dialogue script database. A historical dialogue script database is a rich data resource, including a large amount of dialogue script information from past communications with customers in various business scenarios. This dialogue script information comes from actual business call records, sales conversations, etc. Acquiring a historical dialogue script database can be achieved through various means. For some companies, it could be an internal database storing business communication records from many years ago. Through data mining and organization, these records can be extracted and used to build a historical dialogue script database. For example, an e-commerce company can export all chat records with customers from its customer service system, including pre-sales inquiries and after-sales feedback, and then classify and organize these records to form a historical dialogue script database containing dialogue scripts from various business scenarios.

[0045] For newly established companies or those lacking their own data, purchasing external data services can provide access to historical dialogue script databases. Professional data providers collect and organize vast amounts of dialogue script data across different industries and business scenarios, allowing companies to purchase the appropriate historical dialogue script databases based on their needs. Furthermore, companies can acquire even more historical dialogue script information through data sharing or collaboration with others in the same industry.

[0046] Furthermore, high-performing sales scripts with conversion rates exceeding preset values, along with corresponding product type data, are retrieved from the historical script database to obtain the scripts to be trained. The key to this step lies in accurately selecting these high-performing scripts. The preset values ​​need to be determined based on the company's actual situation and business objectives. Generally, historical order data can be analyzed to identify business scenarios with high order conversion rates and their corresponding scripts. For example, in a cosmetics sales company, data analysis revealed that when customer service representatives used specific scripts to introduce product efficacy and promotional activities, the order conversion rate was significantly higher than with other scripts. These scripts can then be selected as high-performing sales script data.

[0047] Furthermore, data mining and machine learning algorithms can be employed during the screening process. For example, association rule mining algorithms can be used to identify the correlation between sales pitches and order conversion rates, thereby determining which sales pitches are truly effective in improving order conversion rates. Simultaneously, it is also necessary to obtain product type data corresponding to these effective sales pitches, as different products may require different sales pitches for promotion. Combining this data with the corresponding product type data yields the sales pitches to be trained.

[0048] S202: Obtain the original user profile data, and train the initial model with the original user profile data and the script to be trained to obtain the large script model.

[0049] Specifically, this involves acquiring raw user profile data. Raw user profile data describes user characteristics and behaviors, helping the large-scale user profile model better understand user needs and preferences. Raw user profile data can be acquired in various ways. Businesses can build the foundation for raw user profiles by collecting basic user information such as age, gender, occupation, and region. For example, an online education company can use the information users fill in during registration to understand their age, education level, learning goals, etc., thus initially constructing a user profile.

[0050] Furthermore, the original user profile can be further refined by analyzing user behavior data. For example, by analyzing users' browsing history, purchase history, and search history on websites, we can understand their interests and consumption habits. Taking a travel company as an example, by analyzing users' search history, we can understand the travel destinations and travel methods that users are interested in, thereby building a more accurate user profile. At the same time, we can also obtain users' subjective opinions and needs through questionnaires, user feedback, and other methods, further enriching the original user profile data.

[0051] Finally, the initial model is trained using the original user profile data and the training scripts to obtain a large-scale script model. During training, it's crucial to select appropriate machine learning algorithms and model architectures. Common machine learning algorithms include neural networks, decision trees, and support vector machines. Taking neural networks as an example, they can learn the correlation between the original user profile data and the training scripts through the connections and computations of multiple layers of neurons.

[0052] During training, the original user profile data and the training scripts are used as input, and business metrics such as order conversion rate are used as output. By continuously adjusting the model's parameters, the model can generate corresponding interactive scripts based on the user profile data as accurately as possible, thereby improving the order conversion rate. Simultaneously, to avoid overfitting, regularization methods such as L1 and L2 regularization are employed.

[0053] During training, the model is evaluated and optimized. The dataset is divided into training and test sets, with the test set used to evaluate the model's performance. If the model's performance on the test set is unsatisfactory, adjustments and optimizations are needed, such as adjusting model parameters or increasing training data. Through continuous training, evaluation, and optimization, a large-scale dialogue model capable of generating effective interactive dialogue based on user profile data is ultimately obtained.

[0054] In one embodiment, step S20, namely obtaining the batch call-out instruction, involves generating corresponding interactive dialogue data based on user profile data using a preset dialogue script model, specifically including: S21: Obtain the list of users to be connected from the batch call instructions, and obtain the matching profile data corresponding to the list of users to be connected based on the user profile data.

[0055] Specifically, the system retrieves the list of users to be contacted from the batch outbound call instructions. Upon receiving a batch outbound call instruction, initiated by relevant business personnel based on business needs, the instruction is parsed to extract the list of users to be contacted. This list may contain basic user identifiers, such as phone numbers and user IDs. Accurately obtaining this information clarifies the target audience for the outbound call, laying the foundation for subsequent personalized interactive script generation. This process, through the instruction parsing module, extracts key user list information from the instruction, thereby avoiding incorrect targeting, improving the relevance of outbound calls, reducing invalid outbound calls, and saving significant time and resources.

[0056] Furthermore, matching profile data is obtained based on the user profile data and the list of users to be connected. In the previous process, user profile data was generated based on user information and business type data. This user profile data includes various user characteristics, such as age, spending power, and interest tags. Once the list of users to be connected is obtained, these users are matched with the existing user profile data. Specifically, a data matching algorithm compares the identifiers in the user list with the relevant identifiers in the user profile data to find the corresponding user profile, ensuring that the most suitable profile data is found for each user to be connected. The matching profile data reflects each user's unique needs and potential interests, providing an important basis for generating personalized interactive scripts. For example, if a user's profile data shows that they are older and have higher spending power, then the subsequent interactive scripts can target this characteristic by recommending some high-end products or services suitable for older users.

[0057] S22: Input the profile data to be matched into the large dialogue model, and obtain the corresponding potential business type from each profile data to be matched.

[0058] Specifically, the user profile data to be matched is input into the large-scale communication script model, from which the corresponding potential business types are extracted. This large-scale model is trained by integrating a large amount of historical communication script data and user profile data. During training, high-performing communication scripts with order conversion rates exceeding preset values, along with corresponding product type data, are selected from the historical script library and used to train the initial model together with the original user profile data. When the user profile data to be matched is input into this trained large-scale model, the model analyzes the potential business types that each user might be interested in based on the characteristics of the profile data, using its internal algorithms and logic. Through the large-scale model's data analysis and predictive capabilities, potential business needs are mined from the user profile data. For example, for a young user interested in technology products, the model might determine that their potential business type is smart electronic products, software services, etc. By obtaining potential business types, the direction of communication with each user can be further clarified, providing key information for generating more targeted interactive scripts.

[0059] S23: Generate interactive dialogue information based on potential business types and profile data to be matched.

[0060] Specifically, interactive scripts are generated based on potential business types and user profile data. After obtaining the potential business types and user profile data, a script generation algorithm is used to generate interactive scripts, combining these two pieces of information. The algorithm comprehensively considers various characteristics of the potential business type and user profile to generate scripts that both meet business needs and attract users. For example, if the potential business type is tourism services, and the user profile shows that the user likes natural scenery and outdoor activities, the generated interactive script might highlight some scenic tourist destinations suitable for outdoor activities and introduce related travel packages. Personalized interactive scripts can greatly improve the effectiveness of communication with users, increase user interest and engagement in the business, and make users feel that the outbound caller truly understands their needs, thereby establishing a better communication atmosphere and improving business conversion rates.

[0061] In one embodiment, step S30, namely obtaining customer demand data corresponding to the interactive dialogue data, and updating the user profile data based on the customer demand data to obtain profile update data, specifically includes: S31: Obtain user interaction data from customer demand data, and obtain the intention matching degree from user interaction data.

[0062] Specifically, user interaction data is obtained from customer demand data. When the predictive outbound calling system makes outbound calls, all information during the entire call process, including customer demand data, is recorded when a customer answers and interacts with the system. To obtain user interaction data, this customer demand data is filtered and extracted. For example, keyword matching rules can be set up; when business-related keywords such as "product price," "service content," and "purchase intention" appear in the call content, the dialogue portion containing these keywords is extracted as user interaction data. Simultaneously, the voice in the call is recognized and analyzed, converted into text, and then processed. Non-textual information such as the customer's tone, speaking speed, and pauses can also be analyzed. For example, a customer speaking rapidly or with an excited tone may indicate their interest in a particular topic, and this information can also be included as part of the user interaction data. Furthermore, customer actions, such as whether they request further product details or inquire about promotional activities during the call, can be recorded as user interaction data.

[0063] Furthermore, intent matching is obtained from user interaction data. Intent matching measures the strength of a customer's need for a service and how well it aligns with the service type. To obtain intent matching, the frequency and importance of keywords appearing in user interaction data are used for evaluation. For example, when a customer repeatedly mentions keywords such as "buy" or "need immediately," it indicates a strong purchase intention, and the intent matching can be increased accordingly. Simultaneously, the degree to which customers inquire about different service types is considered; the more inquiries about a specific service type, the higher their intent matching for that service. The system can also analyze historical data, comparing the customer's past interaction data and purchasing behavior to determine their current intent matching. If the customer has a history of purchasing similar services and shows high interest in this interaction, the intent matching can be further improved. In addition, customer feedback attitudes, such as positive responses and proactive inquiries, can also be considered as factors that improve intent matching.

[0064] S32: Generate corresponding follow-up strategy data based on the intention matching degree and historical dialogue database, and update the user profile data based on the follow-up strategy data.

[0065] Specifically, the system generates corresponding follow-up strategy data based on intent matching and a historical script database. The historical script database contains all script information accumulated from past outbound calls, including successful and unsuccessful scripts. Suitable scripts are selected from this database based on intent matching levels. When intent matching is high, scripts designed to close the deal are chosen, such as explaining the purchase process and emphasizing the expiration date of promotional offers. These scripts guide customers to make a purchase decision quickly. When intent matching is low, scripts designed to build trust and understand needs are chosen, such as asking customers about their specific expectations for the product and introducing its advantages and features. Furthermore, machine learning and algorithm optimization are performed based on the intent matching and historical script database data to continuously adjust and improve the follow-up strategy data. For example, the system can analyze the success rate of using different scripts under different intent matching levels to determine the most suitable follow-up strategy.

[0066] Furthermore, the user profile data is updated based on the follow-up strategy data. User profile data is a description of customer characteristics obtained through analysis of user information and business type. After obtaining the follow-up strategy data, this data is integrated into the user profile data. If the follow-up strategy is a sales-driving strategy for high-intent customers, a "High-intent, pending-purchase" tag can be added to the user profile data, and customer demand information, such as purchase time and quantity, can be updated. If the follow-up strategy is a trust-building strategy for low-intent customers, the customer's interests and concerns can be updated in the user profile data for more precise outbound calls. The user profile data is further updated based on the execution of the follow-up strategy. If the customer's intent matching degree changes after implementing the follow-up strategy, the user profile data is adjusted promptly to ensure it always reflects the customer's latest needs and characteristics.

[0067] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0068] In one embodiment, a predictive outbound calling device is provided, which corresponds one-to-one with the predictive outbound calling method described in the above embodiments. For example... Figure 2 As shown, the predictive outbound calling device includes a number control module, an outbound calling module, a profile update module, and a monitoring module. Detailed descriptions of each functional module are as follows: The number control module is used to acquire user information data and business type data, and analyze the user information data and business type data to obtain user profile data that is associated with the business type data. The outbound module is used to obtain batch outbound commands and generate corresponding interactive script data based on user profile data through a preset script model. The profile update module is used to obtain customer demand data corresponding to interactive dialogue data, update user profile data based on customer demand data, and obtain profile update data. The supervision module is used to generate user demand progress based on profile update data according to a predetermined time cycle, and generate corresponding user follow-up prompts based on the user demand progress.

[0069] Optionally, the number control module includes: The data acquisition submodule is used to obtain user age data, spending power data, and interest tag data from user information data; The user profile building submodule is used to analyze user age data, spending power data, interest tag data, and business type data to obtain user profile data.

[0070] Optionally, the predictive outbound calling device also includes: The training script acquisition module is used to acquire the historical script library. It retrieves excellent script data with order conversion rates exceeding preset values ​​and corresponding product type data from the historical script library to obtain the scripts to be trained. The model training module is used to acquire raw user profile data, and then use the raw user profile data and the scripts to be trained to train the initial model, resulting in a large script model.

[0071] Optionally, the call-out module includes: The profile matching submodule is used to obtain the list of users to be connected from the batch call instructions, and to obtain the profile data to be matched corresponding to the list of users to be connected based on the user profile data. The business type acquisition submodule is used to input the profile data to be matched into the large dialogue model and obtain the corresponding potential business type from each profile data to be matched; The script production submodule is used to generate interactive script information based on potential business types and profile data to be matched.

[0072] Optionally, the profile update module includes: The matching degree acquisition submodule is used to obtain user interaction data from customer demand data, and obtain the intention matching degree from user interaction data; The profile update submodule is used to generate corresponding follow-up strategy data based on the intention matching degree and the historical dialogue library, and update the user profile data based on the follow-up strategy data.

[0073] Specific limitations regarding the predictive outbound calling device can be found in the limitations of the predictive outbound calling method described above, and will not be repeated here. Each module in the aforementioned predictive outbound calling device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device in hardware form, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module.

[0074] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 3 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The network interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements a predictive outbound calling method.

[0075] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the following steps: Acquire user information data and business type data, analyze the user information data and business type data to obtain user profile data that corresponds to the business type data; Obtain batch call instructions and generate corresponding interactive script data based on user profile data through a preset script model; Obtain customer demand data corresponding to interactive dialogue data, update user profile data based on customer demand data, and obtain profile update data; According to the predetermined time cycle, the user demand progress is generated based on the profile update data, and corresponding user follow-up prompts are generated based on the user demand progress.

[0076] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, the computer program performing the following steps when executed by a processor: Acquire user information data and business type data, analyze the user information data and business type data to obtain user profile data that corresponds to the business type data; Obtain batch call instructions and generate corresponding interactive script data based on user profile data through a preset script model; Obtain customer demand data corresponding to interactive dialogue data, update user profile data based on customer demand data, and obtain profile update data; According to the predetermined time cycle, the user demand progress is generated based on the profile update data, and corresponding user follow-up prompts are generated based on the user demand progress.

[0077] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0078] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0079] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for predicting outbound calls, characterized in that, The method for predicting outbound calls includes: Obtain user information data and business type data, analyze the user information data and business type data, and obtain user profile data associated with the business type data; Obtain batch call instructions, and generate corresponding interactive script data based on the user profile data using a preset script model; Obtain customer demand data corresponding to the interactive dialogue data, and update the user profile data based on the customer demand data to obtain profile update data; According to the predetermined time cycle, a user demand progress is generated based on the profile update data, and a corresponding user follow-up prompt is generated based on the user demand progress.

2. The predictive outbound call method according to claim 1, characterized in that, The process of acquiring user information data and business type data, and analyzing the user information data and business type data to obtain user profile data associated with the business type data, specifically includes: User age data, spending power data, and interest tag data are obtained from the user information data; The user profile data is obtained by analyzing the user age data, the spending power data, the interest tag data, and the business type data.

3. The predictive outbound call method according to claim 1, characterized in that, Before acquiring batch call-out instructions and generating corresponding interactive dialogue data based on the user profile data using a preset dialogue model, the method for training the dialogue model specifically includes: Obtain the historical script library, and extract excellent script data with order conversion rates exceeding a preset value and corresponding product type data from the historical script library to obtain the script to be trained; Obtain the original user profile data, and use the original user profile data and the training script to train the initial model to obtain the large model of the script.

4. The predictive outbound call method according to claim 3, characterized in that, The process of acquiring batch outbound call instructions involves generating corresponding interactive dialogue data based on the user profile data using a preset dialogue script model. Specifically, this includes: Obtain the list of users to be connected from the batch call instructions, and obtain the matching profile data corresponding to the list of users to be connected based on the user profile data; The profile data to be matched is input into the large dialogue model, and the corresponding potential business type is obtained from each profile data to be matched; The interactive script information is generated based on the potential business type and the profile data to be matched.

5. The predictive outbound call method according to claim 3, characterized in that, The step of obtaining customer demand data corresponding to the interactive dialogue data and updating the user profile data based on the customer demand data to obtain profile update data specifically includes: Obtain user interaction data from the customer demand data, and obtain the intention matching degree from the user interaction data; Based on the intention matching degree and the historical dialogue database, corresponding follow-up strategy data is generated, and the user profile data is updated based on the follow-up strategy data.

6. A predictive outbound call device, characterized in that, The predictive outbound call device includes: The number control module is used to acquire user information data and business type data, analyze the user information data and business type data, and obtain user profile data associated with the business type data. The outbound module is used to acquire batch outbound commands and generate corresponding interactive dialogue data based on the user profile data through a preset dialogue model. The profile update module is used to obtain customer demand data corresponding to the interactive dialogue data, and update the user profile data according to the customer demand data to obtain profile update data. The supervision module is used to generate a user demand progress based on the profile update data according to a predetermined time period, and to generate corresponding user follow-up prompts based on the user demand progress.

7. The predictive outbound call device according to claim 6, characterized in that, The number control module includes: The data acquisition submodule is used to acquire user age data, spending power data, and interest tag data from the user information data; The user profile construction submodule is used to analyze the user age data, spending power data, interest tag data, and business type data to obtain the user profile data.

8. The predictive outbound call device according to claim 6, characterized in that, The predictive outbound call device also includes: The training script acquisition module is used to acquire a historical script library, and to acquire excellent script data with order conversion rates exceeding a preset value and corresponding product type data from the historical script library to obtain the script to be trained. The model training module is used to acquire raw user profile data, and to train the initial model with the raw user profile data and the script to be trained to obtain the large script model.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the predictive outbound call method as described in any one of claims 1 to 5.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the predictive outbound call method as described in any one of claims 1 to 5.

Images