Virtual number intelligent scheduling and callback association method, device, equipment and medium

Abstract

The application provides a virtual number intelligent scheduling and callback association method and device, equipment and medium, comprising: constructing a Prophet time series prediction model; constructing a training data set, training the Prophet time series prediction model using the constructed training data set, and obtaining a trained Prophet time series prediction model; using the trained Prophet time series prediction model, estimating the total outbound call number based on the date attribute of the day and the order size of the day, and combining the daily average carrying call volume of a single number to obtain the total number of virtual numbers required on the day.

Description

Technical Field

[0001] This invention relates to the field of logistics management technology, specifically to a method, apparatus, equipment, and medium for intelligent scheduling and callback association of virtual numbers, and more specifically to a method, apparatus, equipment, and medium for intelligent scheduling and callback association of virtual numbers for logistics outbound calling scenarios. Background Technology

[0002] Logistics industry sales representatives commonly use virtual numbers for customer delivery and after-sales inquiries. However, existing virtual number scheduling systems suffer from several flaws, including: a lack of intelligent management of virtual number resources; the entire process of number application, allocation, and recycling relies on manual statistics, making it impossible to monitor number inventory and usage frequency in real time. Relying solely on experience for rough estimates easily leads to management chaos such as number shortages and idle waste; the absence of a unified load constraint mechanism means that if a single virtual number is dialed too frequently, exceeding the operator's risk control threshold, it can be blocked, directly interrupting the sales representative's normal outbound calling business; when customers call back based on the virtual number used for inquiries, the existing system cannot accurately match the initial sales representative, resulting in random transfers to customer service personnel, disrupting communication, requiring customers to repeatedly describe waybill issues, and leading to a poor service experience; and to achieve simple call-back transfers, existing solutions require multiple complex middleware components to build forwarding links, resulting in a bloated system architecture, high maintenance costs, and the inability to link to corresponding logistics waybill information, making business traceability difficult. The overall system relies heavily on manual assistance for transfers, leading to low efficiency. Traditional usage forecasting methods fail to account for call volume fluctuations caused by business cycles, holidays, and shift schedules, resulting in delayed number replenishment or excessive reserve redundancy and poor resource allocation flexibility. In summary, existing virtual number scheduling methods cannot meet the business needs of high-frequency outbound calls, callback traceability, and waybill binding in logistics. The industry urgently needs a virtual number intelligent scheduling and callback association solution that integrates artificial intelligence prediction capabilities, automation, traceability, and anti-blocking features. Summary of the Invention

[0003] The main objective of this invention is to solve the problem that the existing virtual number scheduling method cannot adapt to the business needs of high-frequency outbound calls, callback tracing, and waybill binding in logistics.

[0004] The first aspect of this invention provides a method for intelligent scheduling and callback association of virtual numbers, comprising: Construct a Prophet time series forecasting model; Construct a training dataset and use it to train the Prophet time series prediction model to obtain the trained Prophet time series prediction model. Using the trained Prophet time series prediction model, the total number of outbound calls is estimated based on the date attribute and the order size of the day. Then, combined with the average daily call volume of a single number, the total number of virtual numbers required for the day is obtained.

[0005] Optionally, in a first implementation of the first aspect of the present invention, constructing the training dataset includes: Collect historical time-series business data, and preprocess the collected historical time-series business data to obtain preprocessed historical time-series business data. Business features are extracted based on preprocessed historical time-series business data, including: date and timestamp, total number of outbound calls on the day, date attribute tags, actual number of virtual numbers consumed on the day, call volume during morning and evening shifts, and total number of regional logistics orders; wherein, the date attribute tags include: weekdays, weekends, statutory holidays, e-commerce promotions, and holidays when local logistics are suspended. The training dataset is a set of business feature samples extracted from historical time-series business data within a preset time period.

[0006] Optionally, in a second implementation of the first aspect of the present invention, the constructed training dataset is used to train the Prophet time series prediction model, including: The training dataset is input into the Prophet time series prediction model. The Prophet time series prediction model is trained using a gradient iterative method to learn the relationship between business characteristics and total outbound call count and time-segmented number consumption rate, thus obtaining the trained Prophet time series prediction model.

[0007] Optionally, in a third implementation of the first aspect of the present invention, the method further includes: constructing a logistics-specific virtual number pool and monitoring the status of numbers in the logistics-specific virtual number pool in real time, including: monitoring the number of numbers used, the number of numbers remaining available, and the idle time in the number pool; The predicted total number of virtual numbers needed for the day is compared with the current stock. At the same time, the number of remaining numbers is monitored in real time. When the number of available virtual numbers drops to the warning threshold or below, and the call volume during the morning and evening shifts is predicted to increase the rate of number consumption in the subsequent periods, a number application polling request is sent to the operator interface to replenish the number pool in batches.

[0008] Optionally, in the fourth implementation of the first aspect of the present invention, the method further includes: retrieving the cumulative number of calls made by each number in the number pool on a given day, and using the daily call limit of a single virtual number as a risk control constraint standard, filtering out idle virtual numbers with the lowest daily call frequency and the lowest load pressure, and allocating them to the salesperson terminal that is currently initiating the call according to a load balancing strategy; after the allocation is completed, updating the call count of the current number in real time until the call count of the current number reaches the daily call limit, at which point the current number will no longer be allocated to the salesperson terminal.

[0009] Optionally, in the fifth implementation of the first aspect of the present invention, the method further includes: after each outbound call is successfully established, automatically capturing all information of the call and constructing a standardized association data table for solidified storage; The standardized associated data table stores the following fields: the currently used virtual number, the customer's called mobile phone number, the salesperson's unique ID, the logistics tracking number corresponding to this call, and the time the call was initiated. The standardized associated data table stores records sequentially by call time, forming a traceable data archive for each outbound call. It also establishes a temporary binding relationship between the virtual number and the corresponding salesperson and waybill. The binding record is retained after the call ends, and the temporary binding is automatically released after a preset time if there are no associated calls, thus realizing dynamic unbinding management.

[0010] Optionally, in a sixth implementation of the first aspect of the present invention, the method further includes: triggering an AI routing retrieval program when a customer calls back a virtual number with a binding relationship using a mobile phone number; The program uses the virtual caller ID as the search keyword, traverses the standardized associated data table, filters out the latest call record under that virtual number, extracts the corresponding salesperson ID from the record, and automatically transfers the customer callback call to the salesperson's terminal device corresponding to that ID. At the same time, a pop-up window automatically displays the logistics waybill details bound to this call on the salesperson's operation interface, so that the salesperson can directly contact the customer for consultation based on the waybill information and complete the closed-loop business communication.

[0011] A second aspect of the present invention provides a virtual number intelligent scheduling and callback association device, characterized in that it comprises: The model building module is used to build Prophet time series prediction models; The training module is used to build a training dataset and use the built training dataset to train the Prophet time series prediction model to obtain the trained Prophet time series prediction model. The prediction module uses the trained Prophet time series prediction model to estimate the total number of outbound calls based on the date attributes and the order size of the day. It then combines this with the average daily call volume of a single number to obtain the total number of virtual numbers required for the day.

[0012] Optionally, in a first implementation of the second aspect of the present invention, constructing the training dataset includes: The raw data acquisition submodule is used to collect historical time-series business data and preprocess the collected historical time-series business data to obtain preprocessed historical time-series business data. The feature extraction submodule is used to extract business features based on preprocessed historical time-series business data, including: date and timestamp, total number of outbound calls on the day, date attribute tags, actual consumption of virtual numbers on the day, call volume during morning and evening shifts, and total number of regional logistics orders; wherein, the date attribute tags include: weekdays, weekends, statutory holidays, e-commerce promotions, and holidays when local logistics are suspended. The training dataset construction submodule is used to extract a set of business feature samples based on historical time-series business data for a preset period as the training dataset.

[0013] Optionally, in a second implementation of the second aspect of the present invention, the training module includes: The training dataset is input into the Prophet time series prediction model. The Prophet time series prediction model is trained using a gradient iterative method to learn the relationship between business characteristics and total outbound call count and time-segmented number consumption rate, thus obtaining the trained Prophet time series prediction model.

[0014] Optionally, in a third implementation of the second aspect of the present invention, the device further includes: a number pool monitoring module, used to construct a logistics-specific virtual number pool and to monitor the status of numbers in the logistics-specific virtual number pool in real time, including: monitoring the number of numbers used, the number of numbers remaining available, and the idle time in the number pool; The predicted total number of virtual numbers needed for the day is compared with the current stock. At the same time, the number of remaining numbers is monitored in real time. When the number of available virtual numbers drops to the warning threshold or below, and the call volume during the morning and evening shifts is predicted to increase the rate of number consumption in the subsequent periods, a number application polling request is sent to the operator interface to replenish the number pool in batches.

[0015] Optionally, in the fourth implementation of the second aspect of the present invention, the device further includes: an allocation module, configured to retrieve the cumulative number of calls made by each number in the number pool on a given day, and use the daily call limit of a single virtual number as a risk control constraint standard to filter idle virtual numbers whose daily call frequency has not reached the limit and whose load pressure is the lowest, and allocate them to the salesperson terminal that is currently initiating the call according to a load balancing strategy; after allocation, the call count of the current number is updated in real time until the call count of the current number reaches the daily call limit, at which point the current number is no longer allocated to the salesperson terminal.

[0016] Optionally, in the fifth implementation of the second aspect of the present invention, the device further includes: a data storage module, used to automatically capture all information of the call after each outbound call is successfully established, and construct a standardized association data table to complete the solidified storage; The standardized associated data table stores the following fields: the currently used virtual number, the customer's called mobile phone number, the salesperson's unique ID, the logistics tracking number corresponding to this call, and the time the call was initiated. The standardized associated data table stores records sequentially by call time, forming a traceable data archive for each outbound call. It also establishes a temporary binding relationship between the virtual number and the corresponding salesperson and waybill. The binding record is retained after the call ends, and the temporary binding is automatically released after a preset time if there are no associated calls, thus realizing dynamic unbinding management.

[0017] Optionally, in a sixth implementation of the second aspect of the present invention, the device further includes: an intelligent callback module, used to trigger an AI routing retrieval program when a customer calls back a virtual number with a binding relationship using a mobile phone number; The program uses the virtual caller ID as the search keyword, traverses the standardized associated data table, filters out the latest call record under that virtual number, extracts the corresponding salesperson ID from the record, and automatically transfers the customer callback call to the salesperson's terminal device corresponding to that ID. At the same time, a pop-up window automatically displays the logistics waybill details bound to this call on the salesperson's operation interface, so that the salesperson can directly contact the customer for consultation based on the waybill information and complete the closed-loop business communication.

[0018] A third aspect of the present invention provides an electronic device, the electronic device including a memory and at least one processor, wherein the memory stores instructions; The at least one processor invokes the instructions in the memory to cause the electronic device to perform the various steps of the virtual number intelligent scheduling and callback association method as described above.

[0019] A fourth aspect of the present invention provides a computer-readable storage medium storing instructions that, when executed by a processor, implement the various steps of the virtual number intelligent scheduling and callback association method described above.

[0020] Compared with the prior art, the present invention has the following beneficial effects: 1. This invention introduces the Prophet time series prediction model and incorporates multi-dimensional business features such as date and time stamps, date attribute tags, and call volume during morning and evening shifts into the model training. This allows the model to simultaneously learn the mapping relationship between business features and total outbound calls and time-segmented number consumption rates. It can accurately capture call fluctuations caused by logistics business cycles, holidays, and morning and evening shifts, replacing the traditional manual experience-based number estimation method. By combining the stock threshold and time-segment consumption growth rate for automatic number application, it avoids virtual number shortages and call interruptions during peak business periods and prevents numbers from being idle and stockpiled for a long time during off-peak periods, significantly improving the overall utilization rate of virtual number resources and reducing manual operation and maintenance costs. 2. This invention sets a daily limit on the number of calls a single virtual number can make as a risk control constraint, uses a load balancing strategy to select low-load idle numbers and allocate them to salespersons, updates the call count of each number in real time, and evenly distributes the call pressure of all numbers, effectively avoiding the problem of high-frequency calls from a single number triggering the operator's risk control blocking, and ensuring the uninterrupted and stable operation of logistics outbound call business; 3. This invention establishes a dynamic temporary binding relationship between the number, the salesperson, and the waybill by constructing a standardized time-series association data table containing virtual numbers, customer mobile phone numbers, salesperson IDs, logistics waybill numbers, and call times. The binding is automatically unbound and invalid association records are cleared after a long period of no calls. This reduces data table redundancy, improves retrieval efficiency, and enables the archiving and traceability of all call business data, which is convenient for checking order disputes and business issues. 4. This invention automatically matches the original salesperson corresponding to the customer's callback by searching the associated data table based on virtual number keywords and simultaneously pops up the waybill information. There is no need to build a complex middleware forwarding link, which simplifies the system architecture and reduces operation and maintenance costs. Customers do not need to repeatedly describe waybill issues, resulting in stronger communication continuity. This significantly improves the customer consultation service experience and the efficiency of salesperson communication and processing, forming a complete business closed loop of outbound call-callback. Attached Figure Description

[0021] Other features, objects, and advantages of the present invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a first flowchart of the virtual number intelligent scheduling and callback association method provided in an embodiment of the present invention.

[0022] Figure 2This is a second flowchart of the virtual number intelligent scheduling and callback association method provided in an embodiment of the present invention.

[0023] Figure 3 This is a third flowchart of the virtual number intelligent scheduling and callback association method provided in the embodiments of the present invention.

[0024] Figure 4 This is a schematic diagram of a virtual number intelligent scheduling and callback association device provided in an embodiment of the present invention.

[0025] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0026] This invention provides a method, apparatus, device, and medium for intelligent scheduling and callback association of virtual numbers, comprising: constructing a Prophet time series prediction model; constructing a training dataset, training the Prophet time series prediction model using the constructed training dataset to obtain a trained Prophet time series prediction model; using the trained Prophet time series prediction model, estimating the total number of outbound calls based on the date attributes and the order size of the day, and then combining this with the average daily call volume of a single number to obtain the total number of virtual numbers required for the day. This invention solves the problem that existing virtual number scheduling methods cannot adapt to the business needs of high-frequency outbound calls, callback tracing, and waybill binding in logistics. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0027] For ease of understanding, the specific process of the embodiments of the present invention is described below. Please refer to [link / reference]. Figure 1 The first embodiment of the virtual number intelligent scheduling and callback association method in this invention includes: 101. Construct a Prophet time series forecasting model; This embodiment uses the Prophet time series algorithm to construct a dedicated logistics outbound call number usage prediction model, which differs from general time series prediction models and is specifically adapted to the cyclical fluctuation characteristics of logistics operations. The model includes a built-in trend decomposition module with non-linear segmented trends to adapt to long-term business growth changes; a period fitting module, including weekly and monthly periods, matching logistics weekly schedules and monthly shipping patterns; and a holiday disturbance module, including special labels for e-commerce promotions and local logistics shutdowns during holidays.

[0028] 102. Construct a training dataset and use the constructed training dataset to train the Prophet time series prediction model to obtain the trained Prophet time series prediction model. In this embodiment, historical time-series business data is collected and preprocessed to obtain preprocessed historical time-series business data. For example, historical time-series business data of logistics outbound calls over the past 12 months is selected as the original data source. The data collection frequency is daily, covering business data across all scenarios, including weekdays, weekends, public holidays, e-commerce promotional periods, and regional logistics shutdowns, ensuring comprehensive sample coverage. Preprocessing of the collected original historical time-series business data includes: removing invalid call data, null values, and outlier data caused by system failures; standardizing the data time format and statistical methods; and eliminating data noise interference with model training, resulting in standardized and normalized preprocessed historical time-series business data.

[0029] Business features are extracted based on preprocessed historical time-series business data, including: date and time stamps, accurate to year, month and day; total number of outbound calls on the day; date attribute tags; actual number of virtual numbers consumed on the historical day; call volume during morning and evening shifts; and total number of regional logistics orders. The date attribute tags include: weekdays, weekends, statutory holidays, e-commerce promotions, and holidays when local logistics are suspended. The training dataset is a set of business feature samples extracted from historical time-series business data within a preset time period.

[0030] The training dataset is input into the Prophet time series prediction model, and the Prophet time series prediction model is trained by gradient iteration to learn the relationship between business characteristics and total outbound call count and time-segmented number consumption rate, thus obtaining the trained Prophet time series prediction model.

[0031] 103. Using the trained Prophet time series prediction model, estimate the total number of outbound calls based on the date attribute and the order size of the day, and then combine this with the average daily call volume of a single number to obtain the total number of virtual numbers required for the day.

[0032] In this embodiment, core input parameters for the day are automatically collected at a fixed time in the early morning each day. These parameters include the complete date attribute label, identification of whether it is a weekday, a major promotional period, a period of service suspension, and the predicted regional logistics order volume for the day. These parameters are then input into the trained Prophet prediction model. Based on the business patterns learned during training, the model accurately outputs the estimated total number of outbound calls for the day. Combined with a preset fixed parameter of the average daily call volume per number, the model accurately calculates the total number of virtual numbers required for the day using the formula: "Total number of virtual numbers required for the day = Estimated total number of outbound calls for the day / Average daily call volume per number".

[0033] Please see Figure 2 The second embodiment of the virtual number intelligent scheduling and callback association method in this invention includes: 201. Construct a Prophet time series forecasting model; 202. Construct a training dataset and use the constructed training dataset to train the Prophet time series prediction model to obtain the trained Prophet time series prediction model. 203. Using the trained Prophet time series prediction model, estimate the total number of outbound calls based on the date attributes and order size of the day, and then combine this with the average daily call volume of a single number to obtain the total number of virtual numbers required for the day. 204. Construct a dedicated virtual number pool for logistics and monitor the status of numbers in the dedicated virtual number pool in real time, including: monitoring the number of numbers used, the number of numbers remaining available, and the idle time in the number pool; The total number of virtual numbers required for the day is predicted and compared with the current stock. At the same time, the number of remaining numbers is monitored in real time. When the number of available virtual numbers drops to the warning threshold or below, and the call volume during the morning and evening shifts is predicted to increase the rate of number consumption in the subsequent periods, a number application polling request is sent to the operator interface to replenish the number pool in batches. In this embodiment, a dedicated virtual number resource pool for logistics is independently constructed to achieve centralized and dedicated management and control of logistics outbound call number resources, isolate interference from general industry number resources, and adapt to the business characteristics of high-frequency and time-segmented outbound calls in logistics. The status data of all numbers in the number pool is collected in a polling manner at a rate of seconds, accurately counting and updating in real time three core status data categories: the number of numbers used, the number of numbers remaining available, and the idle time of a single number, thus achieving comprehensive and visualized management and control of number resource status.

[0034] A preset number inventory warning threshold is set, which can be adaptively adjusted according to the scale of logistics business. In this embodiment, the default threshold for remaining available numbers is set to 10% of the predicted total demand for the day. The model compares the total number of virtual numbers required for the day as predicted by the model with the current actual inventory in the number pool in real time. At the same time, it combines the number consumption rate of the model based on the call volume of the morning and evening shifts to predict the number consumption trend in the next 1-2 hours. When the number of remaining available numbers drops to or below the preset warning threshold, and the morning and evening peak periods are identified and the number consumption rate is accelerating, the automatic replenishment mechanism is immediately triggered. The operator's open interface is continuously called through polling requests to apply for compliant virtual numbers in batches and enter them into the number pool at the same time. They are automatically marked as available, and the number resources are dynamically replenished. When the business is at a low point and the consumption rate slows down, the application operation is stopped to avoid excessive hoarding of numbers and to prevent the number supply from being cut off at the source, while also preventing the waste of excessive hoarding of numbers.

[0035] 205. Retrieve the cumulative number of calls made to each number in the number pool for the day, and use the daily call limit for a single virtual number as the risk control constraint standard to filter idle virtual numbers with the lowest daily call frequency and the lowest load pressure, and allocate them to the salesperson terminal that is currently initiating the call according to the load balancing strategy; after the allocation is completed, update the call count of the current number in real time until the call count of the current number reaches the daily call limit, and then no longer allocate the current number to the salesperson terminal; This embodiment addresses the industry pain point of frequent calls from virtual numbers easily triggering operator risk control blocking. It presets a daily call limit for a single virtual number as a hard risk control constraint. Combined with the logistics outbound call business scenario, it sets a maximum daily call limit of 250 calls per number, mitigating the risk of high-frequency risk control from a rule-based perspective. It also dynamically retrieves the cumulative daily call count for each available virtual number in the number pool and dynamically calculates the current load occupancy of each number.

[0036] When a salesperson's terminal initiates an outbound call request, the system automatically iterates through the number pool to filter eligible idle numbers. Priority is given to low-load virtual numbers that have not reached their daily call limit, have the lowest cumulative daily call volume, and the longest idle time. Following a load balancing strategy, the optimal number is assigned to the current salesperson's terminal. After each outbound call is completed, the corresponding virtual number's daily call count, load status, and idle time are updated in real-time, continuously and dynamically updating the load data for each number. When a virtual number's daily call count reaches the preset limit, it is automatically marked as unassignable, suspending outbound allocation and preventing further scheduling for the rest of the day.

[0037] Please see Figure 3 The third embodiment of the virtual number intelligent scheduling and callback association method in this invention includes: 301. Construct a Prophet time series forecasting model; 302. Construct a training dataset and use the constructed training dataset to train the Prophet time series prediction model to obtain the trained Prophet time series prediction model. 303. Using the trained Prophet time series prediction model, estimate the total number of outbound calls based on the date attribute and the order size of the day, and then combine this with the average daily call volume of a single number to obtain the total number of virtual numbers required for the day.

[0038] 304. Construct a dedicated virtual number pool for logistics and monitor the status of numbers in the dedicated virtual number pool in real time, including: monitoring the number of numbers used, the number of numbers remaining available, and the idle time in the number pool; The total number of virtual numbers required for the day is predicted and compared with the current stock. At the same time, the number of remaining numbers is monitored in real time. When the number of available virtual numbers drops to the warning threshold or below, and the call volume during the morning and evening shifts is predicted to increase the rate of number consumption in the subsequent periods, a number application polling request is sent to the operator interface to replenish the number pool in batches. 305. Retrieve the cumulative number of calls made to each number in the number pool for the day, and use the daily call limit for a single virtual number as the risk control constraint standard to filter idle virtual numbers with the lowest daily call frequency and the lowest load pressure, and allocate them to the salesperson terminal that is currently initiating the call according to the load balancing strategy; after the allocation is completed, update the call count of the current number in real time until the call count of the current number reaches the daily call limit, and then stop allocating the current number to the salesperson terminal; 306. After each outbound call is successfully established, all information of the call is automatically captured and a standardized association data table is constructed for fixed storage. The standardized association data table stores fields including: the currently used virtual number, the customer's called mobile phone number, the salesperson's unique ID, the logistics waybill number corresponding to this call, and the call initiation time. The standardized association data table retains records in order of call time, forming a traceable data file for each outbound call business, and simultaneously establishing a temporary binding relationship between the virtual number and the corresponding salesperson and waybill. The binding record is retained after the call ends, and the temporary binding is automatically released after a preset time if there are no associated calls, realizing dynamic unbinding management. In this embodiment, after each outbound call by a salesperson successfully establishes a connection, a data capture mechanism is automatically triggered immediately to collect all dimensions of business data for the call without delay. No manual data entry is required, achieving fully automated collection and archiving of call data. A dedicated standardized relational data table is built, with fixed storage fields, specifically including: the currently used virtual number, the customer's called mobile phone number, the salesperson's unique ID, the logistics tracking number corresponding to this call, and the call initiation time. This achieves structured, standardized, and fixed storage of business data for each outbound call.

[0039] The data table strictly sorts and stores all call records according to the chronological order of call initiation time, forming a fully traceable electronic archive of logistics outbound calls, ensuring that the data for each waybill and each call is searchable and traceable. Simultaneously, a temporary binding mapping relationship is established in real time, with a one-to-one correspondence between the virtual number, salesperson ID, and logistics waybill number. This binding relationship is only effective for the current business cycle and is used for accurate routing matching of subsequent customer callbacks. This embodiment presets a 7-day unbinding threshold for no calls. Once the binding relationship is generated, it is continuously retained. If the virtual number does not generate any associated outbound calls or customer callbacks within 7 days, the business process is automatically determined to be completed, the binding relationship becomes invalid, and the invalid binding mapping is automatically cleared in the background, completing the dynamic unbinding. If associated calls occur during this period, the unbinding timer period is automatically refreshed.

[0040] 307. When a customer calls back a virtual number that is linked to their mobile phone number, the AI ​​routing retrieval program is triggered. The program uses the virtual number as the search keyword, traverses the standardized association data table, filters out the latest call record under that virtual number, extracts the corresponding salesperson ID from the record, and automatically transfers the customer's callback call to the salesperson's terminal device corresponding to that ID. At the same time, the logistics waybill details linked to this call are automatically displayed in a pop-up window on the salesperson's operation interface, so that the salesperson can directly contact the customer for consultation based on the waybill information and complete the closed-loop business communication. In this embodiment, when a customer calls back a previously linked virtual number using their personal mobile phone number, the AI ​​routing retrieval program is immediately triggered, initiating an intelligent callback matching process. The system automatically uses the virtual number as the unique search keyword, traversing all time-series call records corresponding to that virtual number in the standardized association data table. Based on a reverse time order rule, it filters out the latest and most recent valid business call records, accurately extracting the unique ID of the target salesperson bound to those records.

[0041] Based on the extracted salesperson ID, the system automatically transfers customer callback calls to the corresponding salesperson's terminal device, achieving precise targeted callback transfer rather than the traditional random transfer mode. Simultaneously, a pop-up window automatically displays detailed information about the corresponding logistics order in the salesperson's backend, including the waybill address, delivery status, historical communication records, and any issues. This allows salespeople to understand customer needs without secondary queries, directly connect with customers to answer inquiries, resolve problems, and achieve a closed loop for outbound and callback services.

[0042] The above describes the method for intelligent scheduling and callback association of virtual numbers in the embodiments of the present invention. The following describes the device for intelligent scheduling and callback association of virtual numbers in the embodiments of the present invention. Please refer to [link to relevant documentation]. Figure 4 One embodiment of the virtual number intelligent scheduling and callback association device in this invention includes: Model building module 401 is used to build the Prophet time series prediction model; Training module 402 is used to construct a training dataset and use the constructed training dataset to train the Prophet time series prediction model to obtain the trained Prophet time series prediction model. In this embodiment, the training module 402 includes: The raw data acquisition submodule 4021 is used to collect historical time-series business data and preprocess the collected historical time-series business data to obtain preprocessed historical time-series business data. The feature extraction submodule 4022 is used to extract business features based on preprocessed historical time-series business data, including: date and timestamp, total number of outbound calls on the day, date attribute tags, actual consumption of virtual numbers on the historical day, call volume during morning and evening shifts, and total number of regional logistics orders; wherein, the date attribute tags include: weekdays, weekends, statutory holidays, e-commerce promotions, and holidays when local logistics are suspended. The training dataset construction submodule 4023 is used to extract a set of business feature samples based on historical time-series business data for a preset period as the training dataset. The training submodule 4024 is used to train the Prophet time series prediction model based on the input dataset. It uses a gradient iterative approach to train the Prophet time series prediction model, learns the relationship between business characteristics and total outbound call count and time-segmented number consumption rate, and obtains the trained Prophet time series prediction model.

[0043] The prediction module 403 is used to estimate the total number of outbound calls based on the date attributes and order size of the day using the trained Prophet time series prediction model, and then combine it with the average daily call volume of a single number to obtain the total number of virtual numbers required for the day. The number pool monitoring module 404 is used to build a dedicated virtual number pool for logistics and to monitor the status of numbers in the dedicated virtual number pool in real time, including: monitoring the number of numbers used, the number of numbers remaining available, and the idle time of numbers in the number pool; comparing the predicted total number of virtual numbers needed for the day with the current stock, and monitoring the number of remaining numbers in real time; when it is detected that the number of available virtual numbers has dropped to the warning threshold or below, and the call volume during the morning and evening shifts is predicted to accelerate the number consumption rate in the subsequent periods, a number application polling request is sent to the operator interface to replenish available numbers to the number pool in batches; The allocation module 405 is used to retrieve the cumulative number of calls made by each number in the number pool for the day, and uses the daily call limit of a single virtual number as the risk control constraint standard to filter idle virtual numbers with the lowest daily call frequency and the lowest load pressure, and allocates them to the salesperson terminal that is currently initiating the call according to the load balancing strategy; after the allocation is completed, the call count of the current number is updated in real time until the call count of the current number reaches the daily call limit, at which point the current number will no longer be allocated to the salesperson terminal; The data storage module 406 is used to automatically capture all information of each outbound call after it is successfully established, and build a standardized related data table to complete the solidified storage. The standardized associated data table stores the following fields: the currently used virtual number, the customer's called mobile phone number, the salesperson's unique ID, the logistics tracking number corresponding to this call, and the time the call was initiated. The standardized associated data table stores records sequentially by call time, forming a traceable data archive for each outbound call. It also establishes a temporary binding relationship between the virtual number and the corresponding salesperson and waybill. The binding record is retained after the call ends, and the temporary binding is automatically released after a preset time if there are no associated calls, thus realizing dynamic unbinding management.

[0044] The intelligent callback module 407 is used to trigger the AI ​​routing retrieval program when a customer calls back a virtual number that is bound to their mobile phone number. The program uses the virtual caller ID as the search keyword, traverses the standardized associated data table, filters out the latest call record under that virtual number, extracts the corresponding salesperson ID from the record, and automatically transfers the customer callback call to the salesperson's terminal device corresponding to that ID. At the same time, a pop-up window automatically displays the logistics waybill details bound to this call on the salesperson's operation interface, so that the salesperson can directly contact the customer for consultation based on the waybill information and complete the closed-loop business communication.

[0045] above Figure 4The virtual number intelligent scheduling and callback association device in the embodiments of the present invention will be described in detail from the perspective of modular functional entities. The electronic device in the embodiments of the present invention will be described in detail from the perspective of hardware processing.

[0046] Figure 5 This is a schematic diagram of the structure of an electronic device 700 provided in an embodiment of the present invention. The electronic device 700 can vary significantly due to different configurations or performance characteristics. It may include one or more central processing units (CPUs) 710 (e.g., one or more processors) and a memory 720, and one or more storage media 730 (e.g., one or more mass storage devices) for storing application programs 733 or data 732. The memory 720 and storage media 730 can be temporary or persistent storage. The program stored in the storage media 730 may include one or more modules (not shown in the diagram), each module including a series of instruction operations on the electronic device 700. Furthermore, the processor 710 may be configured to communicate with the storage media 730 and execute the series of instruction operations in the storage media 730 on the electronic device 700.

[0047] Electronic device 700 may also include one or more power supplies 740, one or more wired or wireless network interfaces 750, one or more input / output interfaces 750, and / or one or more operating systems 731, such as Windows Server, Mac OS X, Unix, Linux, FreeBSD, etc. Those skilled in the art will understand that... Figure 5 The illustrated electronic device structure does not constitute a limitation on electronic devices and may include more or fewer components than illustrated, or combine certain components, or have different component arrangements.

[0048] The present invention also provides a computer-readable storage medium, which can be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium, wherein the computer-readable storage medium stores instructions that, when executed on a computer, cause the computer to perform the steps of the virtual number intelligent scheduling and callback association method.

[0049] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the system, device, or unit described above can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0050] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0051] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention 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 the present invention.

Claims

1. A method for intelligent scheduling and callback association of virtual numbers, characterized in that, include: Construct a Prophet time series forecasting model; Construct a training dataset and use it to train the Prophet time series prediction model to obtain the trained Prophet time series prediction model. Using the trained Prophet time series prediction model, the total number of outbound calls is estimated based on the date attribute and the order size of the day. Then, combined with the average daily call volume of a single number, the total number of virtual numbers required for the day is obtained.

2. The method for intelligent scheduling and callback association of virtual numbers according to claim 1, characterized in that, The construction of the training dataset includes: Collect historical time-series business data, and preprocess the collected historical time-series business data to obtain preprocessed historical time-series business data. Business features are extracted based on preprocessed historical time-series business data, including: date and timestamp, total number of outbound calls on the day, date attribute tags, actual number of virtual numbers consumed on the day, call volume during morning and evening shifts, and total number of regional logistics orders; wherein, the date attribute tags include: weekdays, weekends, statutory holidays, e-commerce promotions, and holidays when local logistics are suspended. The training dataset is a set of business feature samples extracted from historical time-series business data within a preset time period.

3. The method for intelligent scheduling and callback association of virtual numbers according to claim 1, characterized in that, The constructed training dataset is used to train the Prophet time series prediction model, including: The training dataset is input into the Prophet time series prediction model. The Prophet time series prediction model is trained using a gradient iterative method to learn the relationship between business characteristics and total outbound call count and time-segmented number consumption rate, thus obtaining the trained Prophet time series prediction model.

4. The method for intelligent scheduling and callback association of virtual numbers according to claim 1, characterized in that, The method also includes: constructing a logistics-specific virtual number pool and monitoring the status of numbers in the logistics-specific virtual number pool in real time, including: monitoring the number of numbers used, the number of numbers remaining available, and the idle time in the number pool; The system compares the predicted total number of virtual numbers needed for the day with the current stock, and monitors the remaining number of numbers in real time. When the number of available virtual numbers drops to or below the warning threshold, and the call volume during morning and evening shifts is predicted to increase the rate of number consumption in subsequent periods, a number application polling request is sent to the operator interface to replenish the number pool in batches.

5. The virtual number intelligent scheduling and callback association method according to claim 4, characterized in that, The method further includes: retrieving the cumulative number of calls made by each number in the number pool on the same day, and using the daily call limit of a single virtual number as the risk control constraint standard, filtering out idle virtual numbers with the lowest daily call frequency and the lowest load pressure, and allocating them to the salesperson terminal that is currently initiating the call according to the load balancing strategy; after the allocation is completed, updating the call count of the current number in real time until the call count of the current number reaches the daily call limit, at which point the current number will no longer be allocated to the salesperson terminal.

6. The method for intelligent scheduling and callback association of virtual numbers according to claim 1, characterized in that, The method also includes: after each outbound call is successfully established, automatically capturing all information of the call and constructing a standardized association data table for solidified storage; The standardized associated data table stores the following fields: the currently used virtual number, the customer's called mobile phone number, the salesperson's unique ID, the logistics tracking number corresponding to this call, and the time the call was initiated. The standardized associated data table stores records sequentially by call time, forming a traceable data archive for each outbound call. It also establishes a temporary binding relationship between the virtual number and the corresponding salesperson and waybill. The binding record is retained after the call ends, and the temporary binding is automatically released after a preset time if there are no associated calls, thus realizing dynamic unbinding management.

7. The method for intelligent scheduling and callback association of virtual numbers according to claim 6, characterized in that, The method further includes: when a customer calls back a virtual number that is bound to the same number using their mobile phone number, triggering an AI routing retrieval program; The program uses the virtual caller ID as the search keyword, traverses the standardized associated data table, filters out the latest call record under that virtual number, extracts the corresponding salesperson ID from the record, and automatically transfers the customer callback call to the salesperson's terminal device corresponding to that ID. At the same time, a pop-up window automatically displays the logistics waybill details bound to this call on the salesperson's operation interface, so that the salesperson can directly contact the customer for consultation based on the waybill information and complete the closed-loop business communication.

8. A virtual number intelligent dispatch and callback association device, characterized in that, include: The model building module is used to build Prophet time series prediction models; The training module is used to build a training dataset and use the built training dataset to train the Prophet time series prediction model to obtain the trained Prophet time series prediction model. The prediction module uses the trained Prophet time series prediction model to estimate the total number of outbound calls based on the date attributes and the order size of the day. It then combines this with the average daily call volume of a single number to obtain the total number of virtual numbers required for the day.

9. An electronic device comprising a memory and at least one processor, wherein the memory stores instructions; The at least one processor invokes the instructions in the memory to cause the electronic device to perform the steps of the virtual number intelligent scheduling and callback association method as described in any one of claims 1-7.

10. A computer-readable storage medium storing instructions thereon, characterized in that, When the instruction is executed by the processor, it implements each step of the virtual number intelligent scheduling and callback association method as described in any one of claims 1-7.

Images