Agent offline operation method and device based on artificial intelligence, equipment and medium

Abstract

The invention relates to the technical field of artificial intelligence development and the technical field of internet, in particular to an artificial intelligence-based agent offline operation method, device and equipment and a medium, and the artificial intelligence-based agent offline operation method comprises the following steps: obtaining task information corresponding to a plurality of to-be-processed tasks; according to the task information and a trained machine learning model, obtaining a processing probability of the seat end for each to-be-processed task; if the network between the server side and the seat side is connected, obtaining a preset number of to-be-processed tasks according to the processing probability from large to small, and caching the preset number of to-be-processed tasks to a local database of the seat side; and if the network between the server side and the seat side is disconnected, the seat side obtains the task to be processed from the local database so as to perform offline operation. According to the method, the working efficiency of the seat end is ensured, and the stability and reliability of the operation service selling system are also improved.

Description

Technical Field

[0001] The present invention relates to the fields of artificial intelligence development technology and Internet technology, and in particular to an agent offline operation method, device, equipment and medium based on artificial intelligence. Background Art

[0002] In financial insurance companies, agents are set up to conduct business exchanges with customers. The current agent service and sales operation system usually adopts the Browser / Server architecture. In this architecture mode, agents interact with the server through the browser to complete the agent tasks, which is extremely dependent on the stability of the network. Due to practical reasons, network instability is inevitable. When the network is unstable or even disconnected, it will interrupt the normal work process of the agents, seriously affecting the stability and reliability of the operation and sales system and the efficiency of the agent operation, which will extend the agent response time and reduce the customer experience. Summary of the invention

[0003] The present invention provides an artificial intelligence-based agent offline operation method, device, computer equipment and readable storage medium to solve the technical problems of low efficiency of agent operation and insufficient stability and reliability of the operation service system under unstable network conditions.

[0004] In a first aspect, an agent offline operation method based on artificial intelligence is provided, and the agent offline operation method based on artificial intelligence includes:

[0005] Get task information corresponding to multiple pending tasks;

[0006] Obtaining the processing probability of the agent terminal for each of the tasks to be processed according to the task information and the trained machine learning model;

[0007] If the network between the server end and the agent end is connected, a preset number of the tasks to be processed are obtained from large to small according to the processing probability, and the preset number of the tasks to be processed are cached in the local database of the agent end.

[0008] In a second aspect, an agent offline operation device based on artificial intelligence is provided, and the agent offline operation device based on artificial intelligence includes:

[0009] A task information acquisition module is used to obtain task information corresponding to multiple tasks to be processed;

[0010] A task prediction module, used to obtain the processing probability of each of the pending tasks by the agent terminal according to the task information and the trained machine learning model;

[0011] A task cache module, for obtaining a preset number of the tasks to be processed in order from large to small according to the processing probability when the network between the server end and the agent end is connected, and caching the preset number of the tasks to be processed in a local database of the agent end;

[0012] The offline operation module is used to enable the agent end to obtain the to-be-processed task from the local database to perform offline operation if the network between the server end and the agent end is disconnected.

[0013] In a third aspect, 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 implements the steps of the above-mentioned artificial intelligence-based agent offline operation method when executing the computer program.

[0014] In a fourth aspect, a computer-readable storage medium is provided, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned artificial intelligence-based agent offline operation method are implemented.

[0015] In the scheme implemented by the above-mentioned artificial intelligence-based agent offline operation method, device, computer equipment and readable storage medium, the processing probability of the agent end for each of the pending tasks can be obtained through the trained machine learning model; when the network is connected, a preset number of pending tasks are obtained from large to small according to the processing probability of each pending task and cached in the local database of the agent end; so that when the network is disconnected, the agent end can obtain the pending tasks from the local database to perform offline operations. In the present invention, in order to solve the technical problem that the agent end cannot operate normally when the network is unstable, the processing probability of all pending tasks is first obtained, and then a number of pending tasks are cached in the local database in advance according to the processing probability, so that when the network is disconnected, the agent end can perform operations through the pending tasks in the local database. Since the pending tasks in the local database are tasks with a high probability of being processed by the agent end, the agent end can also work according to the normal process and operation habits when working offline, which ensures the work efficiency of the agent end and improves the stability and reliability of the operation service system. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0017] Figure 1The present invention is a schematic diagram of an application environment of an agent offline operation method based on artificial intelligence provided by an embodiment of the present invention.

[0018] Figure 2 It is a flowchart of a first implementation of an agent offline operation method based on artificial intelligence provided by an embodiment of the present invention.

[0019] Figure 3 yes Figure 2 FIG. 1 is a flow chart of a first specific implementation method of step S30.

[0020] Figure 4 yes Figure 2 FIG. 1 is a flow chart of a second specific implementation manner of step S30 in FIG.

[0021] Figure 5 yes Figure 2 FIG. 5 is a flow chart of a third specific implementation manner of step S30 in FIG.

[0022] Figure 6 It is a flowchart of a second implementation of an agent offline operation method based on artificial intelligence provided in one embodiment of the present invention.

[0023] Figure 7 It is a flowchart of a third implementation of an agent offline operation method based on artificial intelligence provided by an embodiment of the present invention.

[0024] Figure 8 It is a schematic diagram of an agent offline operation device based on artificial intelligence in one embodiment of the present invention.

[0025] Fig. 9 It is a structural diagram of a computer device in one embodiment of the present invention. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0027] The artificial intelligence-based agent offline operation method provided in the embodiment of the present invention can be applied in the following aspects: Figure 1 In the application environment, Figure 1 A schematic diagram of an application environment of an agent offline operation method based on artificial intelligence provided by an embodiment of the present invention, such as Figure 1As shown, in this application environment, an agent end 100 and a server end 200 are included. The agent end 100 is a terminal device used by the agent staff, and the server end 200 obtains the agent work tasks through information mining or user input. The agent end 100 and the server end 200 jointly execute the agent offline operation method based on artificial intelligence. It should be noted that the agent end 100 and the server end 200 can be smart phones, tablet computers, laptop computers, desktop computers, etc., but are not limited to these. The agent end 100 and the server end 200 can be connected via Bluetooth, USB (Universal Serial Bus) or other communication connection methods, and the present invention is not limited here.

[0028] Figure 2 A flowchart of a first implementation of an agent offline operation method based on artificial intelligence provided by an embodiment of the present invention is shown in FIG. Figure 2 As shown, the artificial intelligence-based agent offline operation method may specifically include the following steps:

[0029] S10: Obtain task information corresponding to multiple tasks to be processed.

[0030] S20, obtaining the processing probability of the agent terminal for each of the tasks to be processed according to the task information and the trained machine learning model;

[0031] For different pending tasks, the agent has different processing tendencies based on work experience and task importance. The agent's tendency to process different pending tasks is materialized as probability based on task information and trained machine learning models. For example, it is predicted that the processing probability of task A is 80%, the processing probability of task B is 40%, and the processing probability of task C is 60%. Among them, task A may be insurance business, task B may be loan business, and task C may be insurance claims business.

[0032] S30, if the network between the server end and the agent end is connected, obtaining a preset number of the tasks to be processed from large to small according to the processing probability, and caching the preset number of the tasks to be processed in the local database of the agent end;

[0033] When the network between the server and the agent is connected, a preset number of tasks to be processed are obtained from large to small according to the processing probability, and then the preset number of tasks to be processed are cached in the local database of the agent. For example, if the processing probability of task A is 80%, the processing probability of task B is 40%, and the processing probability of task C is 60%, the above tasks are sorted from large to small according to the processing probability: ACB. If the preset number is 2, tasks A and C are obtained in order and cached in the local database of the agent. The local database uses localStorage or IndexedDB storage.

[0034] S40: If the network between the server and the agent is disconnected, the agent acquires the pending task from the local database to perform offline operation.

[0035] When the network is disconnected, the agent can perform offline operations through the pending tasks in the local database. For example, the pending tasks in the local database are Task A and Task C, which are more likely to be processed by the agent. When the network is disconnected, the agent can obtain the task data of Task A and Task C from the local database. Since Task A and Task C are the pending tasks with a higher probability of being processed by the agent, when the agent performs offline operations through the local database, not only will the work efficiency not be affected by the network disconnection, but the work rhythm of the agent can also be guaranteed, which improves the stability and reliability of the operation service system as a whole.

[0036] In addition, it is periodically detected whether the network between the server and the agent is connected; if the network is connected, step S30 is executed, and if the network is disconnected, step S40 is executed. The judgment basis for network connectivity is that the network status between the server and the agent supports the cache requirement of "caching the preset number of pending tasks in the local database of the agent" in step S30.

[0037] In some implementations, for some tasks that require direct communication with customers, the agent can contact the customer through other means such as mobile communication mode through the customer contact information in the pending tasks in the local database.

[0038] Figure 3 yes Figure 2 The first specific implementation flow diagram of step S30 is as follows: Figure 3 As shown, in some implementations, step S30 may include:

[0039] S31, when a preset number of the tasks to be processed are obtained from large to small according to the processing probability, corresponding signature information of the tasks to be processed is simultaneously obtained;

[0040] S32: When caching the preset number of tasks to be processed into the local database of the agent terminal, verifying the signature information.

[0041] In the above steps, each task to be processed corresponds to signature information. When a preset number of tasks to be processed are obtained, the signature information of the preset number of tasks to be processed are obtained simultaneously; when the tasks to be processed are cached in the local database of the agent end, the signature information is verified.

[0042] In a specific implementation, a public key and a private key are generated for the agent end through an encryption algorithm, and the signature information is the signature information obtained by digitally signing the task to be processed according to the private key of the agent end; the signature information is verified by the public key of the agent end, and after the verification, the task to be processed is cached in the local database of the agent end; the encryption algorithm of the digital signature is an asymmetric encryption algorithm, including but not limited to the RSA algorithm, the national encryption SM2 algorithm and other algorithms.

[0043] In another specific implementation, there are multiple agent terminals, each of which corresponds to a different local database. Different public keys and private keys are generated for each agent terminal. The pending tasks are digitally signed by different private keys, and the pending tasks are cached after signature verification by the corresponding public keys. While caching the pending tasks, it is also ensured that each agent terminal can only obtain the corresponding pending tasks, ensuring data security while achieving accurate allocation of pending tasks. For example, agent terminal A handles loan business, and agent terminal B handles insurance business. Public key A and private key A are generated for agent terminal A, and public key B and private key B are generated for agent terminal B through an asymmetric encryption algorithm. The pending tasks include Task 1, Task 2, Task 3, and Task 4, where Task 1 and Task 2 are loan business, and Task 3 and Task 4 are insurance business. Task 1 and Task 2 are digitally signed by private key A, and Task 3 and Task 4 are digitally signed by private key B. When acquiring Task 1, Task 2, Task 3, and Task 4, their signature information is acquired synchronously. When caching them in the local database, Agent A verifies the signature information of the above tasks through Public Key A. Task 1 and Task 2 are cached in Agent A's local database after passing the signature verification, while Task 3 and Task 4 cannot pass the signature verification. Similarly, Agent B's local database can only cache Task 3 and Task 4, but not Task 1 and Task 2. This ensures the accurate allocation of tasks to be processed and the security of task data of different agents.

[0044] Figure 4 yes Figure 2 The second specific implementation flow diagram of step S30 is as follows: Figure 4 As shown, in some implementations, step S30 may include:

[0045] S33, before caching the preset number of tasks to be processed into the local database of the agent terminal, encrypting the preset number of tasks to be processed;

[0046] S34: before the agent terminal obtains the to-be-processed task from the local database to perform offline operations, decrypt the task obtained from the local database.

[0047] In the above steps, before caching the preset number of pending tasks in the local database of the agent, the preset number of pending tasks needs to be encrypted. Before the agent obtains the pending tasks from the local database for offline operations, the tasks obtained from the local database need to be decrypted in advance.

[0048] In a specific embodiment, the preset number of pending tasks are encrypted by a symmetric algorithm and then cached in a local database. Before the agent obtains the pending tasks from the local database to perform offline operations, the pending tasks in the local database are first decrypted by a symmetric algorithm to ensure the data security of the pending tasks in the local database. The symmetric algorithm can be one or more of the CBC128 algorithm, the CBC256 algorithm, and the SM4 algorithm. The above-mentioned CBC refers to Cipher Block Chaining. SM4 is a symmetric cryptographic algorithm.

[0049] Figure 5 yes Figure 2 A third specific implementation flow diagram of step S30 is shown in FIG. Figure 5 As shown, in some implementations, step S30 may include:

[0050] S35, if the network between the server and the agent is connected, the agent obtains the real-time task from the server to perform online operation;

[0051] S36: If the local database includes the real-time tasks, the real-time tasks are deleted from the local database, and the to-be-processed tasks having the same number as the real-time tasks are cached from the server side to the local database.

[0052] In the above steps, if the network between the server side and the agent side is connected, the agent side can actively obtain real-time tasks from the server side online. The real-time tasks may not belong to the tasks cached in the local database, or they may belong to the tasks cached in the local database. When the real-time task does not belong to the tasks cached in the local database, the agent side's behavior of actively obtaining the real-time task from the server side online has no effect on the artificial intelligence-based agent offline operation method in the present invention. When the real-time task belongs to the task cached in the local database, there are pending tasks in the local database that can no longer be processed by the agent side when offline. Therefore, it is necessary to delete the pending tasks in the local database that are the same as the real-time tasks, and cache the same number of pending tasks from the server side to the local database.

[0053] Figure 6 A flowchart of a second implementation of an agent offline operation method based on artificial intelligence provided by an embodiment of the present invention is shown in FIG. Figure 6 As shown, after step S40, the artificial intelligence-based agent offline operation method further includes the following steps:

[0054] S50, storing the job data generated by the offline job in the local database;

[0055] S60: When the network between the server and the agent is reconnected, the operation data in the local database is sent to the server.

[0056] In the above steps, the job data generated after the agent performs offline operations is first stored in the local database. When the network is reconnected, the job data in the local database is sent to the server. For example, the task of the agent to perform offline operations is to conduct insurance claims for a certain customer, and the job data is the insurance claims results. When the network is reconnected, the insurance claims results are sent to the server, which can maximize the work efficiency of the agent.

[0057] Figure 7The flowchart of the third implementation method of the offline agent operation method based on artificial intelligence provided by one embodiment of the present invention, the machine learning model includes an XGBoost model. The XGBoost algorithm is a prediction algorithm that includes classification trees and regression trees, and mainly uses CART regression trees to form weak classifiers. At the same time, the idea of ​​the XGBoost algorithm is to continuously add trees and continuously perform feature splitting to grow a tree. The essence of adding a tree each time is to learn a new function to fit the residual of the previous prediction. When the training is completed and k trees are obtained, to predict the score of a sample, it is necessary to fall into a corresponding leaf node in each tree based on the characteristics of the sample. Each leaf node corresponds to a score. Finally, you only need to add up the scores corresponding to each tree to get the predicted value of the sample. Figure 7 As shown, the artificial intelligence-based agent offline operation method also includes the following steps:

[0058] S71, obtaining historical task data; the historical task data includes historical customer basic information, historical customer consumption information, historical customer contact information, historical task status information and historical task appointment information. Among them, the historical customer basic information includes the historical customer's age, occupation, marital status, vehicle information and other basic information; the historical customer consumption information includes the historical customer's historical consumption data, etc.; the historical customer contact information includes the number of contacts between the agent and the historical customer, the contact duration, etc.; the historical task status information includes whether the task status is a normal queue status or an expedited status, etc.; the historical task appointment information includes the historical customer's appointment time period, the appointment agent, etc.

[0059] The historical task data used to train the XGBoost model are all data that influence the agent's tendency to handle the historical task.

[0060] S72: Clean the historical task data.

[0061] Data cleaning includes deduplication, missing value supplementation, one-hot encoding, and normalization. For example, deleting duplicate data, supplementing missing data, one-hot encoding the product numbers purchased by customers, and normalizing the total amount of products purchased by customers.

[0062] S73. Perform dimensionality reduction processing on the cleaned historical task data through a principal component analysis method to obtain feature data, and divide the feature data into a training set and a test set.

[0063] The linear discriminant analysis method can be used to map all high-dimensional feature vectors in the historical task data into a low-dimensional space for dimensionality reduction. This operation maximizes the separability of the feature data, while also improving the classification efficiency of the model and enhancing its real-time performance in practical applications.

[0064] S74. Train the XGBoost model using the training set.

[0065] The above steps may include first adjusting the tree structure parameters to obtain a baseline, then adjusting the random sampling parameters and regularization parameters to improve the generalization ability of the model and reduce the risk of overfitting, and finally reducing the learning rate to obtain the optimal number of trees.

[0066] S75. Use the test set to evaluate the performance of the XGBoost model, and train the XGBoost model again to obtain a trained XGBoost model.

[0067] Use the test set data to evaluate the performance indicators of the model, including accuracy, precision, and recall, etc. Based on the performance evaluation model, whether parameters need to be adjusted and trained again.

[0068] In some embodiments, the task information in step S10 may include:

[0069] Basic customer information, customer consumption information, customer contact information, task status information, and task appointment information. Among them, basic customer information includes basic information such as the customer's age, occupation, marital status, and vehicle information; customer consumption information includes the customer's historical consumption data; customer contact information includes the number of contacts between the agent and the customer, the duration of the contact, etc.; task status information includes whether the task is in a normal queue or an expedited state, etc.; task appointment information includes the customer's appointment time period, the agent for the appointment, etc.

[0070] The artificial intelligence-based agent offline operation method provided in the embodiment of the present invention can be constructed based on artificial intelligence, and relevant data can be acquired and processed based on artificial intelligence technology to realize unattended artificial intelligence-based agent offline operation. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results. Basic artificial intelligence technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. Artificial intelligence software technology mainly includes computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0071] In one embodiment, an agent offline operation device based on artificial intelligence is provided, and the agent offline operation device based on artificial intelligence corresponds to the agent offline operation method based on artificial intelligence in the above embodiment. Figure 8 As shown, the artificial intelligence-based agent offline operation device includes:

[0072] The task information acquisition module 101 is used to acquire task information corresponding to a plurality of tasks to be processed;

[0073] The task prediction module 102 is used to obtain the processing probability of each of the tasks to be processed by the agent terminal according to the task information and the trained machine learning model;

[0074] The task cache module 103 is used to obtain a preset number of the tasks to be processed in order from large to small according to the processing probability when the network between the server end and the agent end is connected, and cache the preset number of the tasks to be processed in the local database of the agent end;

[0075] The offline operation module 104 is used to enable the agent end to obtain the to-be-processed task from the local database to perform offline operation if the network between the server end and the agent end is disconnected.

[0076] Specifically, the task cache module includes a signature authentication unit.

[0077] The signature authentication unit is used to synchronously obtain the corresponding signature information of the tasks to be processed when obtaining a preset number of the tasks to be processed from large to small according to the processing probability; and to verify the signature information when caching the preset number of the tasks to be processed into the local database of the agent end.

[0078] Specifically, the task cache module also includes a data encryption unit and a data decryption unit.

[0079] The data encryption unit is used to encrypt the preset number of tasks to be processed before caching the preset number of tasks to be processed in the local database of the agent end.

[0080] The data decryption unit is used to decrypt the task obtained from the local database before the agent obtains the task to be processed from the local database to perform offline operations.

[0081] Specifically, the task cache module also includes an online operation unit and a data synchronization unit.

[0082] The online operation unit is used to enable the agent end to obtain real-time tasks from the server end to perform online operations if the network between the server end and the agent end is connected.

[0083] A data synchronization unit is used to delete the real-time tasks from the local database if the local database includes the real-time tasks, and cache the to-be-processed tasks having the same number as the real-time tasks from the server side to the local database.

[0084] Specifically, the artificial intelligence-based agent offline operation device also includes a data storage module and a data sending module.

[0085] The data storage module is used to store the job data generated by the offline job in the local database.

[0086] The data sending module is used to send the operation data in the local database to the server end when the network between the server end and the agent end is reconnected.

[0087] Specifically, the artificial intelligence-based agent offline operation device also includes a training sample acquisition module, a data cleaning module, a feature engineering module, a model training module and a model evaluation and optimization module.

[0088] The training sample acquisition module is used to acquire historical task data; the historical task data includes historical customer basic information, historical customer consumption information, historical customer contact information, historical task status information and historical task reservation information.

[0089] The data cleaning module is used to clean the historical task data.

[0090] The feature engineering module is used to perform dimensionality reduction processing on the cleaned historical task data through a principal component analysis method to obtain feature data, and divide the feature data into a training set and a test set.

[0091] The model training module is used to train the XGBoost model using the training set.

[0092] The model evaluation and optimization module is used to evaluate the performance of the XGBoost model using the test set, and train the XGBoost model again to obtain a trained XGBoost model.

[0093] For the specific definition of the agent offline operation device based on artificial intelligence, please refer to the definition of the agent offline operation method based on artificial intelligence above, which will not be repeated here. Each module in the above-mentioned agent offline operation device based on artificial intelligence can be implemented in whole or in part through software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0094] Fig. 9 is a schematic diagram of the structure of an embodiment of a computer device according to the present application, such as Fig. 9 As shown, the present application also provides a computer device, including:

[0095] A memory and a processor, wherein computer-readable instructions are stored in the processor, and when the computer-readable instructions are executed by the processor, the processor executes any step of the above-mentioned artificial intelligence-based agent offline operation method.

[0096] The present application also provides a computer-readable storage medium. When the computer-readable instructions are executed by one or more processors, the one or more processors execute any step in the artificial intelligence-based agent offline operation method. It can be understood that the readable storage medium in this embodiment can be a volatile readable storage medium or a non-volatile readable storage medium.

[0097] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. An agent offline operation method based on artificial intelligence, characterized in that: include: Get task information corresponding to multiple pending tasks; Obtaining the processing probability of the agent terminal for each of the tasks to be processed according to the task information and the trained machine learning model; If the network between the server and the agent is connected, a preset number of tasks to be processed are obtained from large to small according to the processing probability, and the preset number of tasks to be processed are cached in a local database of the agent; If the network between the server and the agent is disconnected, the agent acquires the pending task from the local database to perform offline operations.

2. The artificial intelligence-based agent offline operation method according to claim 1, characterized in that: The artificial intelligence-based agent offline operation method further includes: When a preset number of the tasks to be processed are obtained from large to small according to the processing probabilities, corresponding signature information of the tasks to be processed is simultaneously obtained; When the preset number of tasks to be processed are cached in the local database of the agent terminal, the signature information is verified.

3. The artificial intelligence-based agent offline operation method according to claim 1, characterized in that: The artificial intelligence-based agent offline operation method further includes: Before caching the preset number of tasks to be processed into a local database of the agent terminal, encrypting the preset number of tasks to be processed; Before the agent terminal obtains the to-be-processed task from the local database to perform offline operations, the task obtained from the local database is decrypted.

4. The artificial intelligence-based agent offline operation method according to claim 1, characterized in that: The artificial intelligence-based agent offline operation method further includes: If the network between the server and the agent is connected, the agent obtains the real-time task from the server to perform online operation; If the local database includes the real-time tasks, the real-time tasks are deleted from the local database, and the to-be-processed tasks having the same number as the real-time tasks are cached from the server to the local database.

5. The artificial intelligence-based agent offline operation method according to claim 1, characterized in that: After enabling the agent terminal to obtain the to-be-processed task from the local database to perform offline operations, the method further includes: storing the job data generated by the offline job in the local database; When the network between the server end and the agent end is reconnected, the operation data in the local database is sent to the server end.

6. The artificial intelligence-based agent offline operation method according to claim 1, characterized in that: The machine learning model includes an XGBoost model; the training steps of the XGBoost model include: Acquire historical task data, wherein the historical task data includes historical customer basic information, historical customer consumption information, historical customer contact information, historical task status information, and historical task reservation information; Performing data cleaning on the historical task data; Performing dimensionality reduction processing on the cleaned historical task data by a principal component analysis method to obtain feature data, and dividing the feature data into a training set and a test set; Using the training set to train the XGBoost model; The performance of the XGBoost model is evaluated using the test set, and the XGBoost model is trained again to obtain a trained XGBoost model.

7. The artificial intelligence-based agent offline operation method according to claim 1, characterized in that: The task information includes: Customer basic information, customer consumption information, customer contact information, task status information, and task appointment information.

8. An agent offline operation device based on artificial intelligence, characterized in that: include: A task information acquisition module is used to obtain task information corresponding to multiple tasks to be processed; A task prediction module, used to obtain the processing probability of each of the pending tasks by the agent terminal according to the task information and the trained machine learning model; A task cache module, for obtaining a preset number of the tasks to be processed in order from large to small according to the processing probability when the network between the server end and the agent end is connected, and caching the preset number of the tasks to be processed in a local database of the agent end; The offline operation module is used to enable the agent end to obtain the to-be-processed task from the local database to perform offline operation if the network between the server end and the agent end is disconnected.

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, the steps of the artificial intelligence-based agent offline operation method as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the agent offline operation method based on artificial intelligence as described in any one of claims 1 to 7 are implemented.

Images