Customic digital employee system and working method

Abstract

The invention relates to a customs digital employee working method. The method comprises the steps that original information of an external data source is monitored or acquired in real time through a communication interface layer; performing semantic understanding and intention recognition on the original information by using a natural language processing module, and extracting business entity information; providing a preset customs affair knowledge graph module, and determining a matched customs affair task needing to be executed through a decision reasoning module; and based on the selected matched customs affair task, decomposing the matched customs affair task into ordered executable atomic operation instructions through a task planning module, and sending the atomic operation instructions to an external communication object for operation, or based on the atomic operation instructions, performing automatic operation on the customs affair management system through an automatic operation engine. According to the invention, after the natural language processing module extracts the external data, the external data is matched with the preset customs affair knowledge graph module, and finally the matched customs affair task is generated to be manually processed or automatically executed, so that the customs affair operation efficiency is improved.

Description

Technical Field

[0001] This invention relates to the fields of artificial intelligence and enterprise digital technology, and in particular to a method for customs digital employees to work. Background Technology

[0002] The description in this section provides only background information related to the disclosure of this invention and does not constitute prior art.

[0003] Currently, customs procedures in international trading companies are complex, highly specialized, and repetitive, heavily reliant on manual operations. Employees need to handle large volumes of emails, frequently communicate with multiple freight forwarders and customs brokers, and manually operate the customs management system (such as entering data, checking status, and generating reports). This model has many significant pain points. For example, manual handling of emails and communications is time-consuming and cannot be done 24 / 7, resulting in low efficiency. Furthermore, the accuracy requirements for customs data (such as HS codes, amounts, and quantities) are extremely high; manual entry and verification are prone to errors, potentially leading to customs clearance delays, administrative penalties, and other risks, resulting in a high error rate. In particular, email, instant messaging tools, and the customs management system are disconnected, requiring manual information transfer and creating data silos, resulting in poor inter-system collaboration.

[0004] It should be noted that the above description of the technical background is only for the purpose of providing a clear and complete explanation of the technical solutions of the present invention and facilitating understanding by those skilled in the art. It should not be assumed that the above technical solutions are known to those skilled in the art simply because they have been described in the background section of this invention. Summary of the Invention

[0005] The purpose of this invention is to provide a method for customs digital employees to work, which can extract external data by a natural language processing module, match it with a preset customs knowledge graph module, and finally generate matching customs tasks for manual processing or automated execution, thereby improving the efficiency of customs operations.

[0006] To achieve the above objectives, the present invention discloses a method for operating a customs digital employee, the method comprising the following steps: The communication interface layer can be used to monitor or obtain raw information from external data sources in real time. The natural language processing module is used to perform semantic understanding and intent recognition on the raw information to extract business entity information; A preset customs knowledge graph module is provided. The decision reasoning module matches the business entity information with the process nodes in the customs knowledge graph module and judges the status to determine the matching customs task to be executed. Based on the selected matching customs tasks, the task planning module decomposes the matching customs tasks into ordered executable atomic operation instructions, and sends the atomic operation instructions to external communication objects for operation, or based on the atomic operation instructions, the automation operation engine performs automated operation on the customs management system.

[0007] As a further description of the above technical solution, in the step of "listening to or obtaining the original information of the external data source in real time through the communication interface layer", the external data source comes from the mail server, instant messaging tool or the customs management system.

[0008] As a further description of the above technical solution, in the step of "using the natural language processing module to perform semantic understanding and intent recognition on the original information and extract business entity information", the business entity information includes, but is not limited to, customs declaration number, HS code, container pick-up number, vessel name and voyage number, customs declaration status, release time, reason for cancellation, and inspection notice.

[0009] As a further description of the above technical solution, in the step of "providing a preset customs knowledge graph module, and matching and judging the status of the business entity information with the process nodes in the customs knowledge graph module through the decision reasoning module to determine the matching customs task to be executed", the customs knowledge graph module includes, but is not limited to, customs regulations nodes, HS code classification rules, customs declaration process status transition rules, and standard processing actions corresponding to each status. The matching customs task is selected from the customs regulations nodes, the HS code classification rules, the customs declaration process status transition rules, and the standard processing actions corresponding to each status.

[0010] As a further description of the above technical solution, after the step of "sending the atomic operation instruction to an external communication object for operation, or performing automated operation on the customs management system through the automated operation engine based on the atomic operation instruction", the management monitoring center records the entire operation process in full-link logs and provides a manual intervention interface.

[0011] As a further description of the above technical solution, in the step of "recording the entire operation process through the management and monitoring center and providing a manual intervention interface", the management and monitoring center records the trigger time of this operation task, the original email content, the parsing result, and screenshots or logs of each step of the operation.

[0012] As a further description of the above technical solution, in the step of "recording the entire operation process through the management and monitoring center and providing a manual intervention interface", if the update of the entire log recording fails, the management and monitoring center marks the operation as abnormal and notifies the administrator to intervene.

[0013] As a further description of the above technical solution, in the step "based on the atomic operation instructions, perform automated operations on the customs management system through the automated operation engine", the automated operation engine identifies the UI elements of the customs management system and simulates the user's click, input, and selection operations, including automatically logging into the customs management system, automatically extracting data from emails or chat information and filling it into the specified form fields of the customs management system, and automatically performing operations such as querying, submitting, and exporting in the customs management system.

[0014] A customs digital employee work system for the aforementioned customs digital employee work method, wherein the customs digital employee work system comprises: The communication interface layer is configured to receive and send raw information from external data sources; The AI ​​core engine, connected to the communication interface layer, is configured to perform semantic understanding, intent recognition, and task decision-making on the received information. An automated operation engine, connected to the AI ​​core engine, is configured to execute automated operations on the customs management system based on decision commands from the AI ​​core engine. The management and monitoring center is connected to the communication interface layer, the AI ​​core engine, and the automated operation engine. It is used to record the entire operation process in full-link logs and provide a manual intervention interface.

[0015] As a further description of the above technical solution, a natural language processing module is provided for semantic understanding and intent recognition of the original information and for extracting business entity information; A customs knowledge graph module used to store process nodes of customs tasks; A decision reasoning module for matching the business entity information with the customs tasks; A task planning module for decomposing the matching customs task into a series of sequentially executable atomic operation instructions.

[0016] By employing the above technical solutions, the beneficial effects of the present invention are as follows: The customs digital employee work method of this invention improves customs operation efficiency by extracting external data through a natural language processing module, matching it with a preset customs knowledge graph module, and ultimately generating matching customs tasks for manual or automated execution. Specifically, this invention connects and matches raw information from external data sources with process nodes in the preset customs knowledge graph module, minimizing error rates at a low cost. It not only executes preset scripts but also understands natural language communication content, makes decisions, and breaks down barriers between communication (external emails, etc.) and the customs management system (internal software, etc.), achieving end-to-end automation from information reception to system operation.

[0017] To further understand the features and technical content of the present invention, please refer to the following detailed description and drawings of the present invention. However, the drawings provided are for reference and illustration only and are not intended to limit the present invention. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the drawings used in the description of the embodiments or prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 This is an organizational diagram illustrating a customs digital employee work method provided in the embodiments of this specification. Detailed Implementation

[0020] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0021] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can understand the advantages and effects of the present invention from the content disclosed in this specification. The present invention can be implemented or applied through other different specific embodiments, and various details in this specification can also be modified and changed based on different viewpoints and applications without departing from the concept of the present invention. Furthermore, the accompanying drawings of the present invention are for simple illustrative purposes only and are not depictions of actual dimensions; this is stated in advance. The following embodiments will further describe the relevant technical content of the present invention in detail, but the disclosed content is not intended to limit the scope of protection of the present invention.

[0022] It should be understood that while terms such as "first," "second," and "third" may be used in this document to describe various components or signals, these components or signals should not be limited by these terms. These terms are primarily used to distinguish one component from another, or one signal from another. Furthermore, the term "or" as used herein should, as appropriate, include any combination of one or more of the related listed items.

[0023] Please see Figure 1 This embodiment provides a method for operating a customs digital employee, which includes the following steps: The communication interface layer can be used to monitor or obtain raw information from external data sources in real time. The natural language processing module is used to perform semantic understanding and intent recognition on the raw information to extract business entity information; A preset customs knowledge graph module is provided. The decision reasoning module matches the business entity information with the process nodes in the customs knowledge graph module and judges the status to determine the matching customs task to be executed. Based on the selected matching customs tasks, the task planning module decomposes the matching customs tasks into ordered executable atomic operation instructions, and sends the atomic operation instructions to external communication objects for operation, or based on the atomic operation instructions, the automation operation engine performs automated operation on the customs management system.

[0024] Based on the above steps of this invention, it is mainly implemented through the work of customs digital employees. The system includes a communication interface layer configured to receive and send raw information from external data sources; an AI core engine connected to the communication interface layer configured to perform semantic understanding, intent recognition, and task decision-making on the received information; an automated operation engine connected to the AI ​​core engine configured to execute automated operations on the customs management system based on the decision instructions of the AI ​​core engine; and a management monitoring center connected to the communication interface layer, the AI ​​core engine, and the automated operation engine, used to record the entire operation process in a full-link log and provide a manual intervention interface.

[0025] The main function of the communication interface layer is to exchange data with the external world (email server, instant messaging tool, customs management system), and is responsible for sending and receiving emails, reading IM messages, receiving system callbacks. It can also provide APIs for external triggering to receive messages or instructions from the external data source, such as email server, instant messaging tool or customs management system.

[0026] The core AI engine is the most important component, including: a natural language processing module for semantic understanding and intent recognition of the raw information and extraction of business entity information; a customs knowledge graph module for storing process nodes of customs tasks; a decision reasoning module for matching the business entity information and the customs tasks; and a task planning module for decomposing the matched customs tasks into a series of sequentially executable atomic operation instructions.

[0027] The Natural Language Processing (NLP) module can convert unstructured text in emails, chats, and attachments into structured information. For example, it can use entity extraction methods to identify business entity information such as "customs declaration number / container number / vessel name and voyage / time / amount / cargo description." Furthermore, it can use intent recognition methods to determine whether the email is a "release notice," "rejection notice," "request for additional documents," or "booking confirmation." When attachments are present, OCR functionality can be used to scan images or PDF documents, or to parse and split formats such as Excel.

[0028] The customs knowledge graph module stores and retrieves structured knowledge in the customs field, including but not limited to legal provisions, HS code classification rules, process nodes, role relationships, and commonly used declaration standards for companies, serving as a professional basis for decision-making. For example, it can access authoritative sources such as customs announcements, tariff codes, corporate training databases, and historical precedents. Finally, through semantic retrieval, the output of the natural language processing module is mapped to graph nodes. Then, using the decision reasoning module, the structured output of the natural language processing module and the retrieval results are used as input to determine the appropriate operation. Finally, the task planning module determines the matching customs task to be executed.

[0029] In one embodiment, the atomic operation sequence can refer to breaking down a complete matching customs task into six steps: "login → query → modify → save → notify".

[0030] The automated operation engine allows users to simulate user clicks / inputs / selections in the customs management system, or directly call APIs to complete operations (entry, query, export, submission). This can include two forms of automation: UI automation (simulating clicks / inputs), which mainly involves recognizing UI elements (XPath / visual positioning) and then simulating manual operations; and API automation, which directly and automatically executes login and operations, and is the preferred option.

[0031] In other words, in the solution of this invention, the natural language processing module extracts external data and matches it with a preset customs knowledge graph module to ultimately generate matching customs tasks for manual processing or automated execution, thereby improving customs operation efficiency. Specifically, this invention connects and matches the original information from the external data source with the process nodes in the preset customs knowledge graph module, minimizing the error rate at a low cost. It can not only execute preset scripts but also understand natural language communication content, make decisions, and break down the barriers between communication (external emails, etc.) and the customs management system (internal software, etc.), achieving end-to-end automation from information reception to system operation.

[0032] This invention also includes a management and monitoring center. After the step of "sending the atomic operation command to an external communication object for operation, or automating the operation of the customs management system through an automated operation engine based on the atomic operation command," the management and monitoring center records the entire operation process in a full-link log and provides a manual intervention interface. The management and monitoring center records the trigger time of this operation task, the original email content, the parsing result, and screenshots or logs of each step of the operation. If the full-link log update fails, the management and monitoring center marks the operation as abnormal and notifies the administrator to intervene. The management and monitoring center can be configured with workflows, manage digital employee instances, control access permissions, audit logs, issue alarms, and provide a manual intervention entry point. Through its settings, the management and monitoring center undertakes functions such as workflow configuration, digital employee instance management, hierarchical access control, log auditing, issue alarms, and manual intervention. By recording logs containing complete link operation information and combining screenshots and videos of key steps, the center achieves a clear distinction between automated system operations and manual confirmation, while providing auditing and backtracking capabilities based on task numbers or customs declaration numbers.

[0033] The following is a complete demonstration of an operational scenario.

[0034] The first step is the triggering phase. The communication interface layer monitors the designated email address and receives a new email from the cooperating customs broker with the subject "Release Notice for Customs Declaration Form 12345678 of XX Company".

[0035] Then, the system analyzes and understands the email. The natural language processing module extracts key entities from the email body: customs declaration number (12345678), status (released), release time, and remarks. Next, the decision reasoning module matches the "released" status with process nodes in the customs knowledge graph, determining that the email's intent is to "notify customs declaration status update," and the task to be performed is to "update the status of the corresponding customs declaration to 'released' in the customs management system and notify the business personnel." Then, the task planning module generates an operation sequence, as shown in the following example: a. Log in to the customs management system; b. Go to the "Customs Declaration Query" page; c. Enter the customs declaration number "12345678" to query; d. In the query results, find the status field and modify it to "released"; e. Save the changes; f. (Optional) Send a notification to the customs specialist via the company's internal communication software: "Customs declaration 12345678 has been released, and the system has automatically updated." Then, the execution step begins. The automated operation engine is activated and automatically manipulates the interface elements of the customs management system according to the planned sequence to complete the data update operation. The entire process requires no manual intervention.

[0036] Finally, the process includes monitoring and backtracking. The management monitoring center meticulously records the task's trigger time, original email content, parsing results, and screenshots or logs of each step. If the update fails, the management monitoring center automatically marks the task as abnormal and notifies the administrator to intervene.

[0037] The content disclosed above is only a preferred and feasible embodiment of the present invention, and is not intended to limit the scope of the patent application of the present invention. Therefore, all equivalent technical changes made using the contents of the present invention specification and drawings are included in the scope of the patent application of the present invention.

[0038] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0039] Although this application has been described by way of examples, those skilled in the art will know that this application has many modifications and variations without departing from the spirit of this application, and it is intended that the appended embodiments include these modifications and variations without departing from this application.

Claims

1. A method for customs digital employee work, characterized in that, The customs digital employee working method comprises the following steps: Real-time monitoring or obtaining original information of an external data source through a communication interface layer; Using a natural language processing module to perform semantic understanding and intent recognition on the original information and extract business entity information; Providing a preset customs knowledge graph module, matching the business entity information with process nodes in the customs knowledge graph module through a decision reasoning module, and determining the matching customs task to be executed; Based on the selected matching customs task, the matching customs task is decomposed into an ordered executable atomic operation instruction through a task planning module, and the atomic operation instruction is sent to an external communication object for operation, or based on the atomic operation instruction, an automatic operation engine is used to automatically operate the customs management system.

2. The customs digital worker job process of claim 1, wherein: In the step of "real-time monitoring or obtaining original information of an external data source through a communication interface layer", the external data source comes from a mail server, an instant messaging tool, or the customs management system.

3. The customs digital worker job process of claim 1, wherein: In the step of "using a natural language processing module to perform semantic understanding and intent recognition on the original information and extract business entity information", the business entity information includes, but is not limited to, a customs declaration number, an HS code, a container number, a ship name and voyage, a customs declaration status, a release time, a reason for returning a declaration, and an inspection notification.

4. The customs digital worker working method of claim 3, wherein: In the step of "providing a preset customs knowledge graph module, matching the business entity information with process nodes in the customs knowledge graph module through a decision reasoning module, and determining the matching customs task to be executed", the customs knowledge graph module includes, but is not limited to, a customs regulation node, an HS code classification rule, a customs declaration process state transition rule, and a standard processing action corresponding to each state, and the matching customs task is selected from the customs regulation node, the HS code classification rule, the customs declaration process state transition rule, and the standard processing action corresponding to each state.

5. The customs digital worker job process of claim 1, wherein: After the step of "sending the atomic operation instruction to an external communication object for operation, or based on the atomic operation instruction, using an automatic operation engine to automatically operate the customs management system", the entire operation process is recorded through a management monitoring center, and a manual intervention interface is provided.

6. The customs digital worker working method of claim 5, wherein: In the step of "recording the entire operation process through a management monitoring center, and providing a manual intervention interface", the management monitoring center records the trigger time of the operation task, the original mail content, the analysis result, and the screenshot or log of each operation performed.

7. The customs digital worker working method of claim 5, wherein: In the step of "recording the entire operation process through a management monitoring center, and providing a manual intervention interface", if the full-link log recording fails, the management monitoring center marks the operation as abnormal and notifies an administrator to intervene.

8. The customs digital worker job process of claim 1, wherein: In the step of "automating operations on the customs management system through the automated operation engine based on the atomic operation instructions", the UI elements of the customs management system are identified by the automated operation engine, and the clicking, inputting, and selecting operations of the user are simulated, including automatically logging in to the customs management system, automatically extracting data from emails or chat messages and filling in the specified form fields of the customs management system, automatically querying, submitting, exporting, and other operations in the customs management system.

9. A customs digital employee working system for the customs digital employee working method of any one of claims 1-8, characterized in that, The customs digital employee work system comprises: a communication interface layer configured to receive and send raw information from external data sources; an AI core engine connected with the communication interface layer, configured to perform semantic understanding, intent recognition, and task decision on the received information; an automated operation engine connected with the AI core engine, configured to perform automated operations on the customs management system based on the decision instructions of the AI core engine; a management monitoring center connected with the communication interface layer, the AI core engine, and the automated operation engine, for full-link log recording of the entire operation process and providing a manual intervention interface.

10. The customs digital worker system of claim 9, wherein, The AI core engine comprises: a natural language processing module for semantic understanding and intent recognition of the raw information and extraction of business entity information; a customs knowledge graph module for storing flow nodes of customs tasks; a decision reasoning module for matching the business entity information and the customs tasks; a task planning module for decomposing the matched customs tasks into a series of sequentially executable atomic operation instructions.

Images