Intelligent demand structuring method and system and application thereof

Abstract

The invention discloses an intelligent demand structuring method and an intelligent demand structuring system, which are used for generating a structured demand document which can be used for subsequent software development based on business demand description of a natural language. For the simple requirement of a single module, one sentence input is adopted; for complex demands of multiple modules, a business modeling agent is called to carry out modeling understanding, natural language demands are converted into a business concept set, and the business concept set is further decomposed into specific structured demands by a demand structured agent. In the whole process, the user can adjust and refine the business software demand in real time through interaction with the business modeling agent and the demand structuring agent. According to the method, the structured demand document capable of being used for subsequent software development can be completed through natural language description, an efficient, flexible and intelligent new mode is provided for the structured demand of multi-field personalized service software and even software development, and the efficiency and quality of software development can be greatly improved.

Description

Technical Field

[0001] The present invention relates to the field of software development, and in particular to an intelligent structural demand generation method, system and application thereof. The technology can intelligently convert a natural language demand of a sentence into a structured demand document that can be used for subsequent software development through natural language processing, modeling understanding and structural generation, and is suitable for business software development in different fields. Background Art

[0002] With the acceleration of the digitalization process, the demand for business software in various industries continues to grow. However, the traditional software development method has many limitations. It usually relies on business personnel to put forward requirements in the form of natural language, and then developers manually convert these requirements into program code. This process is not only time-consuming and complicated, but also prone to misunderstandings due to manual intervention, resulting in the final developed software not fully matching the actual business needs. In addition, the traditional method is highly dependent on the professional knowledge and experience of developers, which is particularly evident when the requirements change frequently. The traditional development model is difficult to adapt to changes quickly, and the development quality is difficult to guarantee. The progress and quality of the project are also easily affected by changes in technical personnel. Therefore, there is an urgent need for a technology that can intelligently process natural language requirements and automatically generate structured requirements documents to cope with changing business needs and ensure the efficiency and quality of subsequent software development. Summary of the invention

[0003] The present invention aims to provide an intelligent method for generating structured requirements. By combining natural language processing, large language model (LLM), retrieval enhanced generation (RAG) and other technologies, a business modeling agent and a requirement structuring agent are constructed, thereby realizing the intelligent conversion of natural language requirements into structured requirement documents that can be used for subsequent software development. The technical solution of the present invention is as follows: An intelligent demand structuring method, characterized in that the method comprises the following steps: (1) Business software requirements input: Users input business requirements through natural language. The input can be in the form of simple phrases or detailed business background, goals, functional requirements, user roles, operation processes and other related information. These inputs serve as the initial data of the business software structured requirements document, and the system will carry out subsequent processing based on them.

[0004] Simple input: For clear and single expression of requirements, users can describe them in a simple sentence or phrase.

[0005] Complex input: For complex requirements involving multiple details, users can enter a detailed business description, including background, functional requirements, user roles, processes, etc.

[0006] The system will record these inputs as the basis for subsequent business modeling and demand structuring. This process is mainly to allow user needs to be expressed in a simple way.

[0007] (2) Determine the complexity of the business software requirement. The system first determines the complexity of the input business software requirement. If the requirement only involves a single business module and the module is not related to other business modules, the system classifies it as a simple requirement. For simple requirements, the system skips the steps of business modeling agent processing and business concept set generation and directly jumps to step (5) structured requirement document generation. On the contrary, if the requirement involves multiple business modules or there is an interaction requirement with other modules, the system determines it as a complex requirement and executes the subsequent processing steps sequentially. This judgment ensures that requirements of different complexities can be properly processed, avoiding unnecessary processing steps, and ensuring that complex requirements can be fully understood and refined.

[0008] (3) Business modeling agent processing: The system uses LLM (large language model) and RAG (retrieval-augmented generation) technologies to build a business modeling agent to perform deep semantic understanding and modeling of natural language requirements input by users.

[0009] The business modeling agent understands the business domain through modeling and semantic analysis, parses natural language requirements and constructs an orthogonal coordinate system for the knowledge of the target domain to achieve business modeling. The business modeling agent is a system component based on artificial intelligence technology, including large language model (LLM) and retrieval augmented generation (RAG) technology. The main function is to intelligently convert business requirements described in natural language into structured business models. This agent decomposes the business requirements input by users into operational technical requirements through deep semantic analysis, knowledge retrieval and modeling processing, forming a basic business model to support further demand refinement. The main features include: natural language processing, which can parse the natural language input by users and identify the key business elements therein; semantic understanding and modeling, through modeling processing, the business elements are constructed into a structured model to ensure the system's comprehensive understanding and accurate implementation of business requirements; dynamic update, as business requirements change, the agent can dynamically update the business model to ensure that the generated software modules are consistent with the latest business requirements.

[0010] Semantic parsing: The system identifies and extracts key elements from the input and converts them into actionable business elements.

[0011] Knowledge retrieval: The system retrieves best practices, standard processes, and reference models from the knowledge base of related fields to ensure that the generated business model meets industry standards.

[0012] Model-based processing: The system builds a structured business model to convert natural language requirements into a set of business concepts, providing a basis for subsequent processing.

[0013] (4) Business concept set generation: The business concept set generated by the system has a hierarchical structure, which gradually expands from high dimension to low dimension. Each level contains multiple concept sets and defines the relationship between them.

[0014] High-dimensional concept set: defines the core objects and macro-operation processes of the business, and provides an overall understanding of the business logic.

[0015] Low-dimensional concept set: It is refined into specific data flows, user interactions, role permissions, etc. to ensure that the business logic is operational and has a clear structure when implemented in software.

[0016] Relationships and dependencies between concept sets are systematically marked to ensure seamless integration in subsequent development.

[0017] (5) Business concept set adjustment. In this step, when the business concept set generated by the system does not fully meet the user's expected needs, the user can adjust it through natural language. The system allows users to directly enter new instructions or modify existing business descriptions to ensure that the final business concept set is more in line with the user's actual needs. These user inputs will be parsed by the system, and the system will make real-time adjustments to the generated business concept set based on the content of these new inputs. This process ensures that changes in user requirements can be reflected in the business concept set in a timely manner, thereby maintaining the consistency and accuracy of requirements in the subsequent requirements document generation and software development process. This adjustment mechanism improves the flexibility of the system and can effectively respond to changes in requirements in complex business scenarios.

[0018] (6) Repeat steps (3) to (5). Users can enter some new business descriptions or fine-tune existing businesses. The system will call the business modeling agent to parse, understand and process the new descriptions, and update the business concept set until the business concept set meets the user's actual needs.

[0019] (7) Requirements structured agent processing: Using LLM, RAG and requirements structured design, a requirements structured agent is constructed to further process the business concept set, decompose it into specific structured requirements content, and convert it into a structured data flow description. LLM is responsible for generating requirements details, and RAG retrieves the requirements standards for specific business scenarios, ultimately forming a structured requirements document.

[0020] ·Requirement decomposition: The system further decomposes the high-dimensional and low-dimensional concepts in the concept set to generate specific functional requirements, data requirements, user interface requirements, etc.

[0021] Data flow description: The data flow description generated by the system lists in detail the data exchange and processing flow between modules to ensure the accurate implementation of data interaction in software development.

[0022] Document generation: The system generates structured requirement documents, which contain all necessary requirement specifications and functional descriptions, and provide clear guidance for subsequent software generation.

[0023] (8) Generation of structured requirements documents: This includes formal descriptions of business software nodes, processes, and interfaces. As the standard input for business software generation, structured requirements documents define the specific requirements for software implementation and ensure that the generated software is highly consistent with business requirements.

[0024] Document content: The structured requirements document includes node definition, operation process, user interface design, etc. Each element is described in detail and associated with other related elements to ensure the integrity of the logic.

[0025] Document format: Documents are in a standardized format that is easy to parse and use to ensure consistency and correctness in subsequent processes, including but not limited to Json, Excel, etc.

[0026] (9) Adjustment of structured requirements documents. In this step, when the structured requirements documents generated by the system do not fully meet the user's expected requirements, the user can adjust them through natural language. The system allows users to directly enter new instructions or modify existing requirements descriptions. Based on the new information provided by the user, the system will adjust the structured requirements documents in real time to ensure that their content can more accurately reflect the user's actual needs. This adjustment mechanism ensures the flexibility and accuracy of the structured requirements documents, so that the final generated software can better meet the user's business goals and functional requirements. Through such interactions, the system can effectively respond to changes in requirements and ensure that the requirements documents are always consistent with user expectations.

[0027] (10) Repeat steps (7) to (9) to intelligently process the newly added or changed requirement descriptions through interaction with the intelligent agent and automatically update the structured requirement document. Repeat multiple iterations to ensure that the final generated business software structured requirement document meets the actual needs of users and can be used for subsequent software development.

[0028] The present invention also discloses an intelligent demand structuring system, which is characterized by: (1) business modeling processing: this module is used to describe business requirements based on natural language, model the business field through business modeling agents, and convert natural language requirements into structured business expressions. In this process, the system uses large language models (LLM) and retrieval enhanced generation (RAG) technology to deeply analyze the natural language content input by users, identify key business elements therein, and perform modeling processing in combination with knowledge bases in related fields. The modeled business expression form provides a basis for subsequent processing, ensuring that business requirements can be accurately understood and implemented by the system; (2) demand structuring processing: this module receives the business expression results generated from the business modeling processing module, and further refines them into clear and specific structured demand content. The demand structuring agent converts these concepts into actionable demand details through in-depth processing of business concepts. Specifically, the system decomposes the high-level business model into functional requirements, data requirements, user interface requirements, etc., and generates corresponding structured data flow descriptions to ensure that these requirements can be accurately implemented during the software development process; (3) Interactive adjustment: This module allows users to interact with the system through natural language throughout the process. When the structured requirements document generated by the system does not fully meet the user's expected requirements, the user can enter new instructions or modify the existing requirements description through natural language, and the system will make real-time adjustments and updates based on the new information provided by the user. Through this module, users can iterate and optimize business requirements multiple times to ensure that the final structured requirements document is highly consistent with the user's actual needs, thereby providing reliable guidance for subsequent software development. Through the collaborative work of the above modules, the intelligent requirements structuring system can effectively convert business requirements described in natural language into accurate and detailed structured requirements documents, significantly improving the accuracy of business requirements analysis and the efficiency of software development. This system can not only dynamically adjust requirements, but also adapt to personalized requirements in multiple fields, providing an efficient, flexible and intelligent new way for the development of complex business software.

[0029] The present invention also discloses a non-volatile storage medium, characterized in that the non-volatile storage medium includes a stored program, wherein when the program is running, the device where the non-volatile storage medium is located is controlled to execute the above method.

[0030] The present invention also discloses an electronic device, characterized in that it comprises a processor and a memory; the memory stores computer-readable instructions, and the processor is used to execute the computer-readable instructions, wherein the computer-readable instructions execute the above method when they are executed. Beneficial Effects

[0031] The method of the present invention can realize intelligent conversion of a natural language requirement of a sentence into a structured requirement document that can be used for subsequent software development, greatly improving the efficiency and quality of software development. Specifically, it has the following beneficial effects: (1) Improve development efficiency: Through the collaborative work of business modeling agents and demand structuring agents, business requirements described in natural language can be quickly converted into structured business expressions, reducing the time for conversion from "idea-requirement-design" to other links.

[0032] (2) Improve development quality: LLM and RAG technologies are used to ensure accurate understanding of natural language requirements, reducing development risks caused by deviations in requirements understanding. At the same time, structured requirements documents are used to ensure that subsequent software development is highly matched with business requirements.

[0033] (3) Support dynamic adjustment: Users can adjust their requirements through natural language throughout the development process. The system can process these new requirements in real time and generate corresponding software structured requirements document versions, enabling rapid iteration of the software and adapting to changing business needs.

[0034] (4) Wide applicability: The methods, devices, and systems of the present invention are highly versatile and applicable to software development in different business fields. They can be customized according to specific business needs to meet the personalized needs of multiple fields and industries. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 is a flow chart of the intelligent demand structuring method in the present invention; Figure 2 It is a simplified schematic diagram of the overall architecture provided by the present invention; Figure 3 It is a schematic diagram of the use of the intelligent demand structuring method in the present invention. DETAILED DESCRIPTION

[0036] The following further describes the technical solution in the embodiment of the present invention in detail with reference to the intelligent generation of an office management system as an example.

[0037] An office management system is a software tool used to improve the efficiency of an enterprise's internal office work. The system helps enterprises achieve efficient and standardized management by integrating and automating various office affairs, such as document management, meeting arrangements, task allocation, approval processes, employee attendance, etc. An office management system can adapt to enterprises of different sizes, from small teams to large corporate groups, and can optimize internal workflows, reduce human errors, and improve work efficiency through the system.

[0038] In actual applications, the office management needs of each company are different, especially for those companies that deeply integrate business and software (i.e., companies where business is software and software is business), their needs are more complex and unique. Traditional office management system development usually relies on pre-made functional modules, but these modules cannot fully meet customized needs. For companies that require high personalization, all functions need to be customized and developed according to specific business scenarios, which leads to long development cycles, high costs, and difficulty in large-scale replication.

[0039] Taking the construction of an automated office management system by a company as an example, the intelligent demand structuring method of the present invention makes the customized development of the office management system very efficient and simple by using intelligent natural language processing, modeling understanding and structured document generation. Under this method, the development of an office management system only requires one person who understands the business to lead, which greatly reduces the reliance on the business understanding of domain experts and technical personnel, and can better meet the specific needs of the enterprise.

[0040] The specific workflow is as follows: Figure 1 As shown: (1) Business software requirements input In this step, users input their requirements for the office management system into the system through natural language. The input can be in the form of a simple phrase or a detailed description, covering the business background, goals, functional requirements, user roles, operation procedures and other information. These inputs constitute the starting point of the entire business software generation process, and the system will carry out subsequent processing and generation work based on these inputs.

[0041] Simple input, suitable for clear and single expression of requirements: Example: A user may enter a simple requirement description, such as "I need a conference management system." This type of input is suitable for users who have clear requirements for a certain functional module and do not need to involve complex business logic or multiple goals.

[0042] ·Processing method: The system’s natural language processing module will immediately identify “meeting management” as a core business requirement and extract it for subsequent processing. Based on this input, the system will generate a basic meeting management module covering functions such as meeting scheduling, meeting notifications, and meeting minutes.

[0043] Complex input, suitable for comprehensive requirements description including detailed business background, multiple objectives and operation processes: Example: A user may enter a detailed description, such as "Our company needs an office management system that must include meeting management, task allocation, employee attendance, and approval process functions. In addition, the system needs to support permission management for multiple user roles and be able to integrate with the existing ERP system." ·Processing method: The system's natural language processing module analyzes user input sentence by sentence and identifies multiple business elements (such as meeting management, task allocation, employee attendance, approval process, permission management, ERP integration, etc.). The system extracts these key elements and builds a comprehensive business requirements model that covers the detailed requirements of each functional module.

[0044] (2) Complexity judgment. In this step, the system first judges the complexity of the input business software requirements. If the requirement only involves a single business module and the module is not related to other business modules, the system classifies it as a simple requirement. For simple requirements, the system skips the steps of business modeling agent processing and business concept set generation and directly jumps to step (5) structured requirement document generation. On the contrary, if the requirement involves multiple business modules or there are interactive requirements with other modules, the system judges it as a complex requirement and executes the subsequent processing steps sequentially. This judgment ensures that requirements of different complexities can be properly processed, avoiding unnecessary processing steps, while ensuring that complex requirements can be fully understood and refined.

[0045] (3) Business modeling agent processing In this step, the system uses LLM (Large Language Model) and RAG (Retrieval Enhanced Generation) technology to build a business modeling agent to perform deep semantic understanding and modeling of the natural language requirements input by users. This process involves the following key links: Semantic parsing: The system first performs semantic parsing on the natural language requirements input by the user, identifying and extracting key business elements. These elements may include business goals, functional requirements, role definitions, and operation processes described by the user. Through semantic parsing, the system can convert the information in these natural languages ​​into more structured business elements, providing a basis for subsequent modeling processing.

[0046] Example: Assume that the user enters "I need a conference management system that can schedule meetings, send notifications, and record meeting minutes." The system will identify "meeting management", "scheduling meetings", "sending notifications", and "recording meeting minutes" as key business elements.

[0047] Knowledge retrieval, after extracting key business elements, the system will retrieve best practices, standard processes, and reference models from the knowledge base of related fields. This step ensures that the business model generated by the system complies with industry standards and best practices, thereby ensuring its applicability and operability. The results of knowledge retrieval will be combined with the business elements obtained from semantic analysis to provide a broader industry background and reference for modeling processing.

[0048] Example: The system may retrieve best practices for meeting management, including standard meeting scheduling processes, common notification methods, and templates for meeting minutes. These search results will help the system incorporate industry standards when building business models.

[0049] Modeling: After completing semantic analysis and knowledge retrieval, the system will perform modeling to construct the extracted business elements and retrieved reference information into a structured business model. This model is represented in the form of a business concept set, which includes the core objects of the business, operation processes, and role permissions. These business models lay the foundation for the subsequent processing of the system, ensuring that natural language requirements can be accurately converted into system functional implementations.

[0050] Example: For a conference management system, the modeling process may include building a set of business concepts that define concepts such as "meeting scheduling", "notification sending", and "meeting minutes recording", and specify the relationships and operational processes between them.

[0051] Through the processing of business modeling agents, the system can effectively transform users' natural language requirements into structured business models, ensuring that the generated software can accurately reflect user needs and meet industry standards. This process greatly reduces the need for human intervention and improves the accuracy of demand understanding and the efficiency of software development.

[0052] (4) Business concept set generation After the business modeling agent is processed, the system further generates a business concept set. The business concept set is a specific expression of the business model. It gradually expands various aspects of the business through a hierarchical structure to form a comprehensive and detailed business description. This process includes the generation of high-dimensional concept sets and low-dimensional concept sets. Each level contains multiple concept sets, and the system will clearly mark the relationship and dependency between them to ensure seamless integration during the development process.

[0053] High-dimensional concept sets are at the top level of the concept set structure and are mainly used to define the core objects and macro-operation processes of the business. Through high-dimensional concept sets, the system can provide a holistic understanding of the business logic. These concepts usually represent the key elements of the business, such as modules such as "meeting management", "task allocation", and "approval process" in the office management system. These high-dimensional concept sets help the system grasp the main processes of the business in a global scope and ensure the coordination between modules and the integrity of the business logic.

[0054] Example: In an office management system, a high-dimensional concept set may include "meeting scheduling", "task management", "approval process", etc. Each concept represents a core functional module. These modules define business processes at a high level, such as how to create and schedule meetings, how to assign tasks, and how to process approval requests.

[0055] Low-dimensional concept set: Low-dimensional concept set is a further refinement of high-dimensional concept set, involving specific business implementation details. Low-dimensional concept set includes detailed descriptions of data flow, user interaction, role permissions, etc. These concepts ensure that in the software implementation process, each business logic has clear operability and structured execution path. For example, low-dimensional concept set may include the specific data structure of the meeting, the sending logic of the meeting notification, the permission settings of different user roles, etc.

[0056] For example, for the "Conference Management" module, the low-dimensional concept set may include "Conference Data (Conference Time, Location, Participants, etc.)", "Notification Sending (Email, Message Notification, etc.)", "Role Permissions (Administrator, Participant Operation Permissions)", etc. These detailed concept sets define how the conference management function is implemented in the system, including how users interact with the system, how data flows, and how permissions are controlled.

[0057] To ensure the smooth progress of subsequent development, the system will clearly mark the relationships and dependencies between concept sets. This includes data sharing relationships between modules, dependencies between operation processes, and inheritance relationships of role permissions. By clearly defining these relationships, the system can maintain seamless integration between modules during the development process and ensure the correct collaboration and operation of each part of the functions.

[0058] For example, in an office management system, the "Meeting Management" module may depend on the "Task Management" module to assign meeting tasks, or the "Approval Process" module may need to determine which users can perform approval operations based on the "Role Permissions" module. The system marks these dependencies to ensure that during the implementation process, each module can correctly interact with data and function.

[0059] By generating a set of business concepts, the system further transforms the user's business needs from an abstract model into a specific, structured business description. This process not only ensures the clarity and operability of the business logic, but also provides a solid foundation for subsequent code generation and system integration.

[0060] (5) Adjustment of business concept set When the business concept set generated by the system fails to fully meet the user's expected needs, the user can make real-time adjustments to the system through natural language. This process allows the user to directly input new instructions or modify existing business descriptions, so that the system can make real-time adjustments based on the new information provided by the user, ensuring that the business concept set can accurately reflect the user's real needs.

[0061] In specific operations, users only need to input their adjustment requirements through simple natural language. For example, if users want to add a new business process to the existing business concept set or modify the details of an existing process, they can express this requirement in one sentence. The system will automatically parse the natural language instructions entered by the user and instantly adjust the business concept set according to the new requirement information.

[0062] This real-time adjustment mechanism ensures the flexibility and adaptability of the system, and can quickly respond to changes in user needs. For example, if users find that a new approval link needs to be added to a key functional module during project development, they only need to input "add a manager approval link in process A" through natural language. The system will automatically identify and adjust the business concept set, integrate the new approval link into process A, and update the relevant process logic and data flow description to ensure that the system can accurately execute new requirements.

[0063] This real-time adjustment mechanism greatly simplifies the user's operation process. Users can ensure that the generated business model is consistent with their expectations without having to master complex technical knowledge or perform manual configuration. This process not only improves work efficiency, but also reduces development errors caused by changes in requirements or inaccurate descriptions, ensuring that business software can accurately match users' actual business needs.

[0064] (6) Repeat steps (3) to (5). Users can enter some new business descriptions or fine-tune existing businesses. The system will call the business modeling agent to parse, understand and process the new descriptions, and update the business concept set until the business concept set meets the user's actual needs.

[0065] (7) Demand structured agent processing After the business concept set is generated, the system further refines these concept sets into specific structured demand content. This process is completed by the demand structuring agent to ensure that all business requirements can be converted into clear and executable technical specifications. This step includes the following key links: Demand decomposition, in the demand decomposition stage, the system further refines the high-dimensional and low-dimensional concepts in the business concept set to generate specific functional requirements, data requirements, user interface requirements, etc. This process ensures that each high-level business concept can be implemented into a specific system function module and its implementation method is clarified.

[0066] Example: In an office management system, the high-dimensional concept "meeting management" may be decomposed into multiple specific functional requirements, such as "creating a meeting", "sending a meeting notice", "managing meeting participants", "generating meeting minutes", etc. At the same time, the system will also generate data requirements (such as meeting data model) and user interface requirements (such as meeting scheduling interface, notification interface) related to these functions.

[0067] Data flow description, the data flow description generated by the system lists in detail the data exchange and processing flow between modules. These descriptions ensure that data can flow accurately between modules and be processed correctly during the software development process. Data flow description is crucial to the overall architecture design of the system. It defines how each functional module interacts through data to implement business logic.

[0068] For example, in the "Meeting Management" module, the data flow description may specify how data flows from "Creating a Meeting" to "Sending a Notification" to "Generating Meeting Minutes". The system will clearly describe the path and method of data transfer from one module (such as the meeting creation module) to another module (such as the notification sending module) to ensure the consistency and integrity of the data throughout the process.

[0069] Document generation, the structured requirements document generated by the system includes all necessary requirements and functional descriptions. These documents provide clear guidance for subsequent software generation. The structured requirements document not only lists in detail the functions that the system needs to implement, but also contains the specific implementation details of each functional module, such as data model, user interaction process, permission settings, etc.

[0070] Example: For an office management system, the document may include detailed requirements, such as "the meeting management module must support multi-user permission management", "notifications must be sent through multiple channels (such as email, SMS, internal messages)", "meeting minutes need to be automatically generated and archived", etc. These documents provide clear references for developers and systems to ensure that development work can be carried out in accordance with business requirements.

[0071] Through the processing of demand structured agents, the system further transforms abstract business concepts into specific technical requirements, laying the foundation for software generation. The three links of demand decomposition, data flow description and document generation ensure the accuracy of system functions, correctness of data processing, and standardization of development work.

[0072] (8) Generation of structured requirements documents After completing the processing of the demand structured agent, the system organizes the generated content into a standardized structured demand document. This document is the standard input for subsequent software generation, ensuring that business requirements can be accurately converted into software implementation.

[0073] Document content, the structured requirements document is a detailed description of each element of the system, covering the following key aspects: Node definition: The document clearly defines the various functional nodes of the system, which represent the main functional modules in the system. For example, the functional modules such as "meeting arrangement" and "task allocation" in the office management system will be defined as independent nodes.

[0074] Operational process: The document describes in detail the operational process and logical relationship between each node. These processes define how the system connects operations and transfers data between various functional modules. For example, in the "Meeting Arrangement" node, the operational process may involve specific steps from creating a meeting to notifying participants and then recording meeting minutes.

[0075] User interface design: The document also includes the design instructions for the user interface, clarifying the interface layout and user interaction methods for each operation. The system will generate the corresponding user interface based on these design guidelines to ensure the intuitiveness and consistency of user operations.

[0076] These are described in detail and linked to other related elements to ensure that each section in the document is logically rigorous and interrelated, thereby ensuring the overall consistency of the system.

[0077] Document format, structured requirements documents exist in a standardized format, which is easy to parse and use, ensuring consistency and correctness in the subsequent software generation process. These document formats include but are not limited to: Json: It is used to describe data models and operation logic in a structured manner, which facilitates machine parsing and automatic code generation.

[0078] Excel: As an intuitive table format, it is suitable for defining and displaying various configuration data, process steps, user roles and other information.

[0079] By generating structured requirement documents, the system standardizes and structures business requirements, operating procedures, user interfaces, etc., laying a solid foundation for generating high-quality software.

[0080] (9) Adjustment of structured requirements documents When the structured requirements document generated by the system does not fully meet the user's expectations, the user can make real-time adjustments to the document through natural language. The system design allows users to directly enter new instructions or modify existing requirements descriptions. Based on the updated information provided by the user, the system will immediately adjust the document content to ensure that it more accurately reflects the user's actual needs and expectations.

[0081] This adjustment mechanism greatly improves the flexibility and accuracy of structured requirements documents. In actual operation, when users find that some requirements in the generated documents do not fully meet business needs, or the descriptions are not detailed enough, they can simply propose new modification instructions through natural language. For example, users may instruct the system to "add a description of two-factor authentication in the user login module" or "modify the current approval process description to include the supervisor approval link." The system will automatically parse these instructions and update the requirements document to ensure that the document can immediately reflect the latest business needs.

[0082] This interactive mode effectively responds to the dynamic changes in requirements and avoids software development deviations caused by inaccurate or imperfect requirements documents. Through continuous real-time adjustments, the system ensures that the structured requirements documents are always highly consistent with the user's business goals and functional requirements, thereby ensuring that the final generated software is closer to the user's actual application scenarios and improving the efficiency and success rate of software development.

[0083] This design not only simplifies the user's operating process, but also significantly reduces the errors and rework that may be caused by changes in requirements during the development process, thereby improving the execution efficiency and quality of the entire project.

[0084] (10) Repeat steps (7) to (9) to intelligently process the newly added or changed requirement descriptions through interaction with the intelligent agent and automatically update the structured requirement document. Repeat multiple iterations to ensure that the final generated business software structured requirement document meets the actual needs of users and can be used for subsequent software development.

[0085] The simplified schematic diagram of the overall architecture of the present invention is as follows: Figure 2 , explaining the key parts of the intelligent requirements structuring method and their interrelationships.

[0086] Figure 2 The whole process from business requirement input to the generation of final business software requirement structured document is demonstrated through five levels.

[0087] (1) Input layer: On the left, users input business requirements through natural language. These requirements are input into the system’s intelligent processing module, namely the business software agent. This agent is responsible for receiving, converting and confirming user requirements.

[0088] Business software agent: This is the core part of the system, which receives the user's natural language input and performs preliminary semantic analysis. The agent converts the user's requirements into internally processable business models and requirement models, and confirms whether these models accurately reflect the user's intentions.

[0089] (2) Business model layer: After receiving user requirements, the business software agent converts them into content in the business model layer. The business model layer is the abstraction and organization of business logic and high-level concepts.

[0090] Business dimensions: represent the main dimensions involved in the business model, such as different business areas or functional modules.

[0091] Concept set: Contains the core concepts in the business, which are the basis for the system to understand user needs.

[0092] Conceptual relationships: By defining the relationships between concept sets, the system can establish the structure of business logic and ensure coordination and integration between various business function modules.

[0093] At this layer, the system abstracts user needs into high-level business concepts and organizes these concepts to facilitate subsequent processing and demand refinement.

[0094] (3) Requirements model layer: The output of the business model layer is further converted into specific content in the requirements model layer. The requirements model layer is closer to the software implementation level and includes the steps of concretizing the business model into actionable requirements.

[0095] Data, actions, and user roles: These are the core elements of the requirements model, representing the data content that the system needs to process, the operations that users can perform, the roles involved in the system, and how users interact.

[0096] Nodes, processes, permissions, and views: The system specifically defines the behavior of each functional node, the order of business processes, the operating permissions of roles, and the layout of the user interface in the demand model layer. These definitions ensure that the system can execute according to the expected logic, and that the user's operating permissions and interface display can match business requirements.

[0097] The requirement model layer further decomposes the high-level abstraction of the business model into specific and executable requirement items, providing clear guidance for subsequent software generation.

[0098] (4) Requirements structured generation layer: On the right, the output of the requirements model layer is input into the requirements structured generation module, which is responsible for converting the structured requirements model into the specified document.

[0099] Parsing, assembly, and output: The structured generation module first parses the demand model and extracts the functions, data structures, and user interface configurations that need to be implemented. Subsequently, the system automatically assembles the specified document structure and publishes the generated documents.

[0100] (5) Output layer: After verification and release, the final generated structured requirements document can be used as a benchmark for subsequent software development. If it is supplemented by automated development related methods or systems, a one-stop solution from requirements input to software generation can be achieved.

[0101] Examples of the use of agents include Figure 3 In the process of using the agent, the business modeling agent and the demand structuring agent are usually used in combination. The user first inputs the natural language requirements, and the agent parses and converts these requirements into business models and structured requirements documents. Then, the user can adjust the generated content through the feedback link, and the agent will regenerate the corresponding structured requirements document according to the new requirements or adjustments. This process greatly simplifies the complexity of demand understanding and structured generation, shortens the cycle of demand understanding and design, and can maximize the guarantee that the subsequent software development results are highly consistent with the actual business needs of users.

[0102] Figure 3 An example of the use of agents is shown. By combining business modeling agents and requirement structuring agents, efficient transformation of business requirements from initial description to final output is ensured. The following is a detailed description of each part in the figure: (1) User input requirements: A business person is shown on the left. His task is to determine the business objectives and scope of requirements and input relevant information through natural language. The business person only needs to describe the software functions, business processes, user roles, operation processes, etc. he needs, without having a technical background. The input in this process can be a simple phrase or a detailed business background and requirements description.

[0103] (2) Customized Business Software Agent: The middle part shows the core of the system - the customized business software agent. This agent processes and understands the natural language requirements input by users by combining large model (LLM) and RAG (retrieval-augmented generation) technology. The agent performs two main tasks: Business modeling: The intelligent agent converts the requirements described in natural language into structured business models through semantic parsing and knowledge retrieval. These models are abstractions and definitions of business logic, data flow, user interaction and other elements.

[0104] ·Requirement structuring: Based on the business modeling, the agent further transforms the business model into specific requirement documents. These documents describe each functional module, data flow, operation logic, etc. in detail and are stored in a structured format.

[0105] (3) Feedback and adjustment of the agent: The right side shows the process of users providing feedback and adjusting the content generated by the system. After the structured requirements document is generated, users can modify their opinions or add new requirements in natural language based on actual changes in requirements or deficiencies found, and the agent will make real-time adjustments. This process enables the delivery results to be continuously optimized, ultimately achieving a customized solution that is highly consistent with business needs.

[0106] (4) Diversified delivery: The bottom shows the final structured requirements document. Various document formats can be generated according to user needs to meet different user scenarios.

[0107] In summary, through the above process, the development of the office management system only needs business personnel to directly transform these requirements into structured requirement documents, and even generate software, quickly and efficiently generating office management software that meets business needs. When the system needs to be adjusted or expanded, this method can respond quickly and achieve iterative updates of the software through intelligent generation means. This method significantly improves the development efficiency of the office management system, shortens the development cycle, reduces development costs, and ensures the high quality and business applicability of the system.

[0108] By adopting this method to develop an office management system, combined with the powerful processing capabilities of the intelligent body, it is possible to achieve rapid and efficient conversion of business requirements and improve the efficiency and quality of software development. By directly converting natural language into structured requirements documents, it avoids dependence on experts and technicians in specific fields, realizes the decoupling of the system and personnel, and improves the flexibility and maintainability of the office management system.

[0109] The present invention proposes an intelligent method for demand structuring, which constructs business modeling agents and demand structuring agents through large language model (LLM) and retrieval enhanced generation (RAG) technology to convert business requirements into structured document generation. The system achieves an efficient and consistent demand structuring process through the process of "requirement input-business modeling-requirement structuring-iterative update", improves the efficiency of subsequent software development, and finally forms an executable office management system. In addition, the present invention avoids dependence on specific personnel through intelligent processing procedures, and enhances the flexibility and efficiency of system development.

[0110] The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the description in the above embodiments and the specification, which are only used to illustrate the principles of the present invention. Various changes and improvements can be made to the present invention without departing from the spirit and scope of the present invention, and these changes and improvements all fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. An intelligent demand structuring method, characterized in that: The following steps are involved: (1) Business software requirement input: users input business requirements through natural language, which serves as the initial data of the business software structured requirement document; (2) Determine the complexity of the business software requirement. If it is a single business module and has no connection with other business modules, it is considered a simple requirement and jumps to step (5). If not, it is considered a complex requirement and executed sequentially. (3) Business modeling agent processing: Using the large language model (LLM), search enhancement generation (RAG), and business modeling design, we build a business modeling agent that deeply understands and models the business requirements described in natural language and constructs an orthogonal coordinate system for the knowledge in the target domain. (4) Business concept set generation. The business concept set provides an abstract perspective on the overall business and has a hierarchical structure. It is divided from high dimension to low dimension. Each level can contain multiple concept sets and defines the relationship between concept sets. (5) Business concept set adjustment: When the business concept set generated by the system does not fully meet the user's expected needs, the user can adjust it through natural language; the user is allowed to directly enter new instructions or modify the existing business description, and the system will make real-time adjustments based on the new information provided by the user; (6) Repeat steps (3) to (5) until the business concept set meets the user's real needs; (7) Requirements structured agent processing: Using LLM, RAG and requirements structured design, we build a requirements structured agent to further process the business concept set, decompose it into specific structured requirements content, and convert it into a structured data flow description; LLM is responsible for generating requirements details, and RAG retrieves the requirements standards for specific business scenarios, and finally forms a structured requirements document; (8) Generation of structured requirements documents, including formal descriptions of business software nodes, processes, and interfaces; structured requirements documents serve as standard input for business software generation, defining specific requirements for software implementation and ensuring that the generated software is highly consistent with business requirements; (9) Adjustment of structured requirement documents: When the structured requirement documents generated by the system do not fully meet the user's expected requirements, the user can make adjustments through natural language; the user is allowed to directly enter new instructions or modify the existing requirement description, and the system will make real-time adjustments based on the new information provided by the user; (10) Repeat steps (7) to (9) for multiple iterations to ensure that the final generated business software structured requirements document meets the actual needs of users and can be used for subsequent software development.

2. The intelligent demand structuring method according to claim 1, characterized in that: The step 3 further includes the following contents: relevant information is retrieved from the business domain knowledge base through the large language model LLM and combined with retrieval enhancement generation RAG, and the business modeling agent converts natural language requirements into structured business expressions to form a business concept set.

3. The intelligent demand structuring method according to claim 1, characterized in that: The step (3) further includes the following contents: the business modeling agent understands the business domain through modeling and semantic analysis, parses the natural language requirements and constructs an orthogonal coordinate system for the knowledge of the target domain to achieve business modeling; the business modeling agent decomposes the business requirements input by the user into operational technical requirements through deep semantic analysis, knowledge retrieval and modeling processing, and forms a basic business model to support further requirements refinement.

4. The intelligent demand structuring method according to claim 1, characterized in that: The step (4) further includes the following contents: the high-dimensional concept set represents the core objects and macro-operation processes of the business, covering the overall understanding of the business logic, including business goals, key processes, main roles and their interrelationships; the low-dimensional concept set is refined into specific data flows, user interactions, and role permissions, presenting the detailed implementation level of the business through a logically clear structure. The low-dimensional concept set provides specific and operational guidance for subsequent demand refinement and implementation.

5. The intelligent demand structuring method according to claim 1, characterized in that: The step (7) further includes the following contents: the demand structuring agent analyzes the specific model of the business field and converts the natural language into a structured data flow description to generate a structured demand document that is suitable for different business scenarios; the main features of the demand structuring agent include: demand decomposition, the agent can further decompose the high-level concepts in the business model into specific functions and technical requirements, covering business links, including but not limited to data, actions, and permissions; data flow description, generating a detailed data flow description, clarifying the business flow relationship and processing logic between each link, and ensuring the accurate flow of business; document generation, organizing the structured demand content into standardized documents as the basic input for software development. These documents include but are not limited to functional descriptions, data, actions, user roles, business nodes, business flow diagrams, permissions, and user interface design content.

6. An intelligent demand structuring system, the system executing the intelligent demand structuring method according to claim 1, characterized in that: Includes the following modules: (1) Business modeling processing module, which is used to describe business requirements based on natural language, model the business domain through business modeling agents, and convert natural language requirements into modeled business expressions; (2) The demand structuring processing module is used to input the modeled business expression results into the demand structuring agent and further decompose them into clear and specific structured demand contents; (3) The interactive adjustment module is used by users to adjust and refine business software requirements in real time through interaction with the business modeling agent and the requirement structuring agent throughout the entire process, ensuring that the generated structured requirements meet user needs for use in subsequent software development.

7. A non-volatile storage medium, characterized in that: The non-volatile storage medium includes a stored program, wherein when the program is executed, the device where the non-volatile storage medium is located is controlled to execute the method according to any one of claims 1 to 5.

8. An electronic device, characterized in that: It comprises a processor and a memory; the memory stores computer-readable instructions, and the processor is used to execute the computer-readable instructions, wherein the computer-readable instructions execute the method according to any one of claims 1 to 5 when executed.

Images