Business whole-process generation system and method based on enterprise knowledge graph enhancement model
By using a business process generation system based on an enterprise knowledge graph enhancement model, the system analyzes enterprise multimodal data to construct a knowledge graph, and then automatically designs and generates code components. This solves the problem of the disconnect between business cognition and software in traditional development models and SaaS models, and enables high-quality, sustainable software system generation and enterprise-level collaboration.
Patent Information
- Application Number
- CN202610178764.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-02-09
- Publication Date
- 2026-03-17
- Estimated Expiration
- 2046-02-09
AI Technical Summary
In the process of enterprise digital transformation, existing technologies have failed to accurately map the essence of enterprise business in both traditional customized development models and standardized SaaS models. This has led to a serious disconnect between business understanding and software implementation, resulting in "mud balls" of systems or data silos, lacking systematic cognitive methods and safeguard mechanisms.
We adopt a business end-to-end generation system based on an enterprise knowledge graph enhancement model. The system uses a perception and understanding engine to parse multimodal raw data to build a business knowledge graph, a design and planning engine to perform automated design, and a construction and delivery engine to generate code components, ensuring that the generated software system is isomorphic to the enterprise business and can evolve sustainably.
It enables the automatic generation of high-quality application architectures, bridging the cognitive gap between business and technology, ensuring that software systems are aligned with enterprise business, continuously adapting to business evolution, reducing development costs and cycles, breaking down data silos, and building a fully collaborative digital ecosystem.
Smart Images

Figure CN121680791A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of intelligent software code technology, specifically relating to a business process generation system and method based on an enterprise knowledge graph enhancement model. Background Technology
[0002] Currently, enterprises face two main software supply paths during their digital transformation: traditional custom development and standardized SaaS. Traditional custom development, while offering personalized solutions, suffers from inherent flaws: the development process starts with a "feature list" rather than a systematic "business understanding," resulting in a final system that only implements discrete functionalities, failing to accurately map the essence of business needs. Severe "translation losses" occur in the long chain of communication from domain experts to developers. Architectural boundaries are easily breached during development, core models are contaminated, forming a "mud ball" system that loses its ability to evolve with business needs. Standardized SaaS, while offering fast delivery, has fundamental limitations. It employs a pre-defined "general best practice" architecture, forcing enterprises to change their unique business processes to adapt to the software, stifling their differentiated competitiveness. Different SaaS products use incompatible application architectures and data models, leading to the institutionalization of "data silos." Both traditional custom development and standardized SaaS heavily rely on manual translation and conversion of business requirements into technical solutions, a process fraught with bias, resulting in a severe disconnect between business understanding and software implementation. Therefore, there is an urgent need for a business process generation system and method based on an enterprise knowledge graph augmentation model to improve the current situation of a serious disconnect between business cognition and software implementation.
[0003] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of this application, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention
[0004] To provide a basic understanding of some aspects of the disclosed embodiments, a brief summary is given below. This summary is not intended as a general commentary, nor is it intended to identify key / important components or describe the scope of protection of these embodiments, but rather as a prelude to the detailed description that follows.
[0005] This disclosure provides a business process generation system and method based on an enterprise knowledge graph enhancement model, which can improve the current situation of serious disconnect between business cognition and software implementation.
[0006] In some embodiments, a business process generation system based on an enterprise knowledge graph augmentation model includes: The perception and understanding engine is used to receive and use a pre-set industry knowledge enhancement model to analyze the enterprise's multimodal raw data, obtain the analyzed content, and build a business knowledge graph based on the analyzed content. The design and planning engine is used for automated system architecture design based on business knowledge graphs to obtain design blueprints. The build and delivery engine is used to generate multiple code components based on the design blueprint, and to assemble and deploy all the generated code components to obtain the software system.
[0007] In some embodiments, a method for generating a full business process based on an enterprise knowledge graph augmentation model includes: Acquire and utilize a pre-defined industry knowledge enhancement model to analyze the enterprise's multimodal raw data, obtain the analyzed content, and construct a business knowledge graph based on the analyzed content; Automated system architecture design based on business knowledge graphs to obtain design blueprints; Multiple code components are generated based on the design blueprint, and all generated code components are assembled and deployed to obtain the software system.
[0008] The beneficial effects of this invention are as follows: The perception and understanding engine receives and utilizes a pre-defined industry knowledge enhancement model to analyze the enterprise's multimodal raw data, enabling a systematic understanding of the enterprise's operational patterns—that is, obtaining the analyzed content. Based on this content, a machine-readable and reasonable business knowledge graph is constructed, transforming the enterprise's operational patterns into structured business knowledge. The design and planning engine then uses this business knowledge graph for automated system architecture design, converting business understanding into a machine-readable design blueprint, ensuring the generated architecture's business adaptability, technological advancement, and engineering feasibility. Finally, the build and delivery engine generates multiple code components based on the design blueprint, assembling and deploying all generated code components to obtain the software system. This results in a high-quality application architecture that is automatically generated and sustainably evolves based on a systematic understanding of the enterprise's operational patterns, fundamentally bridging the cognitive gap between business and technology and improving the current severe disconnect between business understanding and software implementation.
[0009] The above general description and the description below are exemplary and illustrative only and are not intended to limit this application. Attached Figure Description
[0010] One or more embodiments are illustrated by way of example with reference to the accompanying drawings. These illustrations and drawings do not constitute a limitation on the embodiments. Elements having the same reference numerals in the drawings are shown as similar elements. The drawings are not to be scaled. And wherein: Figure 1 This is a schematic diagram of the structure of the business process generation system based on the enterprise knowledge graph enhancement model provided by the present invention; Figure 2 This is a flowchart of the business process generation method based on the enterprise knowledge graph enhancement model provided by the present invention. Detailed Implementation
[0011] To provide a more detailed understanding of the features and technical content of the embodiments of this disclosure, the implementation of the embodiments of this disclosure will be described in detail below with reference to the accompanying drawings. The accompanying drawings are for illustrative purposes only and are not intended to limit the embodiments of this disclosure. In the following technical description, for ease of explanation, several details are used to provide a full understanding of the disclosed embodiments. However, one or more embodiments may still be implemented without these details. In other cases, well-known structures and devices may be simplified in their depiction to simplify the drawings.
[0012] The terms "first," "second," etc., used in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate for the embodiments of this disclosure described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion.
[0013] Unless otherwise stated, the term "multiple" means two or more.
[0014] In this embodiment of the disclosure, the character " / " indicates that the objects before and after it are in an "or" relationship. For example, A / B means: A or B.
[0015] The term "and / or" describes an association between objects, indicating that three relationships can exist. For example, A and / or B means: A or B, or A and B.
[0016] The term "correspondence" can refer to an association or binding relationship. The correspondence between A and B means that there is an association or binding relationship between A and B.
[0017] In the current process of enterprise digital transformation, the construction of software systems mainly relies on two mainstream technical paths: the traditional custom development model and the standardized SaaS (Software as a Service) model. However, both of these existing technical paths have significant technical limitations in resolving the core contradiction between enterprises' personalized needs and software system construction.
[0018] Specifically, while traditional custom development aims to provide personalized solutions, its technical implementation process has inherent flaws. This model typically starts with a "functional requirements document" as the technical input, upon which the development team implements discrete functionalities. In this process, business requirements must be manually communicated through multiple levels—domain experts, project managers, architects, and developers—inevitably resulting in "translation losses" and cognitive biases. More importantly, this technical path lacks a systematic mechanism to ensure the accurate construction and maintenance of the "application architecture"—the skeleton and neural network of the enterprise's business in the software world. Under development cycle pressure, architectural boundaries are easily breached, technical debt accumulates, and ultimately, a "mud ball" system—which, while functional, is internally highly coupled and difficult to evolve. On the other hand, the standardized SaaS model uses pre-built, general application architectures and data models, attempting to skip the complex business cognition process through "best practices." While this technical path improves delivery efficiency, it essentially comes at the expense of the enterprise's business personalization. When a company's unique operating model conflicts with the pre-set architecture, this technical path requires the company to "cut its feet to fit the shoes" to adapt to the software, rather than the software empowering the business. Furthermore, when enterprises adopt multiple different SaaS products, the closed and heterogeneous application architecture of each system will institutionalize "data silos," making it technically difficult to achieve enterprise-level data interoperability and business collaboration.
[0019] In summary, existing technologies (whether traditional custom development or standardized SaaS) have failed to place "building an application architecture that accurately maps to the essence of enterprise business" at the core of their solutions. They either lack systematic cognitive methods and safeguards, or they have simply abandoned the technical effort to build a customized and precise architecture for enterprises. Therefore, the industry urgently needs a new technological solution that can fundamentally bridge the gap between business understanding and software construction, thereby fundamentally solving the "double failure" dilemma of traditional development and standardized SaaS.
[0020] Combination Figure 1 As shown in the embodiments of this disclosure, a business process generation system based on an enterprise knowledge graph augmentation model is provided, including: a perception and understanding engine, a design and planning engine, and a construction and delivery engine. The perception and understanding engine receives and uses a preset industry knowledge augmentation model to parse the enterprise's multimodal raw data, obtain parsed content, and constructs a business knowledge graph based on the parsed content. The design and planning engine is used for automated system architecture design based on the business knowledge graph to obtain a design blueprint. The construction and delivery engine generates multiple code components based on the design blueprint and assembles and deploys all generated code components to obtain a software system.
[0021] The business process generation system based on an enterprise knowledge graph augmentation model provided in this disclosure receives and analyzes the enterprise's multimodal raw data using a perception and understanding engine and a preset industry knowledge augmentation model. This allows for a systematic understanding of the enterprise's operational patterns, i.e., obtaining the analyzed content. Based on this content, a machine-readable and reasonable business knowledge graph is constructed, transforming the enterprise's operational patterns into structured business knowledge. A design and planning engine then uses the business knowledge graph for automated system architecture design, converting business understanding into a machine-readable design blueprint, ensuring the generated architecture's business adaptability, technological advancement, and engineering feasibility. Finally, a build and delivery engine generates multiple code components based on the design blueprint, assembling and deploying all generated code components to obtain the software system. This results in a high-quality application architecture that is automatically generated and sustainably evolves based on a systematic understanding of the enterprise's operational patterns, fundamentally bridging the cognitive gap between business and technology and improving the severe disconnect between business understanding and software implementation. The software system is compatible with the enterprise's own business processes, eliminating the need for the enterprise to modify its unique business processes to adapt to the software.
[0022] The enterprise's multimodal raw data represents the entire business process data related to the enterprise's operations. For example, the enterprise's multimodal raw data includes contracts, logs, reports, etc. This software system is generated based on the enterprise's entire business process operation.
[0023] Preferably, the parsed content includes business entity information, business rule information, and business process information; the perception and understanding engine includes: a multimodal parsing module, used to receive the enterprise's multimodal raw data and perform deep parsing on the multimodal raw data using a preset industry knowledge enhancement model to obtain business entity information, business rule information, and business process information; and a knowledge graph construction module, used to construct a knowledge graph based on the business entity information, business rule information, and business process information using a preset enterprise unified information model to obtain an industry knowledge graph.
[0024] In this way, the multimodal parsing module receives the enterprise's multimodal raw data and uses a preset industry knowledge enhancement model to deeply mine and analyze the multimodal raw data. This enables a systematic understanding of the enterprise's operational patterns, obtaining business entity information, business rule information, and business process information—that is, obtaining the parsed content. This content is then used to construct a business knowledge graph, thereby transforming the chaotic multimodal raw data into structured business knowledge. This provides a precise and unified cognitive foundation for subsequent automated design, fundamentally bridging the cognitive gap between business and technology.
[0025] Preferably, the industry knowledge enhancement model is obtained by: using domain data to perform incremental pre-training and instruction fine-tuning on a general large language model (such as DeepSeek, Qianwen, Doubao, etc.) to form a domain expert model with deep business semantic understanding capabilities, and using this domain expert model as the industry knowledge enhancement model.
[0026] Preferably, the multimodal parsing module includes: a receiving unit for receiving multimodal raw data from the enterprise; a text parsing unit for using a preset industry knowledge enhancement model to perform named entity recognition and relation extraction on unstructured text within the multimodal raw data to obtain business entity information; a table parsing unit for using a preset industry knowledge enhancement model to perform semantic understanding on structured or semi-structured information within the multimodal raw data to obtain business rule information; and a log parsing unit for analyzing operation logs within the multimodal raw data based on a preset industry knowledge enhancement model and process mining algorithm to obtain business process information.
[0027] In this way, the text parsing unit performs named entity recognition and relation extraction on unstructured texts such as contracts and regulations; the table parsing unit parses structured or semi-structured data such as Excel reports and database schemas to understand their business semantics; and the log parsing unit uses process mining algorithms to analyze system operation logs and reconstruct the actual business processes. This allows for in-depth analysis of enterprise operations from multiple dimensions, including text, tables, and logs, to systematically understand the enterprise's operational patterns.
[0028] Preferably, based on a pre-defined industry knowledge enhancement model and process mining algorithm, the operation logs within the multimodal raw data are analyzed to obtain business process information. This includes: using the pre-defined industry knowledge enhancement model to perform preprocessing such as cleaning, labeling, and repair on the operation logs within the multimodal raw data to obtain enhanced logs rich in business semantics; and using the process mining algorithm to read the enhanced logs to reconstruct the real business process and obtain business process information.
[0029] Preferably, the core definition of the Enterprise Customer Information Facility (ECIF) is a centralized and standardized customer information management system built by an enterprise. By integrating customer data scattered across various business systems (such as CRM, ERP, e-commerce platforms, and offline store systems), it forms a unified, complete, and authoritative single view of customers, providing consistent customer information services across all business scenarios. In other words, it provides a unique, authoritative, and standardized reference system for enterprise data, serving as a common benchmark for data exchange and integration across all systems.
[0030] Preferably, a knowledge graph is constructed based on business entity information, business rule information, and business process information using a pre-defined Enterprise Unified Information Model (ECIF) to obtain an industry knowledge graph. This includes: using the pre-defined ECIF model to force alignment of business entity information, business rule information, and business process information with ECIF (i.e., performing entity standardization, attribute normalization, and relationship unification); each business entity conforming to the ECIF standard becomes a node in the knowledge graph, and directed edges are formed between entities in the knowledge graph according to the relationships defined by ECIF; business rule information is represented as constraints or reasoning logic in the graph, and business processes are represented as temporal paths and dependency chains between nodes in the graph, thereby constructing the knowledge graph and obtaining the industry knowledge graph.
[0031] Preferably, entities are standardized, such as mapping extracted terms like "buyer," "client," and "user" to the standard customer entity in ECIF. Attributes are normalized, such as converting attributes like date and amount to the standard format specified by ECIF. Relationships are unified, such as defining purchase and order placement as contractual relationships in ECIF.
[0032] Preferably, nodes will have type and attributes. For example, create a node of type "Customer" with attributes such as {Customer ID: "CUST-01-12345678", Name: "XX Technology Co., Ltd.", Status: "Activated"}. Edges will also have type and attributes (such as contract effective date). For example, create an edge of type "Signed" between the Customer node and the Contract node with attributes such as {Signing Date: "2023-05-01"}.
[0033] Preferably, business rules, such as a rule "Automatic Approval of Orders from Gold-Level Customers," can be represented in the graph as a pattern: when the customer's level attribute is "Gold" and the order status is "Pending Approval," the system can automatically update the order status to "Approved." The business process, such as from the order node to the payment node, then to the delivery node, and finally to the invoice node, clearly depicts the "order fulfillment" process.
[0034] Preferably, the design and planning engine includes: an intelligent agent module for analyzing the business knowledge graph to determine the business process model, domain map, and database schema; and an architecture decision module for determining the QoS requirements of the business scenario based on the business process model, domain map, and database schema, and determining the architecture pattern and technology stack combination based on the QoS requirements of the business scenario using a preset decision tree, and using the business process model, domain map, database schema, architecture pattern, and technology stack combination as a technical solution blueprint.
[0035] In this way, the intelligent agent module is used to analyze the business knowledge graph, determine the business process model, system service division (i.e., domain map), and database schema, and then the architecture decision module determines the optimal architecture pattern and technology stack combination based on the QoS requirements of the business scenario. The business process model, domain map, database schema, architecture pattern, and technology stack combination are used as a technical solution blueprint to simulate the decision-making process of a senior architect, transforming business understanding into an optimal, machine-readable system solution blueprint, ensuring the business adaptability, technological advancement, and engineering feasibility of the generated architecture.
[0036] Preferably, the intelligent agent module includes: a business process generation intelligent agent, used to analyze the business knowledge graph and automatically design business process models; a system architecture design intelligent agent, used to perform topological analysis of the business knowledge graph using graph neural networks, and to perform domain division and boundary context definition for domain-driven design to determine the domain map; and a data model design intelligent agent, used to analyze the business knowledge graph and design the database schema based on a preset unified coding standard.
[0037] In this way, the business process generation agent is used to automatically design and optimize business process models (such as the BPMN 2.0 model); the system architecture design agent uses graph neural networks (GNNs) to perform topological analysis on the knowledge graph, automatically performing domain-driven design (DDD) domain partitioning and bounded context definition; and the data model design agent uses a unified coding standard to automatically design a standardized database schema. Through the collaboration of multiple AI agents, domain-driven design (DDD) and architecture decisions are completed automatically, thus solving the fundamental problem of "how to correctly understand and abstract complex business", ensuring from the source that the software system can truly reflect and flexibly adapt to business.
[0038] It is understandable that an intelligent agent is a domain expert system that completes corresponding tasks based on predefined professional rules (such as DDD principles, database paradigms, and BPMN standards) and advanced algorithms (such as graph neural networks and process mining). The pre-defined unified coding standard is the enterprise element unified coding standard.
[0039] Preferably, a business process model describes the "dynamic" workflow of a business. For example, an "online insurance application" process includes: user submits application - underwriting - payment - policy generation. It can reveal the operation sequence, critical path, and user interaction points. A domain map describes the "static" conceptual structure of a business, defining core domains (such as application form, policy, customer) and the boundaries and relationships between them. It can reveal the modular boundaries of the system, data consistency boundaries, and dependencies between services. A database schema describes the "persistent" structure of data, such as tables, fields, indexes, and foreign key relationships. It can reveal data volume, access patterns (read-heavy or write-heavy), and association complexity.
[0040] Preferably, the business knowledge graph is analyzed to automatically design a business process model, including: identifying the start and end events of the business process from the business knowledge graph; using the start event as the starting node, exploring paths on the business knowledge graph to find all business activity sequences that can reach the end event; and identifying decision points and parallel points in the process according to the business rules in the business knowledge graph; mapping the start event, end event, business activity sequence, decision point, and parallel point to corresponding elements in the standard business process model, and generating a machine-readable business process model accordingly.
[0041] Preferably, the business knowledge graph is analyzed, and a database schema is designed based on a predefined unified coding standard, including: identifying core business entities in the business knowledge graph; mapping the core business entities and their attributes to tables and fields in the database according to the predefined unified coding standard, wherein the unified coding standard includes at least naming rules, data type mapping rules, and constraint rules; analyzing the relationships between core business entities, and generating inter-table associations in the database based on the relationships, wherein many-to-many relationships are mapped to associated tables; and generating a database schema definition file based on the mapping and inter-table associations, thus obtaining the database schema.
[0042] Preferably, the QoS requirements of the business scenario are determined based on the business process model, domain map, and database schema, including: traversing the process based on the business process model and domain map, mapping each business step to the responsible microservice (bounded context), identifying the call relationships between services, and determining the interaction matrix between the business scenario and the system; analyzing structural features based on the interaction matrix between the business scenario and the system, the domain map, and the database schema to obtain an architectural feature and constraint list; and specifying specific QoS indicators for each interaction point based on the call relationships between services and the architectural features and constraint list, thereby determining the QoS requirements of the business scenario.
[0043] Preferably, the build and delivery engine includes: a code generation module for generating multiple code components based on the design blueprint; a visual assembly platform for visually orchestrating, assembling, data binding, and integration testing of all generated code components to obtain an application delivery package; and a cloud-native deployment module for deploying the application delivery package in a cloud environment to obtain a software system.
[0044] In this way, the code generation module transforms the design blueprint into production-grade code, ensuring code quality. A visual assembly platform then performs visual orchestration, assembly, data binding, and integration testing on the generated code components to improve engineering efficiency. A cloud-native deployment module automates delivery performance, ultimately outputting a ready-to-use software system that meets enterprise-level quality and security standards. This shortens the development cycle of customized enterprise applications from months to weeks or even days, reducing development costs by an order of magnitude, achieving significant cost reduction and efficiency improvement.
[0045] Preferably, the code generation module includes: a template engine unit for generating scaffolding code based on the design blueprint; and an LLM code generation unit for generating business logic code based on the business logic and scaffolding code within the design blueprint, and assembling the generated business logic code with the scaffolding code to obtain multiple code components.
[0046] The template engine unit generates basic, standardized project skeletons, API controller skeletons, entity classes, and other scaffolding code, ensuring code structure and security. The LLM code generation unit then generates corresponding business logic code snippets based on the complex business logic in the design blueprint. This hybrid "template engine + LLM" code generation model, while ensuring code standardization and security, injects intelligence to handle complex business logic.
[0047] Understandably, the LLM code generation unit is a software component that uses a large language model to understand the "complex business logic description in the design scheme" and generate specific source code that implements that logic.
[0048] Preferably, the cloud-native deployment module is specifically used for: generating code IaC files based on the application delivery package; wherein the code files include Dockerfile files and Kubernetes deployment manifests; and executing CI / CD pipelines based on the code IaC files to deploy in the cloud environment.
[0049] In this way, the cloud-native deployment module automatically generates Infrastructure as Code (IaC) files such as Dockerfile and Kubernetes deployment manifest based on the application delivery package, and executes a fully automated CI / CD pipeline to complete the one-click deployment of the application and obtain the software system, thereby ensuring delivery efficiency through automation.
[0050] Here, IaC stands for Infrastructure as Code. An IaC file is a configuration file that uses code (not just manual clicks to configure) to define and manage computing infrastructure. Executing a CI / CD pipeline, after automatically generating the Dockerfile and Kubernetes configuration (IaC file), automatically triggers the following pipeline: CI phase: Automatically executes the `docker build` command to build the image and runs tests. CD phase: Automatically pushes the image to the image repository and then automatically executes the `kubectl apply` command to deploy the application to the Kubernetes cluster. The entire process requires no manual intervention, achieving true "one-click deployment."
[0051] Understandably, the perception and understanding engine system receives multimodal raw data input from enterprises and activates its multimodal parsing module. Within each parsing unit, it utilizes industry knowledge-enhanced models for deep semantic analysis, extracting business entities, rules, and processes. Subsequently, the knowledge graph construction module integrates these elements to build a unified enterprise business knowledge graph, thus achieving a cognitive leap from "perceiving data" to "understanding business." The design and planning engine operates based on the business knowledge graph. Its various agents operate in parallel, generating business process models, system service partitioning (i.e., domain maps), and data models (i.e., database schemas). The architecture decision module integrates the outputs from all parties and business requirements to generate the final optimal technical solution blueprint. This completes the decision-making process from "business understanding" to "system design." The build and delivery engine builds the system according to the design blueprint. Its code generation module generates front-end and back-end code (i.e., code components), which are imported into a visual assembly platform for final assembly and testing. Finally, the cloud-native deployment module automatically deploys the assembled application to the target environment. This achieves the engineering transformation from "design blueprint" to "runnable system." Through the above structure and process, the "AI-driven generation" of enterprise application software is realized. What is delivered is no longer static code, but a "digital twin" that is isomorphic with the business and can evolve sustainably, possessing long-term vitality.
[0052] Furthermore, the software system delivered by the business end-to-end generation system based on the enterprise knowledge graph enhancement model provided in this disclosure has a data model and architecture built upon a unified enterprise element coding standard and a business knowledge graph. When business changes, the system can be rapidly iterated by updating input data and re-triggering the platform's generation process. This mechanism enables the software to evolve synchronously with the business, fundamentally avoiding the risks of "delivery equals obsolescence" and "project failure," transforming digital investment from a high "cost" into a "strategic asset" that can continuously create value. By using the "unified enterprise element coding standard" as a "unified coding specification" throughout the entire generation process, it ensures that each generated application uses a unified "language" from the data level. This allows applications generated at different times and for different business domains to naturally achieve data dialogue and business collaboration, breaking down data barriers at the technological root and providing an unbreakable foundation for enterprises to build a future digital ecosystem that is collaborative and interconnected across all domains. It eliminates "data silos" from the source, laying a solid foundation for enterprises to build a future digital ecosystem. It also creates a data flywheel effect that becomes "smarter with use," forming a dynamically growing technological barrier. New data and optimization needs generated during system operation (after being de-identified) can be fed back to the platform for continuous optimization, creating a data flywheel effect that becomes "smarter the more it is used," forming a dynamically growing technological barrier.
[0053] The business process generation system based on an enterprise knowledge graph augmentation model is not a static tool. New data and optimization requirements generated during system operation (after anonymization) can be used to continuously optimize the industry knowledge augmentation model and enrich the knowledge base. This allows for improvements in business understanding and generation quality with each completed project, forming a self-reinforcing intelligent closed loop. This "data flywheel" effect enables the business process generation system based on an enterprise knowledge graph augmentation model provided in this disclosure to continuously evolve, building a dynamic technological barrier that is difficult for newcomers to overcome in the short term.
[0054] Combination Figure 2 As shown, this disclosure provides a method for generating a full business process based on an enterprise knowledge graph augmentation model, including: Step S101: Obtain and use a preset industry knowledge enhancement model to analyze the enterprise's multimodal raw data, obtain the analyzed content, and construct a business knowledge graph based on the analyzed content.
[0055] Step S102: Automated system architecture design based on business knowledge graph to obtain design blueprint; Step S103: Generate multiple code components based on the design blueprint, and assemble and deploy all generated code components to obtain the software system.
[0056] The business process generation method based on an enterprise knowledge graph enhancement model provided in this disclosure acquires and utilizes a preset industry knowledge enhancement model to analyze the enterprise's multimodal raw data, enabling a systematic understanding of the enterprise's operational patterns—that is, obtaining the analyzed content. Based on this content, a machine-readable and reasonable business knowledge graph is constructed, transforming the enterprise's operational patterns into structured business knowledge. The business knowledge graph is then used for automated system architecture design, converting business understanding into a machine-readable design blueprint, ensuring the generated architecture's business adaptability, technological advancement, and engineering feasibility. Multiple code components are then generated based on the design blueprint, and all generated code components are assembled and deployed to obtain the software system. This results in a high-quality application architecture that is automatically generated and sustainably evolves based on a systematic understanding of the enterprise's operational patterns, fundamentally bridging the cognitive gap between business and technology and improving the current situation of a severe disconnect between business understanding and software implementation.
[0057] Finally, it should be noted that the above preferred embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail through the above preferred embodiments, those skilled in the art should understand that various changes can be made to it in form and detail without departing from the scope defined by the claims of the present invention.
Claims
1. A business process generation system based on an enterprise knowledge graph augmentation model, characterized in that: include: The perception and understanding engine is used to receive and use a pre-set industry knowledge enhancement model to analyze the enterprise's multimodal raw data, obtain the analyzed content, and build a business knowledge graph based on the analyzed content. The design and planning engine is used for automated system architecture design based on business knowledge graphs to obtain design blueprints. The build and delivery engine is used to generate multiple code components based on the design blueprint, and to assemble and deploy all the generated code components to obtain the software system.
2. The system according to claim 1, characterized in that, The parsed content includes business entity information, business rule information, and business process information; the perception and understanding engine includes: The multimodal parsing module is used to receive the enterprise's multimodal raw data and use a preset industry knowledge enhancement model to perform in-depth parsing of the multimodal raw data to obtain business entity information, business rule information and business process information. The knowledge graph construction module is used to construct a knowledge graph based on business entity information, business rule information, and business process information using a pre-defined enterprise unified information model, thereby obtaining an industry knowledge graph.
3. The system according to claim 2, characterized in that, The multimodal parsing module includes: The receiving unit is used to receive the enterprise's multimodal raw data; The text parsing unit is used to perform named entity recognition and relation extraction on unstructured text in multimodal raw data using a pre-set industry knowledge enhancement model to obtain business entity information; The table parsing unit is used to perform semantic understanding of structured or semi-structured information in multimodal raw data using a pre-set industry knowledge enhancement model to obtain business rule information; The log parsing unit is used to analyze operation logs within multimodal raw data based on a preset industry knowledge enhancement model and process mining algorithm to obtain business process information.
4. The system according to claim 1, characterized in that, The design and planning engine includes: The intelligent agent module is used to analyze the business knowledge graph and determine the business process model, domain map, and database schema. The architecture decision module is used to determine the QoS requirements of business scenarios based on business process models, domain maps, and database schemas. Based on the QoS requirements of business scenarios, it uses a pre-defined decision tree to determine the architecture pattern and technology stack combination, and uses the business process model, domain map, database schema, architecture pattern, and technology stack combination as a technical solution blueprint.
5. The system according to claim 4, characterized in that, The intelligent agent module includes: A business process generation intelligent agent is used to analyze business knowledge graphs and automatically design business process models. The system architecture design intelligent agent is used to perform topological analysis of the business knowledge graph using graph neural networks, and to perform domain partitioning and boundary context definition for domain-driven design, thereby determining the domain map; The data model designs an intelligent agent to analyze the business knowledge graph and design the database schema based on a pre-defined unified coding standard.
6. The system according to claim 1, characterized in that, The build and delivery engine includes: The code generation module is used to generate multiple code components based on the design blueprint. A visual assembly platform is used to visually orchestrate, assemble, data bind, and integrate test all generated code components to obtain an application delivery package. The cloud-native deployment module is used to deploy application delivery packages in a cloud environment to obtain the software system.
7. The system according to claim 6, characterized in that, The code generation module includes: Template engine unit, used to generate scaffolding code based on design blueprints; The LLM code generation unit is used to generate business logic code based on the business logic and scaffolding code within the design blueprint, and then assemble the generated business logic code with the scaffolding code to obtain multiple code components.
8. The system according to claim 6, characterized in that, The cloud-native deployment module is specifically used for: The application delivery package generates IaC files; these files include Dockerfiles and Kubernetes deployment manifests. Based on the code IaC file, execute the CI / CD pipeline for cloud environment deployment.
9. A method for generating the entire business process based on an enterprise knowledge graph augmentation model, characterized in that: include: Acquire and utilize a pre-defined industry knowledge enhancement model to analyze the enterprise's multimodal raw data, obtain the analyzed content, and construct a business knowledge graph based on the analyzed content; Automated system architecture design based on business knowledge graphs to obtain design blueprints; Multiple code components are generated based on the design blueprint, and all generated code components are assembled and deployed to obtain the software system.
Citation Information
Patent Citations
Electric power question answering and visualization system based on knowledge graph
CN115269862A
Multi-modal man-machine interaction view and component expression method and system
CN120654284A
Application construction method and device, equipment, storage medium and product
CN120803419A
Enterprise finance and taxation intelligent management system based on data analysis
CN120912153A
Code generation system based on natural language and meta-model driving
CN121209875A
Cited By
DDD-based software automatic design method and apparatus, and electronic device
CN121934820A
