A service processing method, electronic device, storage medium and program product
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- INDUSTRIAL AND COMMERCIAL BANK OF CHINA
- Filing Date
- 2026-02-14
- Publication Date
- 2026-06-02
Smart Images
Figure CN122134281A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of artificial intelligence technology and can be applied to the field of financial technology, particularly to a business processing method, electronic device, storage medium, and program product. Background Technology
[0002] With the rapid development of large language model technology, the financial industry is vigorously introducing large language models to reshape business processes in order to improve the efficiency of business processing.
[0003] However, existing technologies for large language models have a high barrier to entry, and there are discrepancies in the process execution of different specific types of business, making it impossible to use the same large language model to connect to processes of multiple types of business. This results in certain constraints on the financial industry's use of large language model technology across all scenarios, and the actual business results are difficult to match the amount of resources invested.
[0004] Therefore, there is an urgent need to provide a way for multiple types of businesses to share a large language model, reduce the application threshold of large language models, and realize the application of large language model technology in all scenarios of the financial industry. Summary of the Invention
[0005] This invention provides a business processing method, electronic device, storage medium, and program product to enable the reuse of large language models for multiple types of business and lower the barrier to entry for using large language models.
[0006] According to one aspect of the present invention, a business processing method is provided, the method comprising: Obtain the user's business requirement description information, and parse the business requirement description information to obtain the user's business processing requirements under the target business type; Based on the prompt words and / or thought processes corresponding to the target business type, supplement the business processing requirements with requirement description data to obtain business processing related parameters. Based on the business processing requirements and the business processing associated parameters, a trained orchestration model is used to orchestrate the business processing flow and generate a dynamic service plugin invocation scheme. Based on the dynamic service plugin invocation scheme and the business processing association parameters, the corresponding service plugin is invoked to perform business processing under the target business type.
[0007] According to another aspect of the present invention, a business processing apparatus is provided, the apparatus comprising: The business requirement parsing module is used to obtain the user's business requirement description information and parse the business requirement description information to obtain the user's business processing requirements under the target business type. The data supplementation module is used to supplement the business processing requirements with requirement description data based on prompt words and / or thought processes corresponding to the target business type, so as to obtain business processing related parameters. The process orchestration module is used to orchestrate the business process according to the business processing requirements and the business processing related parameters, using a trained orchestration model to generate a dynamic service plugin invocation scheme. The business processing module is used to call the corresponding service plugin to perform business processing under the target business type according to the dynamic service plugin invocation scheme and the business processing association parameters.
[0008] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising: At least one processor; and a memory communicatively connected to said at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, which enables the at least one processor to perform the business processing method described in any embodiment of the present invention.
[0009] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions for causing a processor to execute and implement the business processing method described in any embodiment of the present invention.
[0010] According to another aspect of the present invention, a computer program product is provided, comprising a computer program that, when executed by a processor, implements the business processing method described in any embodiment of the present invention.
[0011] The technical solution of this invention obtains and parses the user's business requirement description information to obtain the user's business processing requirements under the target business type. Based on prompts and / or thought processes corresponding to the target business type, the business processing requirements are supplemented with requirement description data to obtain business processing association parameters. Based on the business processing requirements and business processing association parameters, a trained orchestration model is used to orchestrate the business processing flow, generating a dynamic service plugin invocation scheme. Based on the dynamic service plugin invocation scheme and business processing association parameters, the corresponding service plugin is invoked to perform business processing under the target business type. This solves the problem of applying large language model technology across all scenarios in the financial industry. By supplementing the business processing requirement data, accurate requirement matching can be achieved. By orchestrating the business processing flow, multiple types of businesses can be processed using a large model, lowering the application threshold of the large language model. Embedded service plugins enable highly cohesive and loosely coupled business processing across multiple business types.
[0012] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description
[0013] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0014] Figure 1 This is a flowchart of a business processing method provided according to Embodiment 1 of the present invention; Figure 2 This is a flowchart of a business processing method provided according to Embodiment 2 of the present invention; Figure 3 This is a schematic diagram of a business processing flow for a one-stop dialogue service provided according to Embodiment 2 of the present invention; Figure 4 This is a flowchart illustrating another business processing method provided according to Embodiment 2 of the present invention; Figure 5 This is a schematic diagram of the structure of a business processing device according to Embodiment 3 of the present invention; Figure 6 This is a schematic diagram of the structure of an electronic device that implements the business processing method of the present invention. Detailed Implementation
[0015] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.
[0016] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.
[0017] Example 1 Figure 1 This is a flowchart of a business processing method according to Embodiment 1 of the present invention. This embodiment is applicable to situations in the financial industry where artificial intelligence technology is used for business process processing across all scenarios. The method can be executed by a business processing device, which can be implemented in hardware and / or software. This device can be configured in an electronic device, such as a mobile device like a mobile phone, tablet computer (PAD), or wearable device, or a personal computer (PC), server, controller, or distributed cluster device. Figure 1 As shown, the method includes: Step 110: Obtain the user's business requirement description information and parse the business requirement description information to obtain the user's business processing requirements under the target business type.
[0018] The business requirement description information can be content entered by the user in the dialogue service, or it can be business requirement data identified by the business processing device based on data from the interaction between the user and the business processing device. The acquisition of various data involved in obtaining the business requirement description information is only done with user authorization during the business processing, and the acquisition method is reasonable and legal.
[0019] There are several ways to parse business requirement descriptions. For example, neural network models or keyword recognition technologies can be used to parse the descriptions to obtain the user's business processing requirements under the target business type.
[0020] In business processing, various types of business can be handled. These can include, but are not limited to, credit due diligence report generation, policy Q&A, investment and financing analysis, and risk monitoring. By using a unified orchestration model to orchestrate the business processes across multiple business types, a single large language model can be shared, eliminating the need for separate large language models for each type of business or for generating complex large language models for multiple business types. This lowers the barrier to entry for large language models, enabling their application across all business scenarios in the financial sector.
[0021] Step 120: Based on the prompts and / or thought processes corresponding to the target business type, supplement the business processing requirements with requirement description data to obtain business processing related parameters.
[0022] Hint words can be words generated based on business logic and services, facilitating information completion during business processing and improving the accuracy of determining service plugin invocation schemes. Thinking chains can be chains generated based on business logic and services, such as the connection information of multiple services that need to be called sequentially during business processing, or multiple data that need to be retrieved sequentially. Supplementing requirement description data with thinking chains can compensate for the deficiencies in the chain under strongly rule-based business services.
[0023] Optionally, based on the prompt words and / or thought processes corresponding to the target business type, supplement the requirement description data of the business processing requirements to obtain business processing related parameters, including: generating prompt words based on the prompt word template, prompt word recommendation table, and prompt word evaluation results corresponding to the target business type; generating thought processes based on the service workflow, user adjustment and optimization results, and execution feedback evaluation results corresponding to the target business type; and supplementing the requirement description data of the business processing requirements based on the prompt words and / or thought processes to obtain business processing related parameters.
[0024] The prompt word template can be a pre-set prompt word template based on the business type. The prompt word recommendation table can be a table formed by recommending prompt words associated with business processing based on the business type. The prompt word evaluation result can be the result of evaluating the prompt word template and the prompt word recommendation table based on the business type to optimize the prompt word recommendation. For example, in business processing, users can provide feedback on business processing results or process data. The electronic device evaluates the prompt words based on the data provided by the user to optimize the prompt word recommendation under that business type. For example, in bank draft business processing, the business processing requirements for bank draft acceptance can be identified, and bank draft acceptance can be completed through processes such as prompt word completion of contract element recognition, invoice information recognition, online verification, registration information collection, penalty information collection, due diligence report generation, and review.
[0025] The thought process flow can include service workflows corresponding to the business type. Service workflows can be the items to be processed sequentially in the business process. During actual business execution, users can adjust the service workflows, generating user-adjusted optimization results. Similarly, during business processing, users can provide feedback on business processing results or process data. The electronic device evaluates the thought process flow based on the user feedback data to optimize the recommended thought process flow for that business type.
[0026] In business processing, supplementing the data description of business processing requirements by using prompt words and / or thought processes can refine or extend the processing items, thereby improving or further extending the user's business processing, and enhancing the accuracy of service scheduling and the reliability of business process processing.
[0027] Optionally, based on prompt words and / or thought processes, the business processing requirements are supplemented with requirement description data to obtain business processing related parameters, including: extracting the existing first attribute information in the business processing requirements; determining the missing second attribute information in the business processing requirements based on the basic attribute information required for reasoning of the large-scale orchestration model and the first attribute information; and supplementing the second attribute information with requirement description data based on the attribute information covered by prompt words and / or thought processes to obtain business processing related parameters.
[0028] The dynamic service plugin invocation scheme is determined based on business-related attribute information during large-scale model inference. To ensure the accuracy of the inference, the business-related attribute information needs to be supplemented. For example, the basic attribute information required for large-scale model inference includes, but is not limited to: the task objective of the business process, the role information in the business process, the input data of the business process, the constraints of the business process, the example guidance information of the business process, and the output format of the business process. The attribute information in the business process requirements is usually incomplete. To ensure the reliability of the business process, the attribute information in the business process requirements can be extracted to determine the first attribute information. The first attribute information is compared with the basic attribute information to determine the second attribute information missing in the business process requirements. Then, based on the prompt words corresponding to the business type and / or the attribute information covered by the thought process, the second attribute information is supplemented to obtain the business process related parameters.
[0029] There are several ways to supplement the second attribute information using prompts and / or thought processes. For example, one could directly supplement the missing second attribute information in the business processing requirements based on prompts and / or thought processes corresponding to the business type. Alternatively, one could determine and display a list of supplementary second attribute information based on prompts and / or thought processes corresponding to the business type, allowing users to confirm or supplement information based on the displayed list, thereby supplementing the requirement description data and obtaining the business processing-related parameters.
[0030] When supplementing business processing requirements through prompts and / or thought processes, refining the information down to the level of attribute information to fill in the missing attribute information can ensure the comprehensiveness of business processing requirements, thereby improving the reliability of orchestrating large-scale model reasoning.
[0031] Step 130: Based on business processing requirements and related parameters, use the trained orchestration model to orchestrate the business processing flow and generate a dynamic service plugin invocation scheme.
[0032] The orchestration model can be a pre-trained, large-scale language model used for orchestrating business processes. For example, it can acquire business processes for different business types from historical business processes, and label the corresponding business processing requirements and related parameters to form sample data. This sample data is then divided into training and testing sets. The orchestration model is trained using data from the training set and the pre-defined large-scale language model, and the training results are evaluated using data from the testing set. When the evaluation results meet expectations, the orchestration model is obtained. In business processes, inputting business processing requirements and related parameters into the orchestration model enables the orchestration of business processes, resulting in a dynamic service plugin invocation scheme.
[0033] A dynamic service plugin invocation scheme can define the service plugins required for business processing under a specific business type, along with the invocation order of these plugins. These service plugins can be pre-embedded business processes, such as those developed based on actual business workflows and encountered issues. For example, a service plugin could be a neural network model with fewer parameters than a preset data scale. Optionally, electronic devices can embed service plugins for financial fields such as invoice verification, trade contract recognition, financial analysis, due diligence report generation, policy Q&A, system operation guidance, document generation, voice input, and OCR (Optical Character Recognition) information collection. By setting up service plugins for various business types within the electronic device, high cohesion and low coupling can be achieved. When invoking a service plugin, each plugin can be called individually, and the business processing can be achieved through the invocation order. This low coupling between service plugins avoids the complexity of a large language model for multiple business applications. Furthermore, the high density of service plugins within the electronic device, i.e., high cohesion, enables the device to support business processing for multiple business types, thus realizing a large language model for full-scenario applications in the financial industry.
[0034] Optionally, based on business processing requirements and related parameters, a trained orchestration model is used to orchestrate the business processing flow and generate a dynamic service plugin invocation scheme. This includes: retrieving business-related data from a knowledge base based on the related parameters; and using the trained orchestration model to orchestrate the business processing flow based on the business processing requirements, related parameters, and related data to generate a dynamic service plugin invocation scheme.
[0035] A knowledge base can include a policy database, a case database, an industry information database, external websites, a knowledge base definition, and a knowledge base interface (API). By providing unstructured knowledge data, the knowledge base can intelligently parse the content of various document types into the search engine. Simultaneously, it feeds corpus data into the orchestration model, addressing gaps in the model's expertise and enabling precise orchestration of business processes across multiple business types.
[0036] By retrieving business-related data from a knowledge base, the data required for business processing can be automatically supplemented and improved, reducing user workload. This business-related data can be data pre-stored by the user on their electronic devices. All retrieval of this business-related data is authorized by the user and conducted in a legal and legitimate manner.
[0037] Step 140: Based on the dynamic service plugin invocation scheme and business processing associated parameters, invoke the corresponding service plugin to perform business processing under the target business type.
[0038] The dynamic service plugin invocation scheme determines the required service plugins and their invocation order. Based on this order, the corresponding service plugins are invoked, and their business processing methods are used to process the associated parameters to obtain the final result.
[0039] When each service plugin performs business operations, it can do so through its respective processing method. The coupling between service plugins is low, which avoids the complexity of a large language model for multiple types of business applications. By setting up high-density service plugins (i.e., high cohesion) within electronic devices, the devices can support multiple types of business processing, thereby realizing a large language model for full-scenario applications in the financial industry.
[0040] Optionally, before calling the corresponding service plugin to perform business processing under the target business type, the process also includes: orchestrating the processing methods of business services under each business type to form each service plugin; the service plugin includes the plugin service form, the plugin scope of use, and the plugin calling method.
[0041] Electronic devices can orchestrate the business processing flow of services across multiple business types through specific business type services and decision orchestration service engines, forming various service plugins and saving them. For example, electronic devices can orchestrate the business processing flow and required data of each service plugin to prompt users to input data for processing. Specifically, electronic devices can register the form, scope of application, invocation method, and processing method of service plugins to form full-scenario business services. Plugin service forms can include, but are not limited to: application programming interfaces (APIs), scripts, containers, and small neural networks. The scope of plugin application can be the business type that the service plugin can apply to, or the business operations that the service plugin can handle. The invocation method of service plugins can include, but is not limited to, keywords for service plugin routing and data required for business processing. The processing method of service plugins can be the business processing flow and execution method. For example, electronic devices can orchestrate corresponding service plugins based on historical business processing flows to obtain service plugins. Alternatively, electronic devices can also support administrators to personalize the orchestration of business processes to obtain service plugins. By embedding multiple service plugins in electronic devices, high cohesion and low coupling of service plugins can be achieved, enabling intelligent processing of full-scenario business in the financial field and reducing the application threshold and cost of large language models.
[0042] The technical solution of this invention obtains and parses the user's business requirement description information to obtain the user's business processing requirements under the target business type. Based on prompts and / or thought processes corresponding to the target business type, the business processing requirements are supplemented with requirement description data to obtain business processing association parameters. Based on the business processing requirements and business processing association parameters, a trained orchestration model is used to orchestrate the business processing flow, generating a dynamic service plugin invocation scheme. Based on the dynamic service plugin invocation scheme and business processing association parameters, the corresponding service plugin is invoked to perform business processing under the target business type. This solves the problem of applying large language model technology across all scenarios in the financial industry. By supplementing the business processing requirement data, accurate requirement matching can be achieved. Through business processing flow orchestration, multiple types of businesses can be processed using a large model, lowering the application threshold of the large language model. It eliminates the need to train separate neural network models for each type of business; instead, it achieves business processing across all business scenarios through highly cohesive and loosely coupled service plugin invocation, reducing the cost and complexity of applying the large language model to business processing.
[0043] Example 2 Figure 2 This is a flowchart of a business processing method according to Embodiment 2 of the present invention. This embodiment is a further refinement of the above technical solution, and the technical solution in this embodiment can be combined with various optional solutions in one or more of the above embodiments. Figure 2 As shown, the method includes: Step 210: Obtain the user's business requirement description information and parse the business requirement description information to obtain the user's business processing requirements under the target business type.
[0044] Optionally, obtaining the user's business requirement description information includes: obtaining the user's dialogue service content and using the dialogue service content as the business requirement description information; or, identifying the user's next business processing instruction in the current business processing scenario, and in response to the user's confirmation operation of the next business processing instruction, using the next business processing instruction as the business requirement description information.
[0045] The acquisition of dialogue service content is conducted only after user authorization during business processing, and the acquisition method is reasonable and legal. Dialogue service content can be data generated when a user interacts with an electronic device during business processing. For example, dialogue service content can be voice, text, or image data. For instance, dialogue service content could be something like, "Please generate a due diligence report on the business of [Company Name]" or "Please provide risk information for [Client Name]."
[0046] In practical applications, user interaction with electronic devices can include two service access modes: one-stop dialogue service and scene embedding. The one-stop dialogue service can provide users with a variety of task-aware services. Figure 3 This is a schematic diagram illustrating a business processing flow for a one-stop dialogue service provided according to Embodiment 2 of the present invention. Figure 3 As shown, the one-stop dialogue service can include three types of task-aware services: intelligent dialogue, automatic guidance, and process-triggered.
[0047] Intelligent conversational services provide business services through natural language interaction. For example, users can directly request business processing via voice commands. The large-scale orchestration model within the electronic device can schedule service plugins based on user queries to achieve corresponding business processing, such as scheduling service plugin 1, service plugin 2, ..., and service plugin N. Automatic guidance involves intelligently sensing business instructions and pushing data to guide users to initiate business processing. For example, in automatic guidance, business information, material extraction, risk analysis, and system operation instructions can be automatically processed and pushed to the user, guiding them through subsequent business operations. During business processing, the large-scale orchestration model within the electronic device can schedule service plugins based on the user's business operations to achieve corresponding business processing. Process-triggered services are triggered by the user within the business process through conversational commands and confirmation mechanisms. For example, in bank draft processing, when a user initiates a bank draft acceptance instruction, the electronic device automatically identifies contract elements -> automatically identifies invoice information and completes online verification -> automatically collects information such as business registration and environmental penalties, generates a due diligence report and pushes it to the user -> after the user reviews the due diligence information, it automatically assists the user in completing the bank draft acceptance. In a process-triggered model, during business processing, the large-scale model orchestrated within the electronic device can schedule service plugins to perform preliminary data processing based on business needs. After user confirmation, the service plugins are invoked to complete the final business processing. Users can handle various business transactions with a single click through a one-stop dialogue service, such as generating credit due diligence reports, answering policy questions, analyzing investment and financing, and monitoring risk—all personalized business needs.
[0048] Furthermore, electronic devices can also determine business requirement description information through scene-embedded services. In scene-embedded services, auxiliary operations can be proactively performed for the user within the context of business processing. For example, in scene-embedded services, electronic devices can identify and determine the user's next business processing instruction based on the current business processing scenario. When the user determines to execute the next business processing instruction, this instruction can be used as the business requirement description information.
[0049] Optionally, the user's next business processing instruction can be identified in the current business processing scenario, including: obtaining the user's business processing function in the current business processing scenario; and identifying the user's next business processing instruction based on the business processing function and the current business processing scenario.
[0050] The business processing function can be an authorization function for user business processes. For example, corresponding business processing functions can be set based on user business roles. Under this function, users are guided to the next step of the business process. By proactively identifying the user's next business processing instruction within the embedded service, more in-depth or broader business processing can be provided, enriching the business processing scenarios; and it can save user processing operations, requiring only confirmation of the next business processing instruction. Considering the business processing function in the identification of the next business processing instruction allows for scenario-based limitation of the functions used by the user.
[0051] Optionally, the business requirement description information is parsed to obtain the user's business processing requirements under the target business type, including: determining the information parsing granularity based on the business requirement description information; calling the pre-trained parsing model corresponding to the information parsing granularity to parse the business requirement description information and obtain the user's business processing requirements under the target business type.
[0052] There are multiple ways to parse business requirement descriptions. The granularity of information parsing indicates the method used. For example, business requirement descriptions can be parsed using keyword recognition, document decomposition, image recognition, and content understanding. Corresponding large-scale parsing models can be trained for different parsing methods, and these models are then used to parse the business requirement descriptions. The granularity of information parsing allows different methods to be invoked. The granularity can be the information contained within the business requirement description, or it can be information determined based on the business requirement description. For example, document-based business requirement descriptions can be parsed using a document decomposition-based model. Conversational business requirement descriptions can be parsed using a keyword recognition-based model. Image-based business requirement descriptions can be parsed using an image recognition-based model. For complex content, a content understanding-based model can be used.
[0053] Different types of business requirement descriptions can be parsed using large parsing models with different information parsing granularities, enabling accurate determination of business processing requirements.
[0054] In addition, when parsing business requirement description information, content enhancement processing can be used. By enhancing retrieval and referencing information from external knowledge bases, answers or content can be generated to improve prediction quality and accuracy and enhance the generation effect of various types of content.
[0055] Step 220: Generate prompt words based on the prompt word template corresponding to the target business type, the prompt word recommendation table, and the prompt word evaluation results.
[0056] Step 230: Generate a thought process based on the service workflow corresponding to the target business type, user adjustment and optimization results, and execution feedback evaluation results.
[0057] Step 240: Supplement the business processing requirements with requirement description data based on prompt words and / or thought processes to obtain business processing related parameters.
[0058] Optionally, based on prompt words and / or thought processes, the business processing requirements are supplemented with requirement description data to obtain business processing related parameters, including: extracting the existing first attribute information in the business processing requirements; determining the missing second attribute information in the business processing requirements based on the basic attribute information required for reasoning of the large-scale orchestration model and the first attribute information; and supplementing the second attribute information with requirement description data based on the attribute information covered by prompt words and / or thought processes to obtain business processing related parameters.
[0059] Step 250: Based on business processing requirements and related parameters, use the trained orchestration model to orchestrate the business processing flow and generate a dynamic service plugin invocation scheme.
[0060] Optionally, based on business processing requirements and related parameters, a trained orchestration model is used to orchestrate the business processing flow and generate a dynamic service plugin invocation scheme. This includes: retrieving business-related data from a knowledge base based on the related parameters; and using the trained orchestration model to orchestrate the business processing flow based on the business processing requirements, related parameters, and related data to generate a dynamic service plugin invocation scheme.
[0061] Step 260: Based on the dynamic service plugin invocation scheme and business processing associated parameters, invoke the corresponding service plugin to perform business processing under the target business type.
[0062] Optionally, before calling the corresponding service plugin to perform business processing under the target business type, the process also includes: orchestrating the processing methods of business services under each business type to form each service plugin; the service plugin includes the plugin service form, the plugin scope of use, and the plugin calling method.
[0063] Optionally, based on the dynamic service plugin invocation scheme and business processing associated parameters, the corresponding service plugin is invoked to perform business processing under the target business type, including: determining the missing basic business data in the business processing associated parameters according to the plugin invocation method corresponding to the currently invoked service plugin, and prompting the user to provide the basic business data; and performing business processing under the target business type based on the plugin service form, plugin usage scope, plugin invocation method, business processing associated parameters, and basic business data corresponding to the currently invoked service plugin.
[0064] By identifying missing basic business data in the parameters associated with business processing based on the plugin invocation method and prompting the user to provide it, the comprehensiveness, accuracy, and reliability of information in business processing can be guaranteed. By understanding the corresponding plugin service type, plugin scope, and plugin invocation method, the service plugin can be successfully invoked to process the associated parameters and basic business data.
[0065] Based on the above implementation method, optionally, the method further includes: acquiring user evaluation feedback data in business processing, analyzing the evaluation feedback data to obtain analysis results; and adjusting the trained orchestration model according to the analysis results.
[0066] By analyzing, adjusting, and optimizing the orchestration model based on user feedback data within the application, the application performance of the orchestration model can be improved, and business processing satisfaction can be increased.
[0067] The technical solution of this invention involves obtaining and parsing a user's business requirement description to obtain the user's business processing requirements under the target business type; generating prompts based on prompt templates, prompt recommendation tables, and prompt evaluation results corresponding to the target business type; generating thought processes based on the service workflow, user adjustment and optimization results, and execution feedback evaluation results corresponding to the target business type; supplementing the business processing requirements with requirement description data based on prompts and / or thought processes to obtain business processing related parameters; and using a trained orchestration model to process the business processing requirements based on the business processing requirements and business processing related parameters. The system orchestrates workflows and generates dynamic service plugin invocation schemes. Based on these schemes and business processing parameters, it invokes the corresponding service plugins to perform business processing for the target business type. This solves the problem of applying large language model technology across all scenarios in the financial industry. By supplementing business processing requirements with data, it achieves precise matching of needs. Through workflow orchestration, it enables multiple types of business to be processed using large models, lowering the application threshold of large language models. Instead of training separate neural network models for each type of business, it achieves business processing across all business scenarios through highly cohesive and loosely coupled service plugin invocations, reducing the cost and complexity of applying large language models to business processing.
[0068] Figure 4 This is a flowchart illustrating another business processing method provided according to Embodiment 2 of the present invention. Figure 4 As shown, when implementing a multi-type business reuse intelligent model, electronic devices can set up a service access layer, a business scheduling layer, a service plugin layer, intelligent assistance, and an orchestration model in their software architecture. The service access layer provides unified sessions, multimodal session interaction (supporting rich text interaction such as text, images, audio, and video), access control (divided into tenant permissions, scene permissions, and role permissions), and multi-channel adaptation. It also provides lightweight integration tools, enabling businesses to achieve low-cost, one-click integration. Simultaneously, the service access layer undertakes basic functions such as traffic control, user auditing, and event tracking feedback for corpus feedback in algorithm engine optimization. Users can initiate business processing requests through the service access layer in the electronic device, such as generating user due diligence or solution development requirements.
[0069] The business scheduling layer provides an algorithm engine that distributes and executes algorithmic capabilities through intent recognition algorithms, enabling intelligent semantic task scheduling and determining business processing requirements. The business scheduling layer addresses the illusion problem through a content enhancement processing layer. For complex business processes, electronic devices, including service orchestration layers, utilize bytecode generation technology to create workflow orchestration engines, supporting custom orchestration of plug-in service processes and optimizing large language models. The business scheduling layer can also supplement information with a prompt dictionary and thought link library, enabling one-click association of user input. For example, in user due diligence, multiple business modules are involved, such as financial analysis, public opinion analysis, and equity analysis. The business scheduling layer can identify the intent of the conversation during business processing and intelligently route the services required for business processing. For instance, in user due diligence, intent recognition intelligently routes to the financial analysis service plug-in. Both the service access layer and the business scheduling layer can serve as part of a larger orchestration model.
[0070] The service plugin layer defines atomic business service capabilities and configures service-oriented interfaces for different business types, supporting various service interface configurations. It also supports parameter configuration and intent mapping, facilitating workflow orchestration and reducing repetitive service development. Each service plugin can be implemented using a small network model. For question-and-answer service plugins, this can be either a small network model or a large language model. Specifically, the implementation of each service plugin can be customized according to user requirements.
[0071] When calling service plugins in the service plugin layer, service enhancement can be performed through the content enhancement processing layer. For example, by combining private knowledge bases and online information, user financial information can be improved. The content enhancement processing layer may include a domain-level knowledge retrieval library, and can import data in various document formats. It can also segment and vectorize imported data documents, ultimately outputting enhanced results through the search engine. The content enhancement processing layer can solve problems such as illusions and instability in the output results of large language models.
[0072] In addition, when performing business processing through service plugins, large language models can be invoked to combine, summarize, refine, and output structured information.
[0073] After completing the business processing, the service access layer can obtain user feedback on the business processing process, embed user behavior data into the lake, process the data and feed it back to the algorithm for corpus feedback, and carry out continuous algorithm iteration in the business domain scheduling layer to achieve more satisfactory business processing.
[0074] By integrating dialogue processing, service orchestration, service plugins, and knowledge bases into electronic devices, one-click introduction of business domains can be supported, reducing the access costs for various business parties. The underlying layer encapsulates and shields the abstraction capabilities of large language models, exposing them to business personnel with streamlined service configurations, greatly lowering the threshold for practitioners to use large language models.
[0075] In the financial industry, business processing procedures are often complex, as even a small error can lead to serious financial risks. Therefore, strict adherence to procedures and zero tolerance for errors are essential. However, maintaining extremely high accuracy in intelligent business process processing is a significant challenge. The technical solution of this invention executes business processing flows through business service plugins for various business types. It can accurately execute a series of complex business processes, using contextual queries and dialogic confirmation to execute each step, providing business personnel with an immersive assistant service. Previously, business personnel needed to switch between different business systems to execute operations; now, a single dialogic interaction is sufficient to complete complex business processes, greatly improving efficiency and accuracy. Furthermore, the technical solution of this invention sets up service plugins for multiple business types in electronic devices and connects them to a large orchestration model. This enables multiple business types to reuse the orchestration model and service plugins. Within the electronic device, the high-cohesion, low-coupling service-oriented capabilities lower the barrier to entry for using the intelligent model.
[0076] Example 3 Figure 5 This is a schematic diagram of a business processing device according to Embodiment 3 of the present invention. Figure 5As shown, the device includes: a business requirement analysis module 510, a data supplementation module 520, a process orchestration module 530, and a business processing module 540. Among them: The business requirement parsing module 510 is used to obtain the user's business requirement description information and parse the business requirement description information to obtain the user's business processing requirements under the target business type. The data supplementation module 520 is used to supplement the business processing requirements with requirement description data based on prompt words and / or thought processes corresponding to the target business type, and obtain business processing related parameters. The process orchestration module 530 is used to orchestrate business processes based on business processing requirements and related parameters, using a trained orchestration model to generate dynamic service plugin invocation schemes. The business processing module 540 is used to call the corresponding service plugin to perform business processing under the target business type based on the dynamic service plugin invocation scheme and business processing associated parameters.
[0077] Optional, the business requirements parsing module 510 includes: The first business requirement description information determination unit is used to obtain the user's dialogue service content and use the dialogue service content as the business requirement description information; or... The second business requirement description information determination unit is used to identify the user's next business processing instruction in the current business processing scenario, and in response to the user's confirmation operation of the next business processing instruction, to use the next business processing instruction as business requirement description information.
[0078] Optionally, the second business requirement description information determination unit includes: The business processing function acquisition sub-unit is used to acquire the user's business processing functions in the current business processing scenario. The next business processing instruction recognition subunit is used to identify the user's next business processing instruction based on the business processing function and the current business processing scenario.
[0079] Optional, the data supplementation module 520 includes: The prompt word generation unit is used to generate prompt words based on the prompt word template corresponding to the target business type, the prompt word recommendation table, and the prompt word evaluation results; The thought process generation unit is used to generate thought processes based on the service workflow corresponding to the target business type, user adjustment and optimization results, and execution feedback evaluation results. The data supplementation unit is used to supplement the business processing requirements with requirement description data based on prompt words and / or thought processes, and obtain business processing related parameters.
[0080] Optional, the process orchestration module 530 includes: The business-related data acquisition unit is used to retrieve business-related data from the knowledge base based on business processing related parameters. The process orchestration unit is used to orchestrate business processes based on business processing requirements, business processing related parameters, and business related data, using a trained orchestration model to generate dynamic service plugin invocation schemes.
[0081] Optional, the business requirements parsing module 510 includes: The information parsing granularity determination unit is used to determine the information parsing granularity based on the information described in the business requirements. The business requirement parsing unit is used to call the trained parsing model corresponding to the information parsing granularity to parse the business requirement description information and obtain the user's business processing requirements under the target business type.
[0082] Optional, data supplementation unit includes: The first attribute information determination subunit is used to extract the first attribute information that already exists in the business processing requirements; The second attribute information determination subunit is used to determine the missing second attribute information in the business processing requirements based on the basic attribute information and the first attribute information required for the orchestration of the large model inference. The data supplementation subunit is used to supplement the second attribute information with requirement description data based on the attribute information covered by the prompt words and / or thought process, so as to obtain the business processing related parameters.
[0083] Optionally, the device may also include: The service plugin orchestration module is used to orchestrate the processing methods of business services under each business type before calling the corresponding service plugin to perform business processing under the target business type, forming each service plugin; the service plugin includes the plugin service form, the plugin scope of use, and the plugin calling method.
[0084] Optional, the business processing module 540 includes: The business basic data acquisition unit is used to determine the missing business basic data in the business processing associated parameters based on the plugin invocation method corresponding to the currently invoked service plugin, and prompt the user to provide the business basic data. The business processing unit is used to perform business processing under the target business type based on the plugin service type, plugin usage scope, plugin invocation method, business processing associated parameters, and basic business data of the currently invoked service plugin.
[0085] Optionally, the device may also include: The data analysis module is used to acquire user feedback data during business processing and to analyze the feedback data to obtain analysis results. The model tuning module is used to adjust the trained orchestration model based on the analysis results.
[0086] The business processing apparatus provided in the embodiments of the present invention can execute the business processing method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of executing the method.
[0087] In the technical solutions of this invention, the collection, storage, use, processing, transmission, provision, and disclosure of user personal information (such as business requirement description information, business processing related parameters, dialogue service content, business related data, and business basic data) all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.
[0088] The information collected is information and data authorized by the user or fully authorized by all parties. The collection, storage, use, processing, transmission, provision, disclosure and application of the relevant data all comply with the relevant laws, regulations and standards of the relevant countries and regions, necessary confidentiality measures have been taken, and they do not violate public order and good morals. Corresponding operation portals are provided for users to choose to authorize or refuse.
[0089] When performing business processing for users using the business processing method provided in this embodiment of the invention, a corresponding operation entry is provided for users to choose to agree to or refuse automated business processing; if the user chooses to refuse, the process enters the expert decision-making process.
[0090] Example 4 Figure 6 A schematic diagram of an electronic device 10, which can be used to implement embodiments of the present invention, is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.
[0091] like Figure 6As shown, the electronic device 10 includes at least one processor 11 and a memory, such as a read-only memory (ROM) or random access memory (RAM), communicatively connected to the at least one processor 11. The memory stores computer programs executable by the at least one processor. The processor 11 can perform various appropriate actions and processes based on the computer program stored in the ROM 12 or loaded into the RAM 13 from the storage unit 18. The RAM 13 can also store various programs and data required for the operation of the electronic device 10. The processor 11, ROM 12, and RAM 13 are interconnected via a bus 14. Input / output (I / O) interfaces are also connected to the bus 14.
[0092] Multiple components in electronic device 10 are connected to I / O interface 15, including: input unit 16, such as keyboard, mouse, etc.; output unit 17, such as various types of displays, speakers, etc.; storage unit 18, such as disk, optical disk, etc.; and communication unit 19, such as network card, modem, wireless transceiver, etc. Communication unit 19 allows electronic device 10 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.
[0093] Processor 11 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 11 performs the various methods and processes described above, such as business processing methods.
[0094] In some embodiments, the business processing method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 10 via ROM 12 and / or communication unit 19. When the computer program is loaded into RAM 13 and executed by processor 11, one or more steps of the business processing method described above may be performed. Alternatively, in other embodiments, processor 11 may be configured to perform the business processing method by any other suitable means (e.g., by means of firmware).
[0095] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.
[0096] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.
[0097] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, RAM, ROM, erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.
[0098] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).
[0099] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.
[0100] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.
[0101] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.
[0102] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.
Claims
1. A business processing method, characterized in that, include: Obtain the user's business requirement description information, and parse the business requirement description information to obtain the user's business processing requirements under the target business type; Based on the prompt words and / or thought processes corresponding to the target business type, supplement the business processing requirements with requirement description data to obtain business processing related parameters. Based on the business processing requirements and the business processing associated parameters, a trained orchestration model is used to orchestrate the business processing flow and generate a dynamic service plugin invocation scheme. Based on the dynamic service plugin invocation scheme and the business processing association parameters, the corresponding service plugin is invoked to perform business processing under the target business type.
2. The method according to claim 1, characterized in that, Obtain the user's business requirement description information, including: Obtain the user's dialogue service content and use the dialogue service content as the business requirement description information; or... In the current business processing scenario, the user's next business processing instruction is identified, and in response to the user's confirmation operation of the next business processing instruction, the next business processing instruction is used as the business requirement description information.
3. The method according to claim 2, characterized in that, In the current business processing scenario, the user's next business processing instruction is identified, including: Obtain the user's business processing functions in the current business processing scenario; Based on the business processing function and the current business processing scenario, the user's next business processing instruction is identified.
4. The method according to claim 1, characterized in that, Based on the prompts and / or thought processes corresponding to the target business type, supplement the business processing requirements with requirement description data to obtain business processing association parameters, including: Based on the prompt word template, prompt word recommendation table, and prompt word evaluation results corresponding to the target business type, prompt words are generated; Based on the service workflow corresponding to the target business type, user adjustment and optimization results, and execution feedback evaluation results, a thought process is generated. Based on the prompt words and / or the thought process, the business processing requirements are supplemented with requirement description data to obtain business processing related parameters.
5. The method according to claim 1, characterized in that, Based on the business processing requirements and the associated parameters, a trained orchestration model is used to orchestrate the business processing flow, generating a dynamic service plugin invocation scheme, including: Based on the business processing associated parameters, retrieve business-related data from the knowledge base; Based on the business processing requirements, the business processing associated parameters, and the business associated data, the trained orchestration model is used to orchestrate the business processing flow and generate a dynamic service plugin invocation scheme.
6. The method according to claim 2, characterized in that, The business requirement description information is parsed to obtain the user's business processing requirements under the target business type, including: The information parsing granularity is determined based on the business requirement description information; The trained parsing model corresponding to the parsing granularity of the information is invoked to parse the business requirement description information and obtain the user's business processing requirements under the target business type.
7. The method according to claim 4, characterized in that, Based on the prompt words and / or the thought process, the business processing requirements are supplemented with requirement description data to obtain business processing related parameters, including: Extract the first attribute information that already exists in the business processing requirements; Based on the basic attribute information required for the inference of the orchestration model and the first attribute information, the missing second attribute information in the business processing requirements is determined; Based on the prompt words and / or the attribute information covered by the thought process, the second attribute information is supplemented with requirement description data to obtain business processing related parameters.
8. The method according to claim 1, characterized in that, Before invoking the corresponding service plugin to perform business processing under the target business type, the following is also included: The processing methods of business services under each business type are arranged to form service plugins; each service plugin includes the plugin service form, plugin scope of use, and plugin invocation method.
9. The method according to claim 8, characterized in that, Based on the dynamic service plugin invocation scheme and the business processing association parameters, the corresponding service plugin is invoked to perform business processing under the target business type, including: Based on the plugin invocation method corresponding to the currently invoked service plugin, determine the missing basic business data in the business processing associated parameters, and prompt the user to provide the basic business data; Based on the plugin service type, plugin usage scope, plugin invocation method, business processing associated parameters, and basic business data of the currently invoked service plugin, business processing is performed under the target business type.
10. The method according to claim 1, characterized in that, Also includes: Obtain user feedback data during business processing, and analyze the feedback data to obtain analysis results; Adjust the trained orchestration model based on the analysis results.
11. An electronic device, characterized in that, The electronic device includes: At least one processor; and a memory communicatively connected to said at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the business processing method according to any one of claims 1-10.
12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that cause a processor to execute the business processing method according to any one of claims 1-10.
13. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the business processing method according to any one of claims 1-10.