Service system dynamic interaction method based on AI
By setting up business processes and user interaction parts in the AI business system and combining large language models for information retrieval and response generation, the shortcomings of large models in existing technologies in meeting diverse task requirements are solved, and flexible configuration and efficient response are achieved.
Patent Information
- Application Number
- CN202510670275.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-23
- Publication Date
- 2025-09-16
AI Technical Summary
Existing large models have problems such as insufficient support for long-range conversations, limited reasoning capabilities, incomplete coverage of professional knowledge domain corpus, and lack of timeliness of training data when responding flexibly to diverse task requirements, making it difficult to meet users' diverse task requirements.
Provides an AI-based dynamic interaction method for business systems, including business process settings and user interaction parts. Through business permissions and flow settings, combined with a large language model, information retrieval and response generation are performed. It supports two types of response methods: unified response and detailed response, and flexibly configures business processes to meet user needs.
It enables flexible configuration of business processes based on actual scenarios, meets users' diverse task requirements, improves the flexibility and responsiveness of business systems, and enhances the ability to identify and process user intentions.
Smart Images
Figure CN120654811A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of computer technology, and in particular to an AI-based business system dynamic interaction method. Background Art
[0002] With the rapid digitalization of various industries, artificial intelligence systems are required to process massive amounts of information and make complex decisions across diverse tasks. However, existing large models are significantly limited in their ability to flexibly address diverse tasks due to issues such as insufficient support for long-range conversations, limited reasoning capabilities, incomplete coverage of specialized knowledge domains, and lack of timely training data. Summary of the Invention
[0003] In response to the problems existing in the prior art, the present invention provides an AI-based business system dynamic interaction method, which is flexibly configured according to different task requirements to meet the diverse task needs of users.
[0004] To achieve the above object, the technical solution adopted by the present invention is: An AI-based business system dynamic interaction method, which includes a business process setting part and a user interaction part. The business process setting section is used to set business permissions and business flow according to business needs; The business authority setting is to set whether the general business personnel can initiate the authority, including whitelist and blacklist settings, where whitelist personnel can initiate services, and blacklist personnel are restricted from initiating services; Business flow settings include service type, response method, unified reply content, and task details; The user interaction part includes the following steps: Step 1: User input: The user inputs questions and requirements; Step 2: Obtain the natural language description entered by the user and generate an HTTP request; Step 3: Verify the HTTP request to see if there are any applicable business processes for the current user. If no applicable business processes exist, the knowledge base search mechanism is activated to find matching information and an automatic reply is generated using the large language model. If applicable business processes exist, a list of applicable business processes is obtained. Step 4: Intent recognition is performed on the natural language description in the HTTP request. If the user intent cannot be accurately identified, a search mechanism is activated to find matching information and an automatic response is generated using a large language model. If the user intent is recognized, the business process that best meets the user intent is selected from the list of applicable business processes obtained in Step 3 and selected as the interactive business process. Step 5: Obtain detailed information about the interactive business process, perform targeted processing based on the detailed information, and provide an initial response for the user's reference by combining the large language model; Step 6: Wait for and collect user feedback on the initial reply, and adjust and improve the reply content based on the user's feedback; at the same time, generate an HTTP request based on the user's feedback; Step 7: Based on the Http request generated in step 6, the task scheduling function of the message distribution center is triggered, and the task input by the user in the Http request is assigned to the designated service center node, which then assigns the task to the corresponding service executor.
[0005] The response methods include unified response and detailed response. Unified response is executed as a single task by a single person, and directly flows to the next node after the user initiates it. Detailed response is used to deal with multiple options and multiple requirements, supporting single selection of multiple options and multiple selection of multiple options, and each option has a specific executor. The task details include the execution of each node from business initiation to flow, including the executor of each node and the message configuration template.
[0006] The step 5 is specifically as follows: for the business interaction process of the unified response type, a unified standardized processing method is adopted to provide unified reply content and guide the operation; for the business interaction process of the detailed response type, a unified reply content is provided and a selection button is provided.
[0007] During the process of setting up the business process, you will also configure the node flow that allows cancellation, pre-response, execution, post-review, and whether it can be dispatched.
[0008] The business flow settings also include message configuration, which includes classic message configuration and custom message configuration.
[0009] By adopting the above solution, the present invention provides a business process setup link, allowing users to flexibly configure corresponding general services based on actual scenarios to meet their diverse task requirements. Furthermore, the present invention sets two major categories for business flow settings: unified response type and detailed response type. These two settings allow for maximum flexibility in accessing various business processes. BRIEF DESCRIPTION OF THE DRAWINGS
[0010] Figure 1 Schematic diagram of the business flow of the present invention Figure 1 ; Figure 2 Schematic diagram of the business flow of the present invention Figure 2 ; Figure 3 A schematic diagram of setting the service authority of the present invention; Figure 4Part of the process for user interaction Figure 1 ; Figure 5 Part of the process for user interaction Figure 2 (catch Figure 4 ); Figure 6 Schematic diagram of the classic message setting of the present invention; Figure 7 This is a schematic diagram of customizing message settings in the present invention; Figure 8 This is a schematic diagram of the business process for applying for a seal according to an embodiment of the present invention; Figure 9 A schematic diagram of user interaction for applying for a seal according to an embodiment of the present invention; Figure 10 This is a schematic diagram of the repair service business process configuration according to an embodiment of the present invention; Figure 11 for Figure 10 The task details are displayed in . DETAILED DESCRIPTION
[0011] like Figure 1-Figure 5 As shown, the present invention discloses an AI-based business system dynamic interaction method, which includes a business process setting part and a user interaction part.
[0012] Among them, the business process setting part is to set business permissions and business flow according to business needs before use.
[0013] like Figure 1 and Figure 2 As shown, the business flow settings include service type, response method, unified reply content and task details settings.
[0014] The response method settings offer two main categories: unified response and detailed response. Unified response is a single-person, single-task operation, meaning that once initiated by the applicant, the process flows directly to the next node. For example, if a user initiates a "Reply a Repair" request and the "Daily Repair (Unified Response)" request is identified, the AI will guide the user to provide additional information based on the unified response, such as "Please tell me the specific details and problem of your repair request." Once the user completes the request, the general service request is completed, and the work order flow is established. Detailed response is used to address multiple options and requests, supporting single-select and multiple-select options, with each option assigned a specific person. For example, if a user initiates a car dispatch service and the "Car Dispatch Service" request is identified, the AI dialog will pop up a "Car Dispatch Service" button. After the user clicks it, the AI interface will first respond with the unified response, followed by the corresponding option buttons, such as "3-seater car," "5-seater car," and "7-seater car," along with descriptions of each option. After the user makes a selection, the work order flow moves to the next configured node.
[0015] All of the above processes can be configured with node flows such as cancellation permission, pre-response (approval / rejection), execution, post-review, and dispatchability.
[0016] like Figure 3 As shown, the business authority setting is to set whether general business personnel can initiate permissions, including whitelist and blacklist settings, where whitelist personnel can initiate services, and blacklist personnel are restricted from initiating services.
[0017] like Figure 6 As shown, the user interaction part specifically includes the following steps: Step 1: User input: The user inputs questions and requirements.
[0018] Specifically, users input questions and needs through natural language descriptions on the AI interaction platform.
[0019] Step 2: Obtain the natural language description of the user input and generate an HTTP request.
[0020] After receiving the questions and requirements input by the user, the AI interactive platform generates an HTTP request and sends the HTTP request to the server for processing.
[0021] Step 3: Verify the HTTP request to see if there are any business processes that the current user can apply for. If no business processes exist, start the knowledge base search mechanism to find matching information and generate an automatic reply using the large language model. If there are any business processes that can be applied for, obtain a list of applicable business processes.
[0022] After receiving the HTTP request, the server calls the business permission verification API for verification. If verification fails (i.e., no applicable business process exists), the server automatically responds using a large language model. For example, if the user's question is "I want to report a repair," and the knowledge base stores information such as the repair item and property manager, the AI will perform a semantic search to find relevant knowledge base content and provide an integrated response, such as providing the information and contact number of the person in charge of the side business.
[0023] Step 4: Intent recognition is performed on the natural language description in the HTTP request. If the user intent cannot be accurately identified, the knowledge base retrieval mechanism is activated to find matching information and an automatic reply is generated using the large language model. If the user intent is recognized, the business process that best meets the user intent is screened from the list of applicable business processes obtained in Step 3 based on the user intent and selected as the interactive business process.
[0024] In step 4, starting the knowledge base retrieval mechanism and the automatic reply generated by the large language model is the same as step 3.
[0025] Step 5: Obtain detailed information about the interactive business process, perform targeted processing based on the detailed information, and provide an initial response for the user's reference in combination with the large language model.
[0026] For unified response-type business interaction processes, a unified, standardized approach is adopted, providing unified response content and guiding operations. For example, if a user initiates a "Reply" request and is identified as "Daily Repair (Unified Response)", after the user selects "Initiate", the AI will guide the user to provide additional information based on the unified response content, such as "Please tell me the specific content and problem you want to report". Once the user completes the request, a general business report is completed, forming a work order flow.
[0027] For detailed response-type business interaction processes, unified reply content and selection buttons are provided. For example, if a user initiates a car dispatch service and the "car dispatch service" is recognized, the AI dialogue pops up a "car dispatch service" button. After the user clicks it, the AI page first responds with a unified reply content, and then pops up corresponding option buttons, such as "3-seater car", "5-seater car", "7-seater car", etc., and a description of each button option. After the user makes a selection, the work order is transferred to the next set node.
[0028] Step 6: Wait for and collect user feedback on the initial reply, and adjust and improve the reply content based on the user's feedback; at the same time, generate an HTTP request based on the user's feedback.
[0029] After collecting user feedback, the AI interactive platform generates an HTTP request and sends it to the server to synchronize the latest data and information.
[0030] Step 7: Based on the Http request generated in step 6, the task scheduling function of the message distribution center is triggered, and the task input by the user in the Http request is assigned to the designated service center node, which then assigns the task to the corresponding service executor.
[0031] After receiving the Http request, the server calls the service completion interface to assign the task.
[0032] like Figure 6 and Figure 7 As shown, the business flow settings also include message configuration, which includes classic message configuration and custom message configuration. Classic message configuration means that the node of each work order flow is preset to notify whom and to whom the work order should be transferred for processing; custom message users can set the message content, notification personnel, trigger conditions, whether it can be withdrawn, etc. for each node.
[0033] In summary, this invention provides a business process setup link, allowing users to flexibly configure corresponding general services based on actual scenarios to meet their diverse task needs. Furthermore, this invention sets two major categories for business flow settings: unified response type and detailed response type. These two settings allow for maximum flexibility in accessing various business processes.
[0034] In order to better illustrate the technical solution of the present invention, embodiments will be listed below for detailed description.
[0035] like Figure 8 and Figure 9 As shown, the business process sets up an application for the use of the seal of [Chen Group], but the user enters an application for the use of the seal of [Zheng Group] on the AI interactive platform, and the AI interactive platform generates an automatic reply based on the large language model.
[0036] For example Figure 10 and Figure 11 As shown in the figure, the business process settings include repair service and meal ordering service. The specific configuration is shown below.
[0037] Repair service (unified response); The unified reply content is: "Please tell me the name and location of the damaged item, and how can I report it for you?" Response method: unified response; Whether to select multiple options: No; The task details are (single): Assigned supervisor: A7; Planning docking group (direct contact person for work orders): A7; Assignable person: A8; Assignable: Yes; Process intervention: Post-processing review, pre-processing response; Optional content: Logistics and maintenance group.
[0038] The configuration content of the meal ordering service (detailed response) is: The unified reply content is: "What is your preferred work meal combination?" Response method: detailed response; Whether to select multiple options: No; The task details are (multiple): The assigned supervisor is: A7; Planning docking group (the direct docking person for the work order): A7; Assignable persons: A8, A6, A5; Whether assignable: Yes; Process intervention: No process access; Optional content: Work meal package A.
[0039] The assigned supervisor is: A7; Planning docking group (the direct docking person for the work order): A7; Assignable persons: A8, A6, A5; Whether assignable: Yes; Process intervention: No process access; Optional content: Work meal package B.
[0040] Personnel authority configuration: Xiao Ming (has the permission to report repair services and order meals), Xiao Hong (only has the permission to report repair services), and Xiao Wang (does not have any permission).
[0041] The user interaction part is as follows: Xiao Wang inputs “The classroom light is broken” on the AI interactive platform; The server's business permission verification interface is used for verification. If it is identified that Xiao Wang does not have any service permissions, the knowledge base is called to find similar knowledge and organize the answers. The AI answers that the person in charge of the classroom is xx, and xx can be contacted at xxxxxx.
[0042] Xiaohong enters "I want to order a work meal at noon today" on the AI interactive platform; The server calls the business permission verification interface for verification, recognizes that Xiaohong has the permission to report repairs, and proceeds to the next step of intent recognition. Intent recognition fails, and the user's question does not match the service provided. AI calls the knowledge base for an answer. For example, the answer here will be: "If you need to order food, you can go to the offline xx restaurant, the time is: xx".
[0043] Xiao Ming enters "I want to order a work meal" on the AI interactive platform; The server calls the business permission verification interface for verification, recognizes that Xiao Ming has [Repair Service, Meal Ordering Service], performs intent recognition, recognizes the Meal Ordering Service, and pops up the service activation button; When the user clicks on the meal ordering service, the AI interactive platform provides a pre-guidance prompt: "What is your preferred work meal combination?" and pops up the corresponding selection buttons: "Work Meal Package A, xxx" / "Work Meal Package B, xxx"; After the user selects Package A, the AI conducts a secondary guidance reply: "Please tell me the time and place you want to be delivered." The user says: "Technology Building at 3 pm." The AI replies that it has received the relevant information, and the work order is now complete. The process is carried out according to the preset message rules. For example: In advance response, after the executor receives the work order, a to-do will pop up in the dialogue window. Click the corresponding to-do to respond, for example: Work meal package A is available, I will deliver it on time / or Work meal package A is out, please re-initiate and select package B After the response is completed: After the executor actually delivers the meal, he can complete this to-do item. At this time, it will be transferred to the superior supervisor for review according to the message rules (dynamic configuration).
[0044] The above description is merely an embodiment of the present invention and does not limit the technical scope of the present invention. Therefore, any minor modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention are still within the scope of the technical solution of the present invention.
Claims
1. An AI-based business system dynamic interaction method, characterized by: Including business process setting part and user interaction part, The business process setting section is used to set business permissions and business flow according to business needs; The business authority setting is to set whether the general business personnel can initiate the authority, including whitelist and blacklist settings, where whitelist personnel can initiate services, and blacklist personnel are restricted from initiating services; Business flow settings include service type, response method, unified reply content, and task details; The user interaction part includes the following steps: Step 1: User input: The user inputs questions and requirements; Step 2: Obtain the natural language description entered by the user and generate an HTTP request; Step 3: Verify the HTTP request to see if there are any applicable business processes for the current user. If no applicable business processes exist, the knowledge base search mechanism is activated to find matching information and an automatic reply is generated using the large language model. If applicable business processes exist, a list of applicable business processes is obtained. Step 4: Intent recognition is performed on the natural language description in the HTTP request. If the user intent cannot be accurately identified, a search mechanism is activated to find matching information and an automatic response is generated using a large language model. If the user intent is recognized, the business process that best meets the user intent is selected from the list of applicable business processes obtained in Step 3 and selected as the interactive business process. Step 5: Obtain detailed information about the interactive business process, perform targeted processing based on the detailed information, and provide an initial response for the user's reference by combining the large language model; Step 6: Wait for and collect user feedback on the initial reply, and adjust and improve the reply content based on the user's feedback; at the same time, generate an HTTP request based on the user's feedback; Step 7: Based on the Http request generated in step 6, the task scheduling function of the message distribution center is triggered, and the task input by the user in the Http request is assigned to the designated service center node, which then assigns the task to the corresponding service executor.
2. The AI-based business system dynamic interaction method according to claim 1, characterized in that: The response methods include unified response and detailed response. Unified response is executed as a single task by a single person, and directly flows to the next node after the user initiates it. Detailed response is used to deal with multiple options and multiple requirements, supporting single selection of multiple options and multiple selection of multiple options, and each option has a specific executor. The task details include the execution of each node from business initiation to flow, including the executor of each node and the message configuration template.
3. The AI-based business system dynamic interaction method according to claim 2, characterized in that: The step 5 is specifically as follows: for the business interaction process of the unified response type, a unified standardized processing method is adopted to provide unified reply content and guide the operation; for the business interaction process of the detailed response type, a unified reply content is provided and a selection button is provided.
4. The AI-based business system dynamic interaction method according to claim 1, characterized in that: During the process of setting up the business process, you will also configure the node flow that allows cancellation, pre-response, execution, post-review, and whether it can be dispatched.
5. The AI-based business system dynamic interaction method according to claim 1, characterized in that: The business flow settings also include message configuration, which includes classic message configuration and custom message configuration.