A method, medium, and apparatus for seamless transaction handling in an interactive interface

Abstract

The application discloses a method, medium and device for seamlessly handling a service in an interactive interface, the method comprising: receiving a user service request and loading service rules containing a process state machine; guiding the user to input information according to a current process stage and dynamically generating a structured confirmation component after real-time verification; listening to user confirmation operations to constitute a confirmed information unit; querying the process state machine based on the confirmed information unit to determine a stage transition condition, and performing an interface reconstruction operation when the condition is met; when a final stage is reached and all required information fields are confirmed, automatically packaging and submitting a complete data request package, and displaying a handling result on the same interface. The application overcomes the process interruption problem caused by multi-interface jumping in traditional service handling, ensures the coherence of user operations and the consistency of the interface through state machine driving and dynamic interface reconstruction, and significantly improves the service handling efficiency and user experience.

Description

Technical Field

[0001] This application relates to the field of human-computer interaction technology, specifically to a method, medium, and device for seamlessly handling business within an interactive interface. Background Technology

[0002] In existing business processing systems, especially in complex scenarios such as government services and financial services, users often face problems such as complex business processes and poor interactive experience. For example, Chinese invention patent application with publication number CN120235583A discloses an intelligent guidance method for handling cases. Although it guides information collection by engaging in closed-domain dialogue with users through a large language model (LLM) and ultimately automatically populates the information into the business system, this solution still suffers from the inherent defect of interface switching. After providing information in the consultation and guidance robot interaction system, users still need to jump to the independent case handling interface of the business processing system for final confirmation and submission, resulting in a fragmented user experience.

[0003] More specifically, existing technologies have the following limitations: First, dialogue guidance and business processing are in different interactive environments, and the information provided by users during the dialogue phase cannot receive continuous and coherent visual feedback and confirmation in the business processing interface; second, the process advancement lacks a refined state management mechanism, and cannot dynamically adjust the interface content and interaction methods according to the completeness of the information confirmed by the user; third, the interface reconstruction is disconnected from the process transition, resulting in users having to face multiple isolated operation interfaces during the process, increasing operational complexity and cognitive burden.

[0004] Therefore, there is an urgent need for a technical solution that can achieve seamless integration of the entire business processing process in a single interactive interface. Through unified process status management and dynamic interface reconstruction mechanism, the continuity of user operations can be ensured, unnecessary interface jumps can be avoided, and the efficiency of business processing and user experience can be comprehensively improved. Summary of the Invention

[0005] In view of the above problems, this application provides a technical solution for seamless business processing in an interactive interface, which solves the technical problems of rigid interactive processes, frequent interface switching leading to fragmented user experience, and low operation efficiency in existing business processing systems.

[0006] To achieve the above objectives, in a first aspect, this application provides a method for seamlessly handling business within an interactive interface, the method comprising the following steps:

[0007] S1: Receive the user's business request through the interactive interface, determine the target business type, and load the corresponding business processing rules. The business processing rules include a list of required information fields and a process state machine with multiple predefined ordered process stages.

[0008] S2: Based on the current process stage, guide the user to input information through at least one information input component, and verify the information entered by the user in real time. The input information is used to fill the corresponding information field. Whenever an information field is verified, a structured confirmation component is dynamically generated and rendered in the interactive interface to receive the user's final confirmation operation for the information field.

[0009] S3: Listen for the user's confirmation operation on the structured confirmation component, wherein one or more information fields confirmed by the user constitute a confirmed information unit;

[0010] S4: Based on all currently confirmed information units, query the process state machine to determine whether the conditions for transitioning from the current process stage to the subsequent target process stage are met. If the conditions are met, perform an interface reconstruction operation. The interface reconstruction operation includes: clearing all information input components and structured confirmation components rendered in the current process stage, and rendering information input components corresponding to one or more information fields required by the target process stage according to the definition of the target process stage.

[0011] S5: When the process state machine transitions to the final stage and all necessary information fields are included in the confirmed information units, it automatically encapsulates a complete data request package that conforms to the interface specifications of the external business system based on all confirmed information units.

[0012] S6: Submit the complete data request package by calling the backend application interface of the external business system, and display the business processing result returned by the external business system in the same interactive interface.

[0013] Furthermore, step S4 also includes the following steps:

[0014] S41: Based on all the currently confirmed information units, perform real-time analysis using a pre-trained intent recognition model to predict the user's potential related business intent;

[0015] S42: Query the business knowledge graph to obtain the target business processing rules and entry interfaces corresponding to the potential related business intent;

[0016] S43: When performing the interface reconstruction operation, synchronize the information input components required for the target process stage in the interactive interface, and render at least one quick entry component corresponding to the potential related business intent; wherein, the quick entry component encapsulates the target business processing rules and entry interface, and is pre-filled with reusable information related to the target business extracted from all currently confirmed information units;

[0017] S44: In response to the user's trigger operation on the shortcut entry component, the pre-filled reuse information is used as the initial input, and based on the target business processing rules, a new target business processing flow instance is initialized and started in the current interactive interface.

[0018] Furthermore, after step S1 and before step S2, the following steps are also included:

[0019] S11: Query a locally encrypted cross-business information unit library, which is used to store the confirmed information units that have been authorized and cached after the user has completed a business in the past;

[0020] S12: Automatically fill the corresponding information input component with the confirmed information unit that matches the required information field in the current business processing rules as pre-filled information, and initialize the state of the information input component to be verified and confirmed.

[0021] Furthermore, step S4 also includes:

[0022] S401: Based on all currently confirmed information units, calculate the overall completeness of all confirmed information units relative to all required information fields;

[0023] S402: Query the process state machine to obtain multiple optional transition paths that match the overall completeness, wherein each optional transition path corresponds to a different subsequent target process stage;

[0024] S403: Render a path selection interface in the interactive interface to show the user the multiple optional jump paths;

[0025] S404: In response to the user's selection operation in the path selection interface, determine the final target process stage, and then perform the interface reconstruction operation.

[0026] Furthermore, in step S3, when any information unit is marked as confirmed, the method also includes a predictive resource preloading process, as follows:

[0027] S301: Based on the current process stage, all confirmed information units, and the business processing rules, a pre-trained graph neural network model is used to predict at least one backend service interface whose probability of being called in subsequent process steps exceeds a preset threshold in the current context.

[0028] S302: Initiate an asynchronous call to the backend service interface predicted in step S301 in the background; the call will carry the necessary parameters extracted from the currently confirmed information unit and cache the response data returned by the called backend service interface in memory;

[0029] S303: When the process state machine advances to a new process stage, and the business processing rules under this process stage need to call a certain backend service interface, first check whether there is a preloaded result of the backend service interface in the cache; if it exists, directly use the cached data for interface rendering.

[0030] Furthermore, in step S2, when a user enters content in an information input component but does not trigger a confirmation operation, the method further includes the following steps:

[0031] S201: Capture unconfirmed input values ​​in all current information input components in real time, map the input values ​​to the business processing rules in the current state of the process state machine, and identify the logical gap between the input but unconfirmed information and the final business rules;

[0032] S202: Based on the identified logical gaps, a counterfactual reasoning engine is used to perform multi-path deduction to determine the optimal deduction path. The counterfactual reasoning engine is based on a constraint satisfaction problem model that is isomorphic to the business processing rules. It is used to generate one or more complete information unit combinations containing virtual information that can satisfy all business rules. Each information unit combination represents a feasible deduction path from the current incomplete or non-compliant state to the final compliant state.

[0033] S203: Transform at least one optimal derivation path into a guiding suggestion and display the sequence of operations required to achieve compliance to the user through the interactive interface. The sequence of operations includes supplementing a specific information unit, modifying the value of an input information unit, or selecting one or more of a specific enumeration option.

[0034] Furthermore, after determining the target business type in step S1, and before initializing the interactive interface according to the target business type and the current process stage in step S2, the method further includes the following steps:

[0035] S101: Receive the user's initial service request through a multimodal interaction interface, which integrates a speech recognition unit, an image recognition unit, and an emotion computing unit; the speech recognition unit and the image recognition unit are used to convert the user's speech input information and image input information into text or structured data, respectively; the emotion computing unit is used to analyze the acoustic features and text emotion during the user's input process and output emotion state labels.

[0036] S102: Based on the emotional state tags and the completeness of the initial information provided by the user through the multimodal interaction interface, dynamically determine the guidance strategy for subsequent interaction interfaces, specifically including:

[0037] If the emotional state label is confused or anxious, the decision-making guidance strategy is a simplified guidance model;

[0038] If the image recognition unit has extracted some of the data required by the information unit, the guidance strategy determined is the information pre-filling mode;

[0039] S103: When initializing the interactive interface in step S2, the presentation logic of the interface components is adaptively adjusted according to the determined guidance strategy, specifically including:

[0040] When the guidance strategy is a simplified guidance mode, the generated interface will focus on displaying the core required information units at the current stage of the process, and will use step-by-step, enhanced guidance interactive text.

[0041] When the guidance strategy is in the information pre-fill mode, the generated interface will automatically fill the corresponding information input component with the data extracted through image recognition and set its UI status to await user confirmation.

[0042] Furthermore, after determining the target business type in step S1, and before initializing the interactive interface based on the target business type and the current process stage, the method further includes the following steps:

[0043] S110: Collect contextual information for decision-making, including sentiment state tags, historical interaction preferences and intent recognition results related to the current user and target business type, and the sentiment state tags are obtained by analyzing acoustic features and text sentiment during the user input process;

[0044] S111: Based on the context information, an adaptive decision engine calculates and generates an interaction strategy suitable for the current process stage. This interaction strategy is used to dynamically modulate the attributes of the information input components in step S2, specifically including:

[0045] Based on users' emotional state tags, historical interaction preferences, and intent recognition results, the presentation order of information units, as well as the level of detail and style of the guiding copy, are determined.

[0046] Based on the user's historical operation accuracy and current emotional state label, determine the fault tolerance threshold and prompt strength for real-time verification of user input;

[0047] Determine the tone and style of interface feedback prompts based on emotional state tags;

[0048] S112: When initializing the interactive interface in step S2, the initialization logic of the information input component is adaptively configured according to the interaction strategy generated in step S111.

[0049] In a second aspect, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for seamlessly handling business in an interactive interface as described in the first aspect of this application.

[0050] In a third aspect, this application provides an electronic device having a computer program stored thereon, including a processor and a storage medium, wherein the computer program is stored on the storage medium, and when executed by the processor, the computer program implements the method for seamlessly handling business in an interactive interface as described in the first aspect of this application.

[0051] Unlike existing technologies, the above-mentioned technical solution involves a method for seamlessly handling business within an interactive interface. By introducing a collaborative mechanism between a process state machine and a structured confirmation component, it achieves continuity in the business handling process and seamless interface interaction. The method includes: receiving user business requests and loading business rules containing a process state machine; guiding the user to input information based on the current process stage and dynamically generating a structured confirmation component after real-time verification; monitoring user confirmation operations to form confirmed information units; querying the process state machine based on the confirmed information units to determine stage transition conditions, and executing interface reconstruction operations when the conditions are met; when the final stage is reached and all necessary information fields are confirmed, automatically encapsulating and submitting a complete data request package, and displaying the processing result on the same interface. This invention overcomes the process interruption problem caused by multiple interface jumps in traditional business handling. Through state machine-driven and dynamic interface reconstruction, it ensures the continuity of user operations and interface consistency, significantly improving business handling efficiency and user experience.

[0052] The above description of the invention is merely an overview of the technical solution of this application. In order to enable those skilled in the art to better understand the technical solution of this application and to implement it based on the description and drawings, and to make the above-mentioned objectives and other objectives, features and advantages of this application easier to understand, the following description is provided in conjunction with the specific embodiments and drawings of this application. Attached Figure Description

[0053] The accompanying drawings are only used to illustrate the principles, implementation methods, applications, features, and effects of specific embodiments of this application and other related content, and should not be considered as limitations on this application.

[0054] In the accompanying drawings of the instruction manual:

[0055] Figure 1 A flowchart of a method for seamlessly handling business in an interactive interface as described in the first exemplary embodiment of this application;

[0056] Figure 2 A flowchart illustrating the method for seamlessly handling business in an interactive interface as described in the second exemplary embodiment of this application;

[0057] Figure 3 A flowchart illustrating the method for seamlessly handling business in an interactive interface as described in the third exemplary embodiment of this application;

[0058] Figure 4 This is a flowchart illustrating the method for seamlessly handling business in an interactive interface as described in the fourth exemplary embodiment of this application;

[0059] Figure 5 A flowchart illustrating the method for seamlessly handling business in an interactive interface as described in the fifth exemplary embodiment of this application;

[0060] Figure 6 A flowchart illustrating the method for seamlessly handling business in an interactive interface as described in the sixth exemplary embodiment of this application;

[0061] Figure 7 A flowchart illustrating the method for seamlessly handling business in an interactive interface as described in the seventh exemplary embodiment of this application;

[0062] Figure 8 This is a flowchart illustrating the method for seamlessly handling business in an interactive interface as described in the eighth exemplary embodiment of this application;

[0063] Figure 9 This is a schematic diagram of an electronic device according to an exemplary embodiment of this application;

[0064] The reference numerals used in the above figures are explained as follows:

[0065] 10. Electronic equipment; 101. Processor; 102. Storage medium. Detailed Implementation

[0066] To explain in detail the possible application scenarios, technical principles, specific feasible solutions, and the objectives and effects that this application can achieve, the following detailed description is provided in conjunction with the listed specific embodiments and accompanying drawings. The embodiments described herein are only used to more clearly illustrate the technical solutions of this application, and are therefore only examples, and should not be used to limit the scope of protection of this application.

[0067] like Figure 1 As shown, in a first aspect, this application provides a method for seamlessly handling business within an interactive interface, the method comprising:

[0068] S1: Receive the user's business request through the interactive interface, determine the target business type, and load the corresponding business processing rules. The business processing rules include a list of required information fields and a process state machine with multiple predefined ordered process stages.

[0069] S2: Based on the current process stage, guide the user to input information through at least one information input component, and verify the information entered by the user in real time. The input information is used to fill the corresponding information field. Whenever an information field is verified, a structured confirmation component is dynamically generated and rendered in the interactive interface to receive the user's final confirmation operation for the information field.

[0070] S3: Listen for the user's confirmation operation on the structured confirmation component, wherein one or more information fields confirmed by the user constitute a confirmed information unit;

[0071] S4: Based on all currently confirmed information units, query the process state machine to determine whether the conditions for transitioning from the current process stage to the subsequent target process stage are met. If the conditions are met, perform an interface reconstruction operation. The interface reconstruction operation includes: clearing all information input components and structured confirmation components rendered in the current process stage, and rendering information input components corresponding to one or more information fields required by the target process stage according to the definition of the target process stage.

[0072] S5: When the process state machine transitions to the final stage and all necessary information fields are included in the confirmed information units, it automatically encapsulates a complete data request package that conforms to the interface specifications of the external business system based on all confirmed information units.

[0073] S6: Submit the complete data request package by calling the backend application interface of the external business system, and display the business processing result returned by the external business system in the same interactive interface.

[0074] In this embodiment, the interactive interface refers to the carrier that supports information interaction between the user and the system. It can be a chat window, a web page interface, etc. It is compatible with multimodal input such as text, voice, and images, and can complete the entire business process without jumping to a third-party page.

[0075] Business processing rules refer to the pre-defined processing basis for different business types, including a list of required information fields (such as ID number, purpose, etc. required for applying for a certificate of no criminal record) and a process state machine (a logical model that predefines multiple ordered process stages and transition conditions between stages).

[0076] Information input components refer to interface elements used to receive user information input, such as text input boxes, number input boxes, radio button groups, file upload boxes, etc., and their types match the attributes of the information fields to be input.

[0077] The structured confirmation component refers to the interface component that is dynamically generated after the information field passes real-time verification. It displays the verified information in a structured form and includes interactive elements such as "Confirm" and "Modify" to allow users to complete the final confirmation or initiate modifications.

[0078] An information unit refers to a collection of information consisting of one or more user-confirmed information fields. For example, the "basic enterprise information" unit may contain confirmed fields such as enterprise type and years of establishment.

[0079] The process state machine predefines the logic engine of multiple ordered process stages, stores the process stages and stage transition conditions corresponding to different information combinations, and can determine whether to trigger a process stage jump based on the confirmed information unit.

[0080] A complete data request package refers to a data set that conforms to the interface specifications of an external business system. It is composed of all confirmed information units encapsulated in the format required by the interface and can be directly submitted to the external business system for business processing.

[0081] Backend application programming interfaces (APIs) refer to interfaces provided by external business systems for receiving business requests and returning processing results. The system submits business data and obtains processing results by calling these interfaces.

[0082] In step S1, the user initiates a business request through an interactive interface (such as a chat window), for example, "I want to apply for a certificate of no criminal record for going abroad." The system first performs intent recognition and semantic analysis on the request to determine that the target business type is "application for a certificate of no criminal record." Subsequently, the system loads the corresponding business processing rules from the database or knowledge base. The list of required information fields includes ID number, name, application purpose, and photos of the front and back of the ID card; the process state machine predefines an ordered process stage, including the initial stage (requiring collection of ID number and ID card photo), the information verification stage (verifying the consistency of ID card information), and the final submission stage (information encapsulation and submission).

[0083] In step S2, the system renders corresponding information input components in the interactive interface according to the current process stage (such as the initial stage), such as a text input box for entering the ID card number and a file upload box for uploading ID card photos. It also guides the user to input information through prompts (such as "Please enter your ID card number and upload photos of the front and back of your ID card"). After the user enters information, the system verifies the input in real time, for example, verifying whether the ID card number format conforms to the 18-bit encoding rule and whether the ID card photo is clear and contains complete front and back information. If a certain information field (such as the ID card number) passes verification, the system immediately dynamically generates and renders a structured confirmation component in the interactive interface, such as "ID Card Number: 3501********1234 [Confirm] [Modify]", for the user to confirm the information in that field.

[0084] In step S3, the system continuously monitors the user's actions on the structured confirmation components. When the user clicks the "Confirm" button on a structured confirmation component, the corresponding information field is marked as confirmed. If multiple information fields (such as ID number and name) are confirmed successively, these multiple confirmed information fields together constitute a confirmed information unit (such as a "personal identity information unit").

[0085] In step S4, the system queries the process state machine based on all currently confirmed information units to determine whether the process stage transition conditions are met. For example, when the "personal identity information unit" (including ID number, name, and ID photo) has been confirmed, and the condition for transitioning from the "initial stage" to the "information prefilling stage" in the process state machine is "complete confirmation of personal identity information unit", the system determines that the transition conditions are met. Subsequently, the interface reconstruction operation is performed: the ID number input box, ID photo upload box, and corresponding structured confirmation components rendered in the initial stage of the current interactive interface are cleared. According to the definition of the "information prefilling stage", the information input components required for this stage are rendered, such as the application purpose radio button group ("going abroad", "employment", etc.).

[0086] In step S5, when the process state machine transitions to the final stage (such as the "submission preparation stage"), and the system detects that all necessary information fields (ID number, name, application purpose, ID photo) are included in the confirmed information units, the system automatically encapsulates all confirmed information units in the format required by the external business system (such as the government service platform) interface to generate a complete data request package, which includes user ID, business type identifier, standardized data of each information field, etc.

[0087] In step S6, the system submits a complete data request package to the external business system by calling its backend application programming interface. After processing the request, the external business system returns the business processing result (e.g., "Application submitted successfully, business serial number: 202511050001"). Upon receiving the result, the system displays it to the user in natural language or structured format on the same interactive interface, completing the entire seamless business processing workflow.

[0088] Taking a company applying for "R&D subsidy" type of support funds as an example, the method involved in this embodiment specifically includes the following steps:

[0089] In step S1, the user enters "apply for enterprise support funds" in the interactive interface of the government service platform. The system recognizes the target business type as "enterprise support fund application" and then loads the business processing rules. The list of required information fields is A (applicant identity), B (years of establishment), C (last year's revenue), D (type of support applied for), and E (R&D project plan). The process state machine includes the stages and transition rules of "initial state (requires A) → state 2 (requires B / C / D) → state 3 - R&D (requires E) → final state (pending submission)".

[0090] In step S2, the system is in its initial state (A needs to be collected). The interface displays a "Applicant Identity" radio button group (options: Startup / SME / Large Enterprise) and prompts "Please select your enterprise type". After the user selects "Startup", the system verifies the user's selected information in real time: first, confirming that the option is within the preset range; second, verifying the user's enterprise establishment period is ≤5 years (meeting the definition of a startup) by referring to the enterprise database. After the verification is successful, the interface dynamically generates a structured confirmation component "Enterprise Type: Startup [Confirm] [Modify]".

[0091] In step S3, the user clicks the "Confirm" button, and the system marks "Applicant Identity: Startup" as confirmed. This field constitutes a separate "Enterprise Identity Information Unit" and is temporarily stored in the context memory module.

[0092] In step S4, the system queries the process state machine and finds that the "Enterprise Identity Information Unit (A = Startup)" meets the transition condition of "Initial State → State 2". Then, it performs interface reconstruction: clears the "Applicant Identity" radio button group and confirmation component, and renders the three information input components required for State 2 at once - "Years of Establishment" numeric input box (prompt "Please enter a positive integer"), "Previous Year's Revenue (Ten Thousand Yuan)" numeric input box (prompt "Please enter Arabic numerals"), and "Application Support Type" radio button group (options can be set as: R&D subsidy / rent reduction / talent introduction reward), and prompts "Please fill in the company's years of establishment, previous year's revenue, and select the support type".

[0093] If the system receives the user's information: "Years of Establishment: 2", "Revenue: 300", selects "R&D Subsidy", and clicks "Confirm the above information", the system verifies the validity of the information (years of establishment are positive integers, revenue is in the correct format, and the support type is legal). A unified confirmation component is generated and confirmed by the user, forming the "Enterprise Basic and Demand Information Unit". The system then queries the state machine again. Because "D = R&D Subsidy" satisfies the transition condition of "State 2 → State 3 - R&D", a second interface reconstruction is performed: all components of State 2 are cleared, and only the "R&D Project Plan" file upload box is rendered (with the message "Supports PDF / Word format, size not exceeding 20MB").

[0094] In step S5, the user uploads and confirms the R&D project plan. The system collects the "R&D Material Information Unit (E = R&D Project Plan [File Number])" and detects that all required information fields (A / B / C / D / E) have been confirmed. The process state machine then transitions to the final stage. Following the interface requirements of the government service platform, the system encapsulates the three information units into a complete data request package. The "Applicant Identity" is mapped to "enterprise_type: Startup," and the "R&D Project Plan" is converted to Base64 encoded "project_plan_file: [Encoded Content]."

[0095] In step S6, the system calls the "Enterprise Support Fund Application Submission Interface" of the government service platform to submit a complete data request package. After the external system verifies the data, it returns the result "Application submitted successfully, business serial number: 20241115000123, preliminary review expected to be completed within 3 working days". The system displays this result in the form of a "notification card" on the same interactive interface and provides a "Check Progress" button to complete the entire process.

[0096] By employing the above methods, the entire user transaction process can be completed within a single interactive interface, eliminating the need to navigate to third-party pages and avoiding frequent switching between customer service and business systems, significantly reducing operational steps and processing time. Information input components guide user input, while structured confirmation components clearly display verified information, lowering the familiarity requirement for users with the business process. Even users unfamiliar with the process can quickly complete the operation, while avoiding operational confusion caused by page redirects. A real-time verification mechanism promptly detects user input errors (such as incorrectly formatted ID numbers), and the structured confirmation component further ensures user acceptance of the information, reducing business failures and rework due to incorrect information, thus mitigating business risks. The process state machine can flexibly configure process stages and transition conditions according to different business types, adapting to diverse business scenarios. Furthermore, the interface reconstruction operation can dynamically adjust interface elements according to process progression, giving the system strong flexibility and scalability.

[0097] In some embodiments, such as Figure 2 As shown, when performing step S4, the following steps are also included:

[0098] S41: Based on all the currently confirmed information units, perform real-time analysis using a pre-trained intent recognition model to predict the user's potential related business intent;

[0099] S42: Query the business knowledge graph to obtain the target business processing rules and entry interfaces corresponding to the potential related business intent;

[0100] S43: When performing the interface reconstruction operation, synchronize the information input components required for the target process stage in the interactive interface, and render at least one quick entry component corresponding to the potential related business intent; wherein, the quick entry component encapsulates the target business processing rules and entry interface, and is pre-filled with reusable information related to the target business extracted from all currently confirmed information units;

[0101] S44: In response to the user's trigger operation on the shortcut entry component, the pre-filled reuse information is used as the initial input, and based on the target business processing rules, a new target business processing flow instance is initialized and started in the current interactive interface.

[0102] In this embodiment, the intent recognition model refers to an algorithm model trained based on deep learning (such as the BERT model), which can perform semantic analysis and intent mining on the information units confirmed by the user, and predict the potential business needs that the user may have that are related to the current business.

[0103] A business knowledge graph is a knowledge base that stores entities (such as business types and information fields), entity attributes, and relationships between entities within a business domain. For example, there is a relationship between "certificate of no criminal record" and "passport application," and this relationship and the corresponding business processing rules can be obtained from the knowledge graph.

[0104] Quick access components are UI elements that are rendered synchronously during UI reconstruction. They encapsulate the processing rules and entry interfaces for potential related services and are pre-populated with reusable information extracted from currently confirmed information units (such as the ID number confirmed for the current service can be reused for related services). Users can quickly start processing related services by triggering this component.

[0105] A business processing instance refers to an independent processing process generated for a specific business. It includes data such as the process state machine, information input components, and confirmed information units corresponding to that business. Process instances of different businesses are independent of each other and do not interfere with each other.

[0106] When executing step S4 (process status judgment and interface reconstruction), the system first obtains all currently confirmed information units, such as the "ID number, name, and purpose of travel" information units confirmed when applying for a "certificate of no criminal record". These information units are then input into a pre-trained intent recognition model. The model analyzes the key information (such as "purpose of travel") and semantic relationships within the information units to predict the user's potential related business intent, such as "passport application" or "visa application".

[0107] The system then queries the business knowledge graph and extracts the target business processing rules (such as a list of required information fields: ID number, name, photo, application form, etc.; process state machine: information collection stage, material review stage, etc.) and entry interfaces (such as the API address and request format of the passport processing system) based on the predicted potential related business intents (such as "passport processing").

[0108] When performing interface reconstruction operations (clearing current stage components and rendering target stage input components), the system simultaneously renders quick access components corresponding to potentially related business intents in the interactive interface. Taking "passport application" as an example, the quick access component can be displayed as "Do you need to apply for a passport? Click here to apply quickly." This component encapsulates the business rules and entry interface for passport application and is pre-populated with reusable information extracted from currently confirmed information units (such as ID number and name).

[0109] When a user clicks a shortcut component (such as "Click here to apply quickly"), the system responds by triggering an operation. The pre-filled reusable information in the component is used as the initial input for the related service (passport application), automatically filling the corresponding information input component in the passport application process. Simultaneously, based on the processing rules for the target service (passport application), the system initializes and starts a passport application process instance in the current interface, directly entering the initial stages of the passport application process (such as supplementing photos and filling out the application form), without requiring the user to re-initiate the service request or repeatedly enter the already provided information.

[0110] The above solution predicts potential related business intentions through an intent recognition model, proactively providing users with related business options and improving the initiative and accuracy of services. The quick access component pre-fills reusable information, eliminating the need for users to repeatedly enter confirmed information when initiating related businesses. Furthermore, it seamlessly opens new business processes within the same interactive interface, enhancing the continuity of business processing and further saving user time.

[0111] In some embodiments, such as Figure 3 As shown, after step S1 and before step S2, the following steps are also included:

[0112] S11: Query a locally encrypted cross-business information unit library, which is used to store the confirmed information units that have been authorized and cached after the user has completed a business in the past;

[0113] S12: Automatically fill the corresponding information input component with the confirmed information unit that matches the required information field in the current business processing rules as pre-filled information, and initialize the state of the information input component to be verified and confirmed.

[0114] In this embodiment, the cross-business information unit library refers to a locally encrypted database used to store confirmed information units that have been authorized and cached after a user has completed a business in the past. This database uses encryption algorithms (such as AES encryption) to ensure data security and is only used for information pre-filling for subsequent businesses with the user's authorization.

[0115] Pre-filled information refers to confirmed information unit data retrieved from the cross-business information unit database that matches the required information fields for the current business, and can be directly filled into the information input component of the current business.

[0116] After step S1 (request reception and rule loading) is completed and before step S2 (information guidance input) begins, the system first queries the locally encrypted cross-business information unit database. During the query, the current user identifier (such as user ID) and the list of required information fields for the current business are used as search conditions. For example, when applying for a "certificate of no criminal record", the system searches for confirmed information units in the user's historical authorization cache that contain fields such as "ID number and name".

[0117] The system will automatically populate the corresponding information input components (such as the ID number and name input boxes) with confirmed information units that match the required information fields of the current business (e.g., "ID number: 3501********1234, Name: Zhang San") as pre-filled information. Simultaneously, the system will initialize the status of these information input components to "verified and confirmed," meaning that the pre-filled information is assumed to meet the verification rules (e.g., the ID number format is correct), and the user does not need to click the "confirm" button; it is directly considered that the user has confirmed the information. In subsequent processes, this pre-filled information can be directly incorporated into the information units for process status judgment.

[0118] The above solution, by pre-filling historical authorization information, eliminates the need for users to repeatedly enter information about previously processed transactions, significantly reducing input operations. This is especially beneficial for users who frequently process transactions, as it can substantially shorten processing times. The cross-business information unit database employs local encrypted storage to avoid the risk of information leakage. Furthermore, information pre-filling is based on user authorization, ensuring users' control over and right to know their own information, thereby enhancing user trust in the system.

[0119] In some embodiments, such as Figure 4 As shown, step S4 further includes:

[0120] S401: Based on all currently confirmed information units, calculate the overall completeness of all confirmed information units relative to all required information fields;

[0121] S402: Query the process state machine to obtain multiple optional transition paths that match the overall completeness, wherein each optional transition path corresponds to a different subsequent target process stage;

[0122] S403: Render a path selection interface in the interactive interface to show the user the multiple optional jump paths;

[0123] S404: In response to the user's selection operation in the path selection interface, determine the final target process stage, and then perform the interface reconstruction operation.

[0124] In this embodiment, the overall completeness refers to the proportion of confirmed information units that cover all necessary information fields. The calculation formula is "number of confirmed information fields / total number of necessary information fields × 100%", which is used to measure the progress of information completion in the current business process.

[0125] Optional transition paths refer to the paths in the process state machine that match the current overall completeness and can jump from the current process stage to different subsequent target process stages. Each path corresponds to a different direction of business processing.

[0126] The path selection interface is an interface used to display multiple optional transition paths to users. It includes descriptions of the target process stages corresponding to each path (such as "Path 1: First supplement the application purpose, then complete other information", "Path 2: First upload materials, then supplement basic information") and selection controls (such as radio buttons).

[0127] In step S401, the system first counts the number of information fields contained in the currently confirmed information units, and then, combined with the total number of required information fields for the current business, calculates the overall completeness using the formula: "Overall completeness = Number of confirmed information fields / Total number of required information fields × 100%". For example, if a certain business requires 5 required information fields, and 3 have been confirmed so far, the overall completeness is 60%.

[0128] In step S402, the system queries the process state machine and, based on the calculated overall completeness (e.g., 60%), filters out multiple optional transition paths that match this completeness. Assuming the process state machine specifies that when the overall completeness reaches 50%-70%, the optional transition paths include "Path A: Jump to the priority supplementation of core information stage (requires supplementation of core field 1 out of the remaining 2 fields)" and "Path B: Jump to the balanced supplementation of information stage (can supplement any 1 of the remaining 2 fields)", then the system extracts these two paths as optional transition paths.

[0129] In step S403, the system renders a path selection interface in the interactive interface. The interface clearly displays the target process stage name, the information fields that need to be supplemented in the stage, and the advantages of the stage (such as "Path A: Prioritize supplementing core information and quickly enter the review stage") for each optional jump path. Radio buttons and other selection controls are configured for each path so that users can choose their preferred jump path.

[0130] In step S404, the user clicks the selection control corresponding to a path (e.g., selecting path A) in the path selection interface. The system responds to this selection operation and determines the final target process stage as the "priority core information supplementation stage". Subsequently, the system performs an interface reconstruction operation according to the definition of the target process stage, clears the components of the current stage, and renders the information input components required for the "priority core information supplementation stage" (e.g., the input box corresponding to core field 1).

[0131] The above design allows users to choose their preferred workflow path based on their needs and operating habits, breaking away from the traditional "one-size-fits-all" workflow model and enhancing user participation and control during the business process. Different users have different priorities in business processing (e.g., some users want to quickly advance to the review stage, while others prefer flexible information supplementation). The selectable pathways can meet diverse needs, enhancing the system's adaptability to different user groups. Providing multiple path options based on overall completeness ensures that workflow progression is no longer limited to a single fixed logic, but can be dynamically adjusted according to user choices, improving the flexibility and personalization of business processing.

[0132] In some embodiments, in step S3, when any information unit is marked as confirmed, such as Figure 5 As shown, the method also includes a predictive resource preloading process, as detailed below:

[0133] S301: Based on the current process stage, all confirmed information units, and the business processing rules, a pre-trained graph neural network model is used to predict at least one backend service interface whose probability of being called in subsequent process steps exceeds a preset threshold in the current context.

[0134] S302: Initiate an asynchronous call to the backend service interface predicted in step S301 in the background; the call will carry the necessary parameters extracted from the currently confirmed information unit and cache the response data returned by the called backend service interface in memory;

[0135] S303: When the process state machine advances to a new process stage, and the business processing rules under this process stage need to call a certain backend service interface, first check whether there is a preloaded result of the backend service interface in the cache; if it exists, directly use the cached data for interface rendering.

[0136] Predictive resource preloading refers to a process in which, after an information unit is confirmed, the system anticipates the backend service interfaces that may be called in subsequent processes and initiates calls and caches data asynchronously. The purpose is to reduce the waiting time for interface calls in subsequent processes and improve the efficiency of interface rendering.

[0137] Graph neural network models are deep learning models trained on business scenario data. They can combine contextual information such as the current process stage, confirmed information units, and business processing rules to predict which backend service interfaces will be called in subsequent steps with a probability exceeding a preset threshold (such as 80%).

[0138] Backend service interfaces refer to interfaces that provide data support or functional services for business processing, such as identity verification interfaces, material compliance verification interfaces, and business progress query interfaces, which need to be called through parameters to obtain response data.

[0139] Asynchronous call initiation means that the system initiates an interface call request in the background without blocking the current user interaction. The user can continue to complete subsequent input, confirmation and other actions, avoiding interaction lag caused by waiting for interface calls.

[0140] Cached data refers to temporarily storing the response data returned by the backend service interface in the system memory. When subsequent processes need to use this data, it can be read directly without initiating another interface call.

[0141] In step S301, when any information unit is marked as confirmed, the system takes the current process stage, all confirmed information units, and corresponding business processing rules as input and feeds them into a pre-trained graph neural network model. The model analyzes contextual relationships (such as the strong correlation between "R&D subsidy type application" and "project plan compliance verification") to predict backend service interfaces whose call probability exceeds a preset threshold in subsequent process steps, thus identifying the target interfaces that need to be pre-loaded.

[0142] In step S302, the system asynchronously initiates a call to the predicted backend service interface in the background, carrying necessary parameters extracted from the currently confirmed information unit (such as business type identifier, user ID, confirmed support type, etc.). After the interface returns response data, the system caches the data in memory and associates it with the corresponding process stage and interface identifier for easy subsequent querying and use.

[0143] In step S303, when the process state machine advances to a new process stage, and the business processing rules of this stage require calling a certain backend service interface, the system first checks whether the preloaded result of the interface exists in the memory cache. If matching cached data exists, the data is directly used for interface rendering (such as displaying compliance standards, filling default values, etc.); if not, a real-time interface call is initiated, and interface rendering is performed after obtaining the response data.

[0144] Taking the application for R&D subsidies by startups as an example, the process steps of this embodiment are as follows:

[0145] First, after the user confirms the "Enterprise Basic and Demand Information Unit (B=2 years, C=3 million yuan, D=R&D subsidy)," the system inputs the current process stage (State 2), the information unit, and the business rule (State 3 - R&D needs to collect E and verify compliance) into the graph neural network model. After model analysis, it predicts that the probability of calling the "R&D Project Plan Compliance Verification Interface" in the subsequent "State 3 - R&D" stage is 90% (assuming the preset threshold is set to 80%), thus determining this interface as the pre-loading target.

[0146] The system then asynchronously calls the "R&D Project Plan Compliance Verification Interface" in the background, carrying the parameters "business_type=Enterprise Support Fund Application&support_type=R&D Subsidy&user_id=XXX". The interface returns the response data: "Format requirements: PDF / Word; Content elements: Project background, R&D objectives, technical solutions, budget details; Size limit: ≤20MB". The system caches this data in memory and associates it with the identifier "Interface name=R&D Project Plan Compliance Verification Interface&Process stage=Status 3-R&D".

[0147] The process then transitions to state 3 - R&D, where the system needs to display the upload requirements for the R&D project plan on the interface. At this point, the system checks the memory cache and finds pre-loaded data for the "R&D Project Plan Compliance Verification Interface." It directly uses this data to render the interface prompt: "Please upload the R&D project plan (supports PDF / Word format, must include project background, R&D objectives, technical solutions, budget details, and size not exceeding 20MB)." Without needing to call the interface in real time, the interface renders instantly, reducing user waiting time.

[0148] The above solution avoids waiting delays caused by real-time interface calls during process advancement by preloading the required interface data. It can significantly improve the interface rendering speed, especially for backend interfaces with slow response times, making user interaction smoother.

[0149] By using a preloading mechanism to distribute API call time, the system avoids concurrency pressure caused by multiple users simultaneously initiating API calls when advancing to a certain process stage, thereby improving the overall stability of the system.

[0150] In some embodiments, in step S2, when a user enters content in an information input component but does not trigger a confirmation operation, such as Figure 6 As shown, the method further includes the following steps:

[0151] S201: Capture unconfirmed input values ​​in all current information input components in real time, map the input values ​​to the business processing rules in the current state of the process state machine, and identify the logical gap between the input but unconfirmed information and the final business rules;

[0152] S202: Based on the identified logical gaps, a counterfactual reasoning engine is used to perform multi-path deduction to determine the optimal deduction path. The counterfactual reasoning engine is based on a constraint satisfaction problem model that is isomorphic to the business processing rules. It is used to generate one or more complete information unit combinations containing virtual information that can satisfy all business rules. Each information unit combination represents a feasible deduction path from the current incomplete or non-compliant state to the final compliant state.

[0153] S203: Transform at least one optimal derivation path into a guiding suggestion and display the sequence of operations required to achieve compliance to the user through the interactive interface. The sequence of operations includes supplementing a specific information unit, modifying the value of an input information unit, or selecting one or more of a specific enumeration option.

[0154] In this embodiment, an unconfirmed input value refers to the content that the user has entered in the information input component but has not clicked the "confirm" button, and is in a state of pending verification and confirmation.

[0155] Logical gaps refer to discrepancies or incompleteness between entered but unconfirmed information and the final business rules. For example, the entered "Years of establishment: 10 years" may conflict with the rule "Startups have been established for ≤5 years", or required fields such as "Previous year's revenue" may not be entered.

[0156] Counterfactual reasoning engine refers to a reasoning tool built on a constraint satisfaction problem model that is isomorphic to business processing rules. It can generate multiple complete information unit combinations (i.e., feasible deduction paths) that can satisfy all business rules for the current incomplete or non-compliant input state.

[0157] The optimal deduction path refers to the path selected from multiple feasible deduction paths that has the fewest operation steps and best matches the user's current input intent. For example, a path that only requires adding one required field is better than a path that requires adding two fields.

[0158] An operation sequence refers to the set of specific operations that a user needs to perform to achieve a compliant state, including supplementing specific information, modifying already entered information, and selecting enumerated options.

[0159] In step S201, when a user enters content in the information input component but does not trigger a confirmation operation, the system captures all unconfirmed input values ​​in the information input components in real time. Subsequently, these input values ​​are mapped and compared with the business processing rules in the current state of the process state machine to identify logical gaps between the entered but unconfirmed information and the business rules, including missing information (not filling required fields) and non-compliant information (incorrect format, values ​​out of range, conflicts with other fields, etc.).

[0160] In step S202, the system activates the counterfactual reasoning engine, which performs multi-path deduction based on the identified logical gaps and in conjunction with business processing rules. The engine generates multiple complete information unit combinations containing virtual information, each combination representing a feasible deduction path from the current state to the compliant state. Subsequently, the system selects at least one optimal deduction path based on indicators such as the number of operation steps and the degree of consistency with the user's current input intent.

[0161] In step S203, the system converts the optimal deduction path into user-friendly guiding suggestions, displaying the sequence of operations required to achieve compliance through an interactive interface. The suggestions must be clear and specific, ensuring users understand each step they need to take, helping them quickly correct input errors, supplement missing information, and advance business process compliance.

[0162] Taking the user's application for R&D subsidies in status 2 as an example, the method in this embodiment includes the following steps:

[0163] First, the user entered "6 years" in the "Years of Establishment" input box in Status 2, selected "R&D Subsidy" for "Application Support Type," but did not enter "Previous Year's Revenue" and did not confirm the information. The system captured the two unconfirmed input values ​​"Years of Establishment: 6 years" and "Support Type: R&D Subsidy" in real time, and identified logical discrepancies after comparing them with the business rules: First, the "Startup" status corresponding to "R&D Subsidy" requires "Years of Establishment ≤ 5 years," and "6 years" does not meet the rule; second, "Previous Year's Revenue" is a required field, which was not entered at the moment.

[0164] Then, the counterfactual reasoning engine is activated, generating feasible deduction paths based on rules such as "establishment period ≤ 5 years", "revenue amount required", and "support type is R&D subsidy":

[0165] Path 1: Modify the establishment period to 3 years + supplement revenue of 3 million yuan;

[0166] Path 2: Modify the establishment period to 4 years + supplement revenue of 2 million yuan;

[0167] Path 3: Modify the establishment period to 5 years + supplement revenue of 4 million yuan.

[0168] The system selected Path 1 as the optimal path (with the fewest steps, and a revenue of 3 million yuan that falls within the common revenue range for startup R&D subsidy applications, thus meeting the potential needs of users).

[0169] The system then displays a guiding message on the interactive interface: "To ensure your application complies with the rules, please complete the following operations: (1) Change the establishment period to a positive integer of ≤5 years (e.g., 3 years); (2) Add an Arabic numeral (e.g., 300) to the 'Previous Year Revenue (Ten Thousand Yuan)' input box. After completion, you can click 'Confirm the above information' to continue the process." This helps users quickly correct non-compliant information and supplement missing fields.

[0170] The above solution proactively identifies logical discrepancies between input and rules, providing clear operational guidance to prevent users from repeatedly modifying their input due to unfamiliarity with the rules, thus reducing trial-and-error costs. It is particularly suitable for scenarios with complex business rules. By guiding users to correct input deviations through the optimal deduction path, it ensures that the final submitted information complies with business rules, reducing issues such as application rejection and process rework caused by non-compliant information, thereby improving business processing efficiency.

[0171] In some embodiments, after determining the target business type in step S1, and before initializing the interactive interface according to the target business type and the current process stage in step S2, such as Figure 7 As shown, the method further includes the following steps:

[0172] S101: Receive the user's initial service request through a multimodal interaction interface, which integrates a speech recognition unit, an image recognition unit, and an emotion computing unit; the speech recognition unit and the image recognition unit are used to convert the user's speech input information and image input information into text or structured data, respectively; the emotion computing unit is used to analyze the acoustic features and text emotion during the user's input process and output emotion state labels.

[0173] S102: Based on the emotional state tags and the completeness of the initial information provided by the user through the multimodal interaction interface, dynamically determine the guidance strategy for subsequent interaction interfaces, specifically including:

[0174] If the emotional state label is confused or anxious, the decision-making guidance strategy is a simplified guidance model;

[0175] If the image recognition unit has extracted some of the data required by the information unit, the guidance strategy determined is the information pre-filling mode;

[0176] S103: When initializing the interactive interface in step S2, the presentation logic of the interface components is adaptively adjusted according to the determined guidance strategy, specifically including:

[0177] When the guidance strategy is a simplified guidance mode, the generated interface will focus on displaying the core required information units at the current stage of the process, and will use step-by-step, enhanced guidance interactive text.

[0178] When the guidance strategy is in the information pre-fill mode, the generated interface will automatically fill the corresponding information input component with the data extracted through image recognition and set its UI status to await user confirmation.

[0179] In this embodiment, the multimodal interaction interface refers to an interaction entry point that integrates a speech recognition unit, an image recognition unit, and an emotion computing unit. It supports users to input information through multimodal methods such as voice, image, and text, and can also analyze the user's emotional state.

[0180] Speech recognition unit: A functional module that converts the user's voice input information into text data, facilitating semantic analysis and intent recognition by the system.

[0181] The sentiment computing unit refers to a functional module that analyzes the acoustic features (speech rate and pitch) and text sentiment (such as keywords like "anxious" and "don't understand") during the user input process and outputs sentiment status labels (such as confused type and anxious type).

[0182] Guidance strategies refer to interactive guidance methods determined based on the user's emotional state tags and the completeness of initial information, including simplified guidance modes and information pre-filling modes.

[0183] The simplified guidance mode refers to an interactive mode that focuses on core required information and adopts a step-by-step reinforcement guidance approach for users whose emotional state is confused or anxious, thereby reducing information interference.

[0184] Information pre-filling mode refers to an interactive mode that automatically fills the extracted data into the corresponding input component and sets it to a pending confirmation state when partial information has been extracted from image recognition.

[0185] In step S101, the user initiates an initial service request through a multimodal interaction interface. The speech recognition unit in the interface converts the speech input into text, the image recognition unit converts the image input into structured data, and the sentiment computing unit analyzes the acoustic features and text sentiment of the user input and outputs sentiment state labels (such as confused type, anxious type, and calm type).

[0186] In step S102, the system dynamically decides on the guidance strategy based on the emotional state label and the initial information completeness (the amount of information extracted by image recognition): if the emotional state label is confused (e.g., the text contains "don't understand", "how to do it", etc.) or anxious (e.g., speaking fast, the text contains "anxious", "as soon as possible", etc.), the decision is to use the simplified guidance mode; if the image recognition unit has extracted some of the data required by the information unit (e.g., the enterprise type and establishment date in the business license), the decision is to use the information pre-filling mode; if the emotional state is stable and no valid image information has been extracted, the default guidance mode (fully displaying the fields required for the current stage) is adopted.

[0187] In step S103, when the system executes step S2 to initialize the interactive interface, it adjusts the presentation logic of the interface components according to the determined guidance strategy. Specifically, when the guidance strategy is a simplified guidance mode, the interface focuses on the core required information units of the current process stage, hides non-core fields, and adopts step-by-step and enhanced guidance text to reduce the user's information processing burden. When the guidance strategy is an information pre-filling mode, the data extracted by image recognition is automatically filled into the corresponding information input components, and the component UI status is set to "pending confirmation" (e.g., marked "pre-filled, confirm whether it is correct"). The user can directly confirm or modify.

[0188] Taking a user's application for R&D subsidies through a multimodal approach as an example, the method involved in this application includes the following steps:

[0189] First, the user initiates a request through the interactive interface: They input "I want to apply for R&D subsidies, but I don't know how to fill out the form, I need it urgently" (speaking quickly), and upload a photo of their business license. Then, the multimodal interactive interface processes the data, specifically including:

[0190] The speech recognition unit converts speech into text: "I want to apply for R&D subsidies, but I don't know how to fill it out. I need it urgently."

[0191] The image recognition unit identifies the business license and extracts "Enterprise Type: Startup, Establishment Date: May 2022";

[0192] The sentiment computing unit analyzes texts with fast speech and phrases such as "don't know how to fill in" and "urgent need", and outputs dual sentiment labels of "urgent type + confused type".

[0193] The system then uses a simplified guidance mode based on the emotional state tags of "urgency + confusion" to guide decision-making; at the same time, since the "enterprise type" information has been extracted by image recognition, information pre-filling logic is superimposed.

[0194] When initializing the interactive interface for state 2, adjust according to the following guidance strategy:

[0195] To simplify the guidance strategy, only the core required fields "Company Type" and "Previous Year's Revenue (Ten Thousand Yuan)" will be displayed, while non-core fields will be hidden. The guidance text will be strengthened to "To help you quickly advance your application, please fill in the following core information (required) first."

[0196] If an information pre-filling strategy is used, the "Startup" extracted by image recognition will be automatically filled into the "Enterprise Type" radio button group, and the UI will indicate "Pre-filled (from your uploaded business license), please confirm if it is correct". Users can directly click "Confirm" or reselect.

[0197] The above solutions provide differentiated guidance for users with varying emotional states (confusion, urgency) and information completeness, making the service more aligned with their actual needs. The simplified guidance mode focuses on core information and reduces information clutter, suitable for users unfamiliar with the process or under time pressure; the pre-filled information mode reduces repetitive input and improves operational efficiency, especially suitable for business scenarios requiring document uploads. Simultaneously, the system can sense user emotions (e.g., detect user urgency) and provide convenient guidance, while automatically pre-filling information using image recognition, allowing users to experience the system's intelligence and thoughtfulness, enhancing trust and user satisfaction.

[0198] In some embodiments, after determining the target service type in step S1, and before initializing the interactive interface according to the target service type and the current process stage, such as Figure 8 As shown, the method further includes the following steps:

[0199] S110: Collect contextual information for decision-making, including sentiment state tags, historical interaction preferences and intent recognition results related to the current user and target business type, and the sentiment state tags are obtained by analyzing acoustic features and text sentiment during the user input process;

[0200] S111: Based on the context information, an adaptive decision engine calculates and generates an interaction strategy suitable for the current process stage. This interaction strategy is used to dynamically modulate the attributes of the information input components in step S2, specifically including:

[0201] Based on users' emotional state tags, historical interaction preferences, and intent recognition results, the presentation order of information units, as well as the level of detail and style of the guiding copy, are determined.

[0202] Based on the user's historical operation accuracy and current emotional state label, determine the fault tolerance threshold and prompt strength for real-time verification of user input;

[0203] Determine the tone and style of interface feedback prompts based on emotional state tags;

[0204] S112: When initializing the interactive interface in step S2, the initialization logic of the information input component is adaptively configured according to the interaction strategy generated in step S111.

[0205] Taking a user's re-application for R&D subsidies as an example, the method in this embodiment includes the following steps:

[0206] First, the system collects relevant contextual information about the user, including emotional state tags, historical interaction preferences, and intent recognition results. For example, when a user inputs "Apply for R&D subsidies, follow the previous process," the system can determine that the user's tone is stable and output the emotional state tag as "stable type." Secondly, by querying the user's operation log, it can be found that the user selected "simple guidance" in both of their previous transactions and filled in "last year's revenue" first each time. Therefore, the system can output the user's historical interaction preferences as "prefers simple guidance and habitually fills in revenue first." Thirdly, by analyzing the user's input text, the system can identify the user's core need as "to complete the application efficiently according to the familiar process" and use this core need as the corresponding intent recognition result.

[0207] The adaptive decision engine then generates interaction strategies based on the above information, specifically including:

[0208] Information presentation order: previous year's revenue → enterprise type → type of support applied for (aligning with users' habit of filling in revenue first);

[0209] The onboarding text is concise, only prompting the core operation "Please fill in the following information to complete the application," without any additional rules or explanations (in line with a preference for simplicity).

[0210] Real-time verification parameters: User historical operation accuracy rate is 90% (high), fault tolerance threshold is set to medium (minor formatting errors are allowed, such as "3 million" being automatically converted to "300"), and prompt intensity is set to weak (non-serious errors are only prompted in gray text).

[0211] Feedback prompts should be delivered in a neutral tone, such as "Information confirmed" or "Please enter Arabic numerals".

[0212] Then, the interactive interface is initialized and configured. Specifically, during the initialization of the interactive interface in state 2, the configuration is performed according to the current guidance strategy. The configuration content includes:

[0213] Component order: For example, the interactive interface can display the "Previous Year's Revenue (Ten Thousand Yuan)" input box, the "Enterprise Type" radio button group, and the "Application for Support Type" radio button group from top to bottom;

[0214] Guiding text: For example, display a concise guide at the top of the interactive interface: "Please fill in the following information to complete the application";

[0215] Verification configuration: For example, when a user enters "3 million", the system automatically converts it to "300" without prompting; when the user enters "abc", the gray text "Please enter Arabic numerals" is displayed below the input box.

[0216] Feedback tone: For example, after the user confirms the information, prompt "Information has been confirmed, and we will proceed to the next stage" (neutral tone).

[0217] The above solution generates personalized interaction strategies based on users' historical preferences and the current scenario, allowing users to operate in a familiar and habit-friendly interface, reducing adaptation costs and improving comfort. The information presentation order aligns with user habits, the guiding text conforms to user preferences, and the verification parameters are adapted to user operation accuracy. This allows users to efficiently complete operations according to their own habits without needing to adapt to system rules, reducing cognitive burden.

[0218] In a second aspect, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for seamlessly handling business in an interactive interface as described in the first aspect of the present invention.

[0219] The computer-readable storage medium may be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory.

[0220] The non-volatile memory may be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), a magnetic random access memory (FRAM), a flash memory, a magnetic surface memory, an optical disc, or a compact disc read-only memory (CD ROM); the magnetic surface memory may be a disk storage device or a magnetic tape storage device.

[0221] The volatile memory may be random access memory (RAM), which serves as an external cache. By way of example, but not limitation, many forms of RAM are available, such as static random access memory (SRAM), synchronous static random access memory (SSRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synclink dynamic random access memory (SLDRAM), and direct memory bus random access memory (DRRAM). The computer-readable storage media described in the embodiments of the present invention are intended to include these and any other suitable types of memory.

[0222] like Figure 9 As shown, in a third aspect, the present invention provides an electronic device 10, including a processor 101 and a storage medium 102, wherein a computer program is stored on the storage medium, and the computer program, when executed by the processor, implements the method for seamlessly handling business in an interactive interface as described in the first aspect of the present invention.

[0223] In some embodiments, the processor may be implemented by software, hardware, firmware, or a combination thereof, and may use at least one of the following: circuit, single or multiple application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), central processing units (CPUs), controllers, microcontrollers, and microprocessors, thereby enabling the processor to execute some or all of the steps, or any combination thereof, of the methods for seamlessly handling business in an interactive interface as described in the various embodiments of this application.

[0224] Finally, it should be noted that although the above embodiments have been described in the text and drawings of this application, this should not limit the scope of patent protection of this application. Any technical solutions that are based on the essential concept of this application and utilize the content described in the text and drawings of this application, resulting in equivalent structural or procedural substitutions or modifications, as well as the direct or indirect application of the technical solutions of the above embodiments to other related technical fields, are all included within the scope of patent protection of this application.

Claims

1. A method for seamless transaction handling in an interactive interface, characterized by, The method comprises: S1: receiving a service request of a user through an interactive interface, determining a target service type, and loading a corresponding service handling rule, the service handling rule comprising a list of necessary information fields and a process state machine in which a plurality of ordered process stages are predefined; S2: according to the current process stage, guiding the user to input information through at least one information input component, and performing real-time verification on the information input by the user each time, the input information being used to fill in the corresponding information field, and whenever an information field passes the verification, a structured confirmation component is dynamically generated and rendered in the interactive interface for receiving the user's final confirmation operation on the information field; S3: listening to the user's confirmation operation on the structured confirmation component, wherein one or more information fields confirmed by the user constitute a confirmed information unit; S4: based on all the currently confirmed information units, querying the process state machine to determine whether the conditions for jumping from the current process stage to the subsequent target process stage are met, and if the conditions are met, performing an interface reconstruction operation, the interface reconstruction operation comprising: clearing all the information input components and structured confirmation components rendered in the current process stage, and rendering the information input components corresponding to one or more information fields required by the target process stage according to the definition of the target process stage; S5: when the process state machine jumps to the final stage and all the necessary information fields are included in the confirmed information units, based on all the confirmed information units, automatically packaging a complete data request package conforming to the interface specification of an external service system; S6: submitting the complete data request package by calling the backend application interface of the external service system, and displaying the service handling result returned by the external service system in the same interactive interface; Step S4 further comprises: S401: based on all the currently confirmed information units, calculating the overall completeness of all the confirmed information units with respect to all the necessary information fields; S402: querying the process state machine to obtain a plurality of selectable jump paths matched with the overall completeness, wherein each selectable jump path corresponds to a different subsequent target process stage; S403: rendering a path selection interface in the interactive interface to show the plurality of selectable jump paths to the user; S404: in response to the selection operation of the user in the path selection interface, determining the final target process stage, and then performing the interface reconstruction operation.

2. The method for conducting transactions seamlessly in an interactive interface of claim 1, wherein, When performing step S4, the following steps are further included: S41: based on the currently confirmed information units, performing real-time analysis through a pre-trained intent recognition model to predict the user's potential associated service intent; S42: querying a business knowledge graph to obtain the target service handling rule and entry interface corresponding to the potential associated service intent; S43: In response to the interface reconstruction operation, render at least one shortcut entry component corresponding to the potential associated business intent in the interactive interface in synchronization with the information input components required by the target process stage; wherein the shortcut entry component encapsulates the target business handling rule and entry interface, and is pre-filled with the reusable information related to the target business extracted from all currently confirmed information units; S44: In response to the user's triggering operation on the shortcut entry component, pre-fill the reusable information as initial input, and initialize and start a new target business handling process instance in the current interactive interface based on the target business handling rule.

3. The method for conducting transactions seamlessly in an interactive interface of claim 1, wherein, After step S1 and before step S2, further comprising: S11: Query a locally encrypted cross-business information unit library, which is used to store authorized confirmed information units after the user completes the business; S12: Fill the confirmed information units matched with the necessary information fields in the current business handling rule as pre-filled information into the corresponding information input components automatically, and initialize the state of the information input components as verified and passed through the confirmation logic.

4. The method for conducting transactions seamlessly in an interactive interface of claim 1, wherein, In step S3, when any information unit is marked as confirmed, the method further comprises a predictive resource preloading process, specifically as follows: S301: Based on the current process stage, all confirmed information units, and the business handling rule, predict at least one backend service interface with a call probability exceeding a preset threshold in the subsequent process step in the current context through a pre-trained graph neural network model; S302: Asynchronously initiate a call to the backend service interface predicted in step S301 in the background; the call carries the necessary parameters extracted from the currently confirmed information units, and caches the response data returned by the called backend service interface in the memory; S303: When the process state machine advances to a new process stage, and the business handling rule at the process stage needs to call a certain backend service interface, first check whether there is a preloading result of the backend service interface in the cache; if there is, directly use the cache data for interface rendering.

5. The method for seamless transaction handling in an interactive interface as claimed in claim 1 wherein, In step S2, when the user inputs content in an information input component but does not trigger the confirmation operation, the method further comprises the following steps: S201: Real-time capture the unconfirmed input values in all information input components, and map the input values with the business handling rule at the current state of the process state machine to identify the logical gap between the input but unconfirmed information and the final business rule; S202: Based on the identified logical gap, use a counterfactual reasoning engine to perform multi-path reasoning to determine the optimal reasoning path, the counterfactual reasoning engine is based on a constraint satisfaction problem model isomorphic to the business handling rule, and is used to generate one or more complete information unit combinations containing virtual information that can satisfy all business rules, wherein each information unit combination represents a feasible reasoning path from the current incomplete or irregular state to the final compliant state; S203: convert the at least one optimal deduction path into a guiding suggestion, and show the operation sequence required to achieve compliance to the user through the interactive interface, the operation sequence including any one or more of supplementing a specific information unit, modifying the value of an input information unit, or selecting a specific enumerated option.

6. The method for seamless transaction handling in an interactive interface as claimed in claim 1 wherein, After determining the target service type in step S1, and before initializing the interactive interface according to the target service type and the current process stage in step S2, the method further comprises the following steps: S101: receiving an initial service request of a user through a multi-modal interactive interface, the multi-modal interactive interface integrating a voice recognition unit, an image recognition unit, and an emotion computing unit; the voice recognition unit and the image recognition unit are respectively used to convert voice input information and image input information of the user into text or structured data, and the emotion computing unit is used to analyze acoustic features and text emotions in the user input process and output an emotion state label; S102: dynamically deciding a guiding strategy of a subsequent interactive interface based on the emotion state label and the completeness of initial information provided by the user through the multi-modal interactive interface, specifically including: if the emotion state label is of a confused type or an urgent type, the decided guiding strategy is a simplified guiding mode; if part of data required by an information unit has been extracted through the image recognition unit, the decided guiding strategy is an information pre-filling mode; S103: when initializing the interactive interface in step S2, adaptively adjusting the presentation logic of interface components according to the decided guiding strategy, specifically including: when the guiding strategy is the simplified guiding mode, the generated interface focuses on displaying core mandatory information units in the current process stage, and uses step-by-step and reinforced guiding interactive scripts; when the guiding strategy is the information pre-filling mode, the generated interface automatically fills the data extracted through the image recognition into the corresponding information input component, and sets the UI state thereof to be confirmed by the user.

7. The method for seamless transaction handling in interactive interface as claimed in claim 1 wherein, After determining the target service type in step S1, and before initializing the interactive interface according to the target service type and the current process stage, the method further comprises the following steps: S110: collecting context information for decision-making, the context information including an emotion state label, historical interaction preferences and intent recognition results related to the current user and the target service type, and the emotion state label being obtained based on analysis of acoustic features and text emotions in the user input process; S111: based on the context information, calculating and generating a set of interactive strategies suitable for the current process stage through an adaptive decision engine, the interactive strategies being used to dynamically modulate the attributes of information input components in step S2, specifically including: determining the presentation order of information units, and the detail level and style of guiding scripts based on the emotion state label, historical interaction preferences and intent recognition results of the user; determining a fault tolerance threshold and a prompt intensity for real-time verification of user input based on the historical operation accuracy of the user and the current emotion state label; determining the tone style of interface feedback prompt information based on the emotion state label; S112: When the interactive interface is initialized in step S2, the initialization logic of the information input component is adaptively configured according to the interaction strategy generated in step S111.

8. A computer readable storage medium having stored thereon a computer program, characterized in that, The program is executed by the processor to implement the method for seamlessly handling a service in an interactive interface according to any one of claims 1 to 7.

9. An electronic device on which a computer program is stored, characterized by A device comprising a processor and a storage medium, wherein the storage medium stores a computer program, and the computer program is executed by the processor to implement the method for seamlessly handling a service in an interactive interface according to any one of claims 1 to 7.

Images