Intelligent interaction method and device, electronic equipment and storage medium

Abstract

The invention provides an intelligent interaction method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining to-be-processed information corresponding to an information input operation through an interaction entrance in response to the information input operation; judging the information type of the to-be-processed information through a large model in the large model service platform; based on the information type of the to-be-processed information, processing the to-be-processed information by using a corresponding processing mode to obtain a processing result; wherein the processing modes comprise large model processing, application service processing and information supplement. According to the invention, barriers among traditional enterprise applications are broken through, so that various applications and services can be accessed and used through one entrance, the operation process is greatly simplified, and the working efficiency is improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent interaction technology, and in particular to an intelligent interaction method, device, electronic device and storage medium. Background Art

[0002] With the rapid development of artificial intelligence technology, large-model integrated agents have become a key force in promoting intelligent services and are widely used in intelligent assistants, customer service systems, content creation and other fields. This type of agent is based on advanced machine learning and natural language processing technologies and can understand and process complex data and language information. By integrating multiple models, data and interfaces, it forms a powerful knowledge processing and decision support system that can perform various tasks and provide users with efficient and intelligent service experience.

[0003] In traditional enterprise-level communication and office environments, various applications and services are often scattered, and users need to switch between different systems and platforms to complete different tasks. For example, employees may need to use one application for email communication, another system to manage schedules, and other tools to handle matters such as attendance and leave. This fragmented system architecture leads to the complexity of usage calls, which not only increases the user's operational burden but also reduces work efficiency. In addition, data and functions between different systems are often difficult to communicate, further increasing the complexity and inconvenience of user use. Summary of the invention

[0004] In view of this, the purpose of the present application is to provide a method, device, electronic device and storage medium for intelligent interaction to overcome the problems in the prior art.

[0005] In a first aspect, an embodiment of the present application provides a method for intelligent interaction, the method comprising:

[0006] Construction phase and interaction phase;

[0007] The construction phase includes: in response to the portal setting operation, setting an interactive portal in the original communication tool of the target enterprise;

[0008] In response to the application service building operation, integrating the selected plug-ins and knowledge data sets corresponding to the original multiple independent applications in the target enterprise to form multiple application services;

[0009] In response to the application service configuration operation, the application service is connected to a preset large model service platform, wherein the large model service platform is also connected to a plurality of interfaces to be selected;

[0010] The interaction stage includes: in response to an information input operation, obtaining to-be-processed information corresponding to the information input operation through the interaction portal;

[0011] Determining the information type of the information to be processed by using the big model in the big model service platform;

[0012] Based on the information type of the information to be processed, the information to be processed is processed using a corresponding processing method to obtain a processing result; wherein the processing method includes large model processing, application service processing and information supplementation.

[0013] In some technical solutions of the present application, the above-mentioned information type includes a first type; the processing method corresponding to the first type is application service processing;

[0014] The processing of the information to be processed using a corresponding processing method based on the information type of the information to be processed to obtain a processing result includes:

[0015] The information to be processed of the first type is processed by using an application service processing method to obtain a processing result, including:

[0016] Selecting a target application service according to the matching degree between the information to be processed and each of the application services;

[0017] The target application service is used to process the information to be processed to obtain a processing result.

[0018] In some technical solutions of the present application, the above-mentioned information type includes a second type; and the processing method corresponding to the second type is to supplement the information;

[0019] The processing of the information to be processed using a corresponding processing method based on the information type of the information to be processed to obtain a processing result includes:

[0020] The information to be processed of the second type is processed by using an application service processing method to obtain a processing result, including:

[0021] Retrieve the information to be processed using a retrieval enhancement generation model to obtain supplementary information of the information to be processed;

[0022] Perform matching based on the supplementary information to obtain matched recombinant prompt data;

[0023] The reorganization prompt data is processed using a large model to obtain a processing result.

[0024] In some technical solutions of the present application, the above-mentioned information type includes a third type; the processing method corresponding to the second type is large model processing;

[0025] The processing of the information to be processed using a corresponding processing method based on the information type of the information to be processed to obtain a processing result includes:

[0026] The third type of information to be processed is processed by using an application service processing method to obtain a processing result, including:

[0027] The information to be processed is processed using the large model to obtain a processing result.

[0028] In some technical solutions of the present application, the above-mentioned construction stage also includes: constructing a permission management and security control mechanism;

[0029] The interaction stage also includes: verifying the user who inputs the information, and when the user passes the authority verification, obtaining the to-be-processed information corresponding to the information input operation through the interaction entrance.

[0030] In some technical solutions of the present application, the above-mentioned construction stage also includes: building a user feedback mechanism;

[0031] The interaction stage also includes: obtaining user feedback information, and optimizing the large model according to the feedback information.

[0032] In some technical solutions of the present application, the above-mentioned determining the information type of the information to be processed by the big model in the big model service platform includes:

[0033] The large model is used to extract request parameters in the information to be processed, and the correctness of the request parameters is determined.

[0034] In a second aspect, an embodiment of the present application provides a device for intelligent interaction, the device comprising:

[0035] Building blocks and interaction modules;

[0036] The construction module is used to: in response to the portal setting operation, set an interactive portal in the original communication tool of the target enterprise;

[0037] In response to the application service building operation, integrating the selected plug-ins and knowledge data sets corresponding to the original multiple independent applications in the target enterprise to form multiple application services;

[0038] In response to the application service configuration operation, the application service is connected to a preset large model service platform, wherein the large model service platform is also connected to a plurality of interfaces to be selected;

[0039] The interaction module is used to: in response to an information input operation, obtain the to-be-processed information corresponding to the information input operation through the interaction entrance;

[0040] Determining the information type of the information to be processed by using the big model in the big model service platform;

[0041] Based on the information type of the information to be processed, the information to be processed is processed using a corresponding processing method to obtain a processing result; wherein the processing method includes large model processing, application service processing and information supplementation.

[0042] In a third aspect, an embodiment of the present application provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the above-mentioned intelligent interaction method when executing the computer program.

[0043] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned intelligent interaction method are executed.

[0044] The technical solution provided by the embodiments of the present application may have the following beneficial effects:

[0045] The construction phase of the method of the present application includes: in response to the entry setting operation, setting an interactive entry in the original communication tool of the target enterprise; in response to the application service construction operation, integrating the selected plug-ins and knowledge data sets corresponding to the original multiple independent applications in the target enterprise to form multiple application services; in response to the application service configuration operation, connecting the application service with a preset big model service platform, wherein the big model service platform is also connected to multiple interfaces to be selected; the interaction phase includes: in response to the information input operation, obtaining the information to be processed corresponding to the information input operation through the interactive entry; judging the information type of the information to be processed through the big model in the big model service platform; based on the information type of the information to be processed, using the corresponding processing method to process the information to be processed to obtain a processing result; wherein the processing method includes big model processing, application service processing and information supplementation.

[0046] This application breaks down the barriers between traditional enterprise applications, allowing various applications and services to be accessed and used through one entrance, greatly simplifying the operating process and improving work efficiency.

[0047] In order to make the above-mentioned objects, features and advantages of the present application more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0049] Figure 1 A schematic diagram of a flow chart of an intelligent interaction method provided in an embodiment of the present application is shown;

[0050] Figure 2 A schematic diagram of a global AI assistant architecture provided in an embodiment of the present application is shown;

[0051] Figure 3 A diagram showing the composition of an intelligent application service provided by an embodiment of the present application is shown;

[0052] Figure 4 A user question and answer flow chart provided in an embodiment of the present application is shown;

[0053] Figure 5 A logic diagram of an AI assistant interaction entry provided in an embodiment of the present application is shown;

[0054] Figure 6 A schematic diagram of an intelligent interaction device provided in an embodiment of the present application is shown;

[0055] Figure 7 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0056] To make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the drawings in the present application only serve the purpose of explanation and description and are not used to limit the scope of protection of the present application. In addition, it should be understood that the schematic drawings are not drawn in real proportion. The flowchart used in this application shows the operations implemented according to some embodiments of the present application. It should be understood that the operations of the flowchart can be implemented out of sequence, and the steps without logical context can be reversed in order or implemented simultaneously. In addition, those skilled in the art can add one or more other operations to the flowchart under the guidance of the content of the present application, or remove one or more operations from the flowchart.

[0057] In addition, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. The components of the embodiments of the present application described and shown in the drawings here can be arranged and designed in various configurations. Therefore, the following detailed description of the embodiments of the present application provided in the drawings is not intended to limit the scope of the application claimed for protection, but merely represents the selected embodiments of the present application. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without making creative work belong to the scope of protection of the present application.

[0058] It should be noted that the term "comprising" will be used in the embodiments of the present application to indicate the existence of the features declared thereafter, but does not exclude the addition of other features.

[0059] With the rapid development of artificial intelligence technology, large-model integrated agents have become a key force in promoting intelligent services and are widely used in intelligent assistants, customer service systems, content creation and other fields. This type of agent is based on advanced machine learning and natural language processing technologies and can understand and process complex data and language information. By integrating multiple models, data and interfaces, it forms a powerful knowledge processing and decision support system that can perform various tasks and provide users with efficient and intelligent service experience.

[0060] In traditional enterprise-level communication and office environments, various applications and services are often scattered, and users need to switch between different systems and platforms to complete different tasks. For example, employees may need to use one application for email communication, another system to manage schedules, and other tools to handle matters such as attendance and leave. This fragmented system architecture leads to the complexity of usage calls, which not only increases the user's operational burden but also reduces work efficiency. In addition, data and functions between different systems are often difficult to communicate, further increasing the complexity and inconvenience of user use.

[0061] Based on this, the embodiments of the present application provide a method, device, electronic device and storage medium for intelligent interaction, which are described below through embodiments.

[0062] Figure 1 A flow chart of a method for intelligent interaction provided by an embodiment of the present application is shown, wherein the method comprises steps S101-S102; specifically: S101, construction phase and S102, construction interaction phase;

[0063] The construction phase includes: in response to the portal setting operation, setting an interactive portal in the original communication tool of the target enterprise;

[0064] In response to the application service building operation, integrating the selected plug-ins and knowledge data sets corresponding to the original multiple independent applications in the target enterprise to form multiple application services;

[0065] In response to the application service configuration operation, the application service is connected to a preset large model service platform, wherein the large model service platform is also connected to a plurality of interfaces to be selected;

[0066] The interaction stage includes: in response to an information input operation, obtaining to-be-processed information corresponding to the information input operation through the interaction portal;

[0067] Determining the information type of the information to be processed by using the big model in the big model service platform;

[0068] Based on the information type of the information to be processed, the information to be processed is processed using a corresponding processing method to obtain a processing result; wherein the processing method includes large model processing, application service processing and information supplementation.

[0069] This application breaks down the barriers between traditional enterprise applications, allowing various applications and services to be accessed and used through one entrance, greatly simplifying the operating process and improving work efficiency.

[0070] Some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.

[0071] The present application discloses a method for intelligent interaction, and the method in the present application mainly includes a construction phase and an interaction phase. Among them, the construction phase is mainly to build an intelligent agent (AI assistant), and the interaction phase is the interaction process with the intelligent agent.

[0072] For the construction phase, the embodiment of the present application mainly includes the setting of the interactive portal, the construction of the application service and the configuration of the application service. Specifically, in response to the portal setting operation, the interactive portal is set in the original communication tool of the target enterprise; in response to the application service construction operation, the selected plug-ins and knowledge data sets corresponding to the original multiple independent applications in the target enterprise are integrated to form multiple application services; in response to the application service configuration operation, the application service is connected to the preset large model service platform, wherein the large model service platform is also connected to multiple interfaces to be selected.

[0073] When implementing it, Figure 2As shown, the AI ​​assistant constructed in the embodiment of the present application includes a user side, an entry layer, an application layer, a service layer and a physical layer. Among them, the user layer includes employees, marketers, etc. The entry layer is used to set the interactive entry. When setting the interactive entry, the embodiment of the present application is based on the original communication tools of the target enterprise, that is, the embodiment of the present application sets the interactive entry in the original communication tools of the target enterprise. The interactive entry for users to call out the AI ​​assistant globally in the enterprise-level communication tools is provided on the PC side, mobile side, and large-screen side. The front end uses the native and H5 development of IOS and Android APP to ensure the smoothness and responsiveness of the user interface. Users can interact with the AI ​​assistant through voice or text, and the system transmits data through the WebSocket protocol. The ASR technology based on various sequence neural network structures such as LSTM, AttentionModel, DeepCNN, etc. provided by the enterprise-level PaaS product is used to convert speech recognition into text in real time. The next step is to communicate with the back-end AI large model service platform through WebSocket and HTTPS protocols.

[0074] The application layer is used to build application services: multiple intelligent application service modules are built by integrating API plug-ins and knowledge data sets. The global AI assistant provides intelligent application services to enterprise employees through built-in intelligent application service modules. These modules include but are not limited to: Product application awakening module: using the plug-in capability of the big model based on Function Call, other applications are directly awakened through the global AI assistant to trigger specific functions or operations, providing users with a smarter way of interaction. Unified to-do task query: integrating the to-do query portal of the employee's common task process platform, summarizing and displaying all the user's to-do tasks, and providing a to-do task processing portal. Knowledge base question and answer module: accessing the internal knowledge base and public knowledge base of the enterprise, based on the natural language processing capability of the big model and the RAG knowledge base question and answer capability, to provide users with accurate question and answer services. Calendar event management module: users can manage calendars through voice or text commands, including adding, modifying, deleting and querying events, to ensure that users have a clear view of their schedules. Conference room reservation module: providing real-time reservation and management functions for conference rooms, users can view the use of conference rooms and make reservations. Menu query module: integrating the menu query service of the enterprise restaurant, so that users can view daily meal options. Haircut appointment module: Users can make haircut appointments through the AI ​​assistant, view available time slots and confirm appointments. Accommodation application module: Provides employee accommodation application and management functions to simplify the process of employee accommodation application. Attendance and leave module: Integrates the enterprise attendance system interface and data to realize employee attendance management and leave application, and provides relevant reminders and record inquiries. Address book query module: Real-time synchronization of enterprise address book data, quick query of enterprise address book, help users find colleagues' contact information, and promote internal communication. Marketing operation auxiliary module: Integrates enterprise marketing products to provide marketers with sales data analysis, customer relationship management, customer demand analysis, potential customer recommendation, marketing script generation and other functions to improve the intelligence level of marketing work.

[0075] The above intelligent application service module composition architecture is detailed in the attached Figure 3 , where the key technologies are as follows:

[0076] Preset prompt words: Use natural language to stimulate the AI ​​big model to respond or operate in a specific way, help the AI ​​big model better understand the role positioning of the intelligent application service, help it better understand the user's intentions, and respond in a more accurate and expected way.

[0077] API plug-in set based on Function Call technology: allows AI assistants to call internal or external APIs (application programming interfaces) of the enterprise to expand their functions. The API plug-in set is a set of pre-integrated APIs. API is a set of predefined rules and standards for communication and data exchange between different software. API interfaces are usually provided by the server side, and clients can call these interfaces by sending HTTP requests to implement various functions. The request parameters of the API interface refer to the data that needs to be passed to the server when calling the API interface. These data can be basic types such as strings, numbers, Boolean values, or complex types such as arrays and objects. After understanding the user's intention, the AI ​​big model selects the appropriate plug-in in the API plug-in set through Function Call technology for calling. This technology realizes the seamless connection between the AI ​​big model and real data and services, enhancing its practicality and intelligence level.

[0078] RAG retrieval enhancement technology: This is a technology that retrieves relevant information from massive knowledge data sets and then uses this information to assist the AI ​​large model in generating answers or content. The AI ​​assistant implements vector similarity retrieval based on the vectorized model M3E or GTE, and hybrid retrieval with hybrid keyword full-text retrieval technology, enabling it to quickly find the most relevant information and combine this information to generate more accurate and detailed answers.

[0079] Knowledge data set: Contains internal and public knowledge bases of enterprises, is the information source of intelligent application service modules, and provides rich data support for large AI models.

[0080] The service layer includes the AI ​​big model service platform. The intelligent application service module communicates with the model question and answer API routing of the back-end big model service platform developed based on the Flask framework through the WebSocket and HTTPS protocols. Various APIs in the API plug-in set initiate requests through the back-end big model service platform. The API request is transferred through the open platform responsible for API management and authentication. The API open platform acts as an API request agent, distributes tokens to each application for identity authentication, and forwards the API request to the API server after authentication to obtain services, thereby realizing real-time data acquisition and processing, and task creation and execution. Combining the API response data with the intelligent application service module to reorganize the question and answer prompt, the big model comprehensive intelligent agent provides users with intelligent service responses based on the reorganized question and answer prompt. Specifically, the following steps are included:

[0081] API encapsulation intelligent plug-in: The standardized API that connects to the enterprise's internal application system and database is encapsulated into an intelligent plug-in that can be understood by the large model comprehensive intelligent body. The intelligent plug-in can be used to build intelligent application service modules.

[0082] Create and bind the vectorized application knowledge base: pre-process, format and classify the internal or external knowledge base of the enterprise, use the vectorization model M3E or GTE to vectorize the knowledge base data, and bind it to the intelligent application service module.

[0083] Big model analysis and service call: Use the big model comprehensive intelligent agent to analyze the intention of user interaction issues, and the model selects the appropriate intelligent application service module to respond based on the user's intention.

[0084] During the construction phase, we also built a permission management and security control mechanism. The AI ​​assistant uses the OAuth 2.0 account permission authentication mechanism of enterprise-level communication tools as the permission control base, and builds fine-grained access control for each intelligent application service module, plug-in, and knowledge base in the large model service platform. The AI ​​assistant ensures the reasonable allocation of permissions by setting custom permission roles at different levels. Only users who have obtained corresponding authorization can access and use related functions.

[0085] Build a user feedback mechanism and establish a perfect user feedback mechanism to ensure that the AI ​​assistant can respond to user needs in a timely manner and continuously optimize service quality based on user feedback. The specific steps include:

[0086] Real-time response and feedback: Socket streaming output and verbatim feedback: The AI ​​assistant can respond to user instructions instantly and provide clear feedback information verbatim through socket streaming output to ensure that users understand the operation status and results.

[0087] Response optimization: The server uses multi-node deployment of container clusters to achieve load distribution, optimize concurrent response speed, and improve user experience.

[0088] User feedback collection and processing: User feedback collection channels: Establish multiple user feedback collection channels, such as collecting likes from AI assistant answers, collecting statistics on the number of times various intelligent application services are used, satisfaction questionnaire feedback, email feedback, etc., to facilitate the collection of user suggestions and user preferences for various services.

[0089] Feedback analysis: Categorize and analyze collected user feedback to identify common problems and improvement needs.

[0090] Continuous improvement: Based on feedback analysis results, we continuously optimize the AI ​​assistant's interaction and intelligent application services to improve the overall service quality.

[0091] In the interactive phase, Figure 4 and Figure 5As shown, in response to an information input operation, the information to be processed corresponding to the information input operation is obtained through the interactive portal; the information type of the information to be processed is determined through the big model in the big model service platform; based on the information type of the information to be processed, the information to be processed is processed using a corresponding processing method to obtain a processing result; wherein the processing method includes big model processing, application service processing, and information supplementation. The information type includes a first type; the processing method corresponding to the first type is application service processing; based on the information type of the information to be processed, the information to be processed is processed using a corresponding processing method to obtain a processing result, including: for the information to be processed of the first type, the application service processing method is used to process the information to obtain a processing result, including: according to the matching degree between the information to be processed and each of the application services, a target application service is selected; the information to be processed is processed using the target application service to obtain a processing result.

[0092] The information type includes a second type; the processing method corresponding to the second type is to supplement the information; based on the information type of the information to be processed, the information to be processed is processed using the corresponding processing method to obtain a processing result, including: for the information to be processed of the second type, the information to be processed is processed using an application service processing method to obtain a processing result, including: using a retrieval enhanced generation model to retrieve the information to be processed to obtain supplementary information of the information to be processed; matching based on the supplementary information to obtain reorganized prompt data after matching; using a large model to process the reorganized prompt data to obtain a processing result.

[0093] The information type includes a third type; the processing method corresponding to the second type is large model processing; based on the information type of the information to be processed, the corresponding processing method is used to process the information to be processed to obtain a processing result, including: for the information to be processed of the third type, the application service processing method is used to process it to obtain a processing result, including: using the large model to process the information to be processed to obtain a processing result.

[0094] For example, there are three types of responses for intelligent application service modules: Need to call API plug-in set processing: The large model integrated intelligent agent selects the API plug-in with the highest matching degree, extracts the request parameters from the user's question, encapsulates them into a standardized request supported by the HTTPS protocol, and sends them to the API provider. This process is based on Function Call. Based on the response result of the API, the large model integrated intelligent agent makes a final reply to the user.

[0095] Need to use RAG to retrieve supplementary knowledge data: Use RAG to retrieve supplementary knowledge data: Through the large model agent combined with RAG retrieval enhancement generation technology, use the milvus database to perform semantic retrieval of relevant knowledge data based on vector similarity. Based on the supplementary knowledge data, the large model comprehensive agent makes a final response to the user.

[0096] Based on the pre-trained data and model reasoning capabilities of the large model comprehensive intelligent agent, a final response is given to the user's question.

[0097] Furthermore, after the user of this application asks a question, the big model is first used to optimize and rewrite the question to ensure that the question is expressed more clearly and accurately. Before using the Function Call capability of the big model, a hybrid search of vectorized retrieval and keyword retrieval is performed on the data set containing the full amount of function call information, and the relevance score (Score) of each function call to the user's question is calculated. The top n functions with the highest scores (n is set by the user) are selected to re-arrange the prompt words and pass them to the big model for final judgment and calling. In this way, the problem of poor effect of directly using the Function Call capability due to the excessive total number of functions and the maximum context limit of the underlying big model is solved, and the accuracy of function calls is significantly improved.

[0098] Improved accuracy of parameter extraction: After the big model extracts the request parameters, it first determines the correctness of the request parameters based on the function parameter description and examples. If necessary parameters are missing or parameter extraction errors are found, the user is asked to add or modify the correct parameters in the form of a card. This interactive method ensures the accuracy of the request parameters and improves the success rate of function calls.

[0099] Improved understanding of tool returns: When defining request parameters and return results, a detailed description is added to the YAML file for each request and corresponding parameter, and prompts are used to supplement the parameter description. These descriptions help large models understand the tool's return parameters.

[0100] Through the detailed description of the returned parameters, the big model can better understand and process the return information of the tool, ensuring the accuracy and reliability of the response.

[0101] This application has the following advantages: One-stop intelligent service: Traditional enterprise-level applications and services are usually scattered, and users need to switch between multiple systems, which is time-consuming and laborious. The present invention integrates these services on a single platform through a global AI assistant to provide a one-stop intelligent service. Users can access and use various applications and services through only one portal, which greatly simplifies the operation process and improves work efficiency.

[0102] Intelligent interactive experience: The global AI assistant supports both voice and text interaction. Users can choose the most convenient way to communicate with the AI ​​assistant according to their habits. The large-model comprehensive intelligent agent is based on natural language processing technology, which can accurately understand the user's intentions and provide timely and accurate responses, improving the user's interactive experience.

[0103] Efficient task management: By integrating modules such as to-do task query, calendar management, and conference room reservation, the present invention provides users with an efficient task management tool. Users can easily query, add, and manage various tasks to ensure that no important matters are missed, thus improving work efficiency and organizational ability.

[0104] Powerful knowledge base question and answer function: The global AI assistant of the present invention integrates the internal and public knowledge bases of the enterprise, and uses the natural language processing capabilities of the large model and the RAG knowledge base question and answer capabilities to provide users with accurate and fast knowledge question and answer services. Users can query the required information at any time, reducing the time cost of searching and obtaining information.

[0105] Data-driven intelligent decision-making: The large-scale integrated intelligent agent can conduct in-depth analysis of user interaction data and provide data-driven intelligent decision-making support in combination with the company's internal data resources. This includes sales data analysis, customer relationship management and other functions provided by the marketing operation auxiliary module, helping companies make more informed decisions and improve business operation efficiency.

[0106] Flexible scalability: The global AI assistant adopts a modular design, and each intelligent application service module and plug-in can be flexibly expanded and customized according to the needs of the enterprise. Whether it is adding new functional modules or integrating new data sources and knowledge bases, it can be easily expanded to meet the ever-changing needs of enterprises.

[0107] Higher function call accuracy: The accuracy of function calls is improved by optimizing and rewriting user questions through preprocessing steps and using a hybrid method of vectorized search and keyword search. This improvement ensures that the large model is more intelligent and accurate in selecting functions, reduces call errors, and improves the reliability of the overall service.

[0108] Improved parameter extraction mechanism: When calling a function, the present invention performs self-checking by combining the function parameter description with examples, and actively interacts with the user when a parameter error is found, thus ensuring the correctness of the request parameters. This mechanism significantly reduces call failures caused by parameter errors, improving user experience and system stability.

[0109] Continuous improvement driven by user feedback: The present invention has established a complete mechanism for collecting and processing user feedback, which can collect and analyze user opinions and suggestions in a timely manner. Through continuous improvement of user feedback, the functions and service quality of the AI ​​assistant are continuously optimized, and a high level of user satisfaction and experience is always maintained.

[0110] Improve employee satisfaction and job satisfaction: By providing intelligent services and efficient task management tools, the present invention significantly improves employee job satisfaction. Employees can focus more on core work content, reduce tedious operation and management tasks, and improve overall work efficiency and enterprise productivity.

[0111] Figure 6 A schematic diagram of the structure of an intelligent interaction device provided in an embodiment of the present application is shown, and the device includes:

[0112] Building blocks and interaction modules;

[0113] The construction module is used to: in response to the portal setting operation, set an interactive portal in the original communication tool of the target enterprise;

[0114] In response to the application service building operation, integrating the selected plug-ins and knowledge data sets corresponding to the original multiple independent applications in the target enterprise to form multiple application services;

[0115] In response to the application service configuration operation, the application service is connected to a preset large model service platform, wherein the large model service platform is also connected to a plurality of interfaces to be selected;

[0116] The interaction module is used to: in response to an information input operation, obtain the to-be-processed information corresponding to the information input operation through the interaction entrance;

[0117] Determining the information type of the information to be processed by using the big model in the big model service platform;

[0118] Based on the information type of the information to be processed, the information to be processed is processed using a corresponding processing method to obtain a processing result; wherein the processing method includes large model processing, application service processing and information supplementation.

[0119] The information type includes a first type; the processing method corresponding to the first type is application service processing;

[0120] The processing of the information to be processed using a corresponding processing method based on the information type of the information to be processed to obtain a processing result includes:

[0121] The information to be processed of the first type is processed by using an application service processing method to obtain a processing result, including:

[0122] Selecting a target application service according to the matching degree between the information to be processed and each of the application services;

[0123] The target application service is used to process the information to be processed to obtain a processing result.

[0124] The information type includes a second type; and a processing method corresponding to the second type is to supplement the information;

[0125] The processing of the information to be processed using a corresponding processing method based on the information type of the information to be processed to obtain a processing result includes:

[0126] The information to be processed of the second type is processed by using an application service processing method to obtain a processing result, including:

[0127] Retrieve the information to be processed using a retrieval enhancement generation model to obtain supplementary information of the information to be processed;

[0128] Perform matching based on the supplementary information to obtain matched recombinant prompt data;

[0129] The reorganization prompt data is processed using a large model to obtain a processing result.

[0130] The information type includes a third type; the processing method corresponding to the second type is large model processing;

[0131] The processing of the information to be processed using a corresponding processing method based on the information type of the information to be processed to obtain a processing result includes:

[0132] The third type of information to be processed is processed by using an application service processing method to obtain a processing result, including:

[0133] The information to be processed is processed using the large model to obtain a processing result.

[0134] The construction phase also includes: constructing a rights management and security control mechanism;

[0135] The interaction stage also includes: verifying the user who inputs the information, and when the user passes the authority verification, obtaining the to-be-processed information corresponding to the information input operation through the interaction entrance.

[0136] The construction phase also includes: building a user feedback mechanism;

[0137] The interaction stage also includes: obtaining user feedback information, and optimizing the large model according to the feedback information.

[0138] The determining the information type of the information to be processed by using the big model in the big model service platform includes:

[0139] The large model is used to extract request parameters in the information to be processed, and the correctness of the request parameters is determined.

[0140] like Figure 7 As shown, an embodiment of the present application provides an electronic device for executing the intelligent interaction method in the present application, the device includes a memory, a processor, a bus, and a computer program stored in the memory and executable on the processor, wherein the steps of the intelligent interaction method are implemented when the processor executes the computer program.

[0141] Specifically, the above-mentioned memory and processor may be general-purpose memory and processor, which are not specifically limited here. When the processor runs the computer program stored in the memory, the above-mentioned intelligent interaction method can be executed.

[0142] Corresponding to the intelligent interaction method in the present application, an embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned intelligent interaction method are executed.

[0143] Specifically, the storage medium can be a general storage medium, such as a mobile disk, a hard disk, etc. When the computer program on the storage medium is executed, the above-mentioned intelligent interaction method can be executed.

[0144] In the embodiments provided in the present application, it should be understood that the disclosed systems and methods can be implemented in other ways. The system embodiments described above are merely schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, indirect coupling or communication connection of systems or units, which can be electrical, mechanical or other forms.

[0145] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, and may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0146] In addition, each functional unit in the embodiments provided in the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0147] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0148] It should be noted that similar numbers and letters represent similar items in the following figures. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. In addition, the terms "first", "second", "third", etc. are only used to distinguish the description and are not to be understood as indicating or implying relative importance.

[0149] Finally, it should be noted that the above-described embodiments are only specific implementation methods of the present application, which are used to illustrate the technical solutions of the present application, rather than to limit them. The protection scope of the present application is not limited thereto. Although the present application is described in detail with reference to the above-mentioned embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-mentioned embodiments within the technical scope disclosed in the present application, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application. They should all be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A method of intelligent interaction, characterized in that: The method comprises: a construction phase and an interaction phase; The construction phase includes: in response to the portal setting operation, setting an interactive portal in the original communication tool of the target enterprise; In response to the application service building operation, integrating the selected plug-ins and knowledge data sets corresponding to the original multiple independent applications in the target enterprise to form multiple application services; In response to the application service configuration operation, the application service is connected to a preset large model service platform, wherein the large model service platform is also connected to a plurality of interfaces to be selected; The interaction stage includes: in response to an information input operation, obtaining to-be-processed information corresponding to the information input operation through the interaction portal; Determining the information type of the information to be processed by using the big model in the big model service platform; Based on the information type of the information to be processed, the information to be processed is processed using a corresponding processing method to obtain a processing result; wherein the processing method includes large model processing, application service processing and information supplementation.

2. The method according to claim 1, characterized in that The information type includes a first type; the processing method corresponding to the first type is application service processing; The processing of the information to be processed using a corresponding processing method based on the information type of the information to be processed to obtain a processing result includes: The information to be processed of the first type is processed by using an application service processing method to obtain a processing result, including: Selecting a target application service according to the matching degree between the information to be processed and each of the application services; The target application service is used to process the information to be processed to obtain a processing result.

3. The method according to claim 1, characterized in that The information type includes a second type; and a processing method corresponding to the second type is to supplement the information; The processing of the information to be processed using a corresponding processing method based on the information type of the information to be processed to obtain a processing result includes: The information to be processed of the second type is processed by using an application service processing method to obtain a processing result, including: Retrieve the information to be processed using a retrieval enhancement generation model to obtain supplementary information of the information to be processed; Perform matching based on the supplementary information to obtain matched recombinant prompt data; The reorganization prompt data is processed using a large model to obtain a processing result.

4. The method according to claim 1, characterized in that: The information type includes a third type; the processing method corresponding to the second type is large model processing; The processing of the information to be processed using a corresponding processing method based on the information type of the information to be processed to obtain a processing result includes: The third type of information to be processed is processed by using an application service processing method to obtain a processing result, including: The information to be processed is processed using the large model to obtain a processing result.

5. The method according to claim 1, characterized in that The construction phase also includes: constructing a rights management and security control mechanism; The interaction stage also includes: verifying the user who inputs the information, and when the user passes the authority verification, obtaining the to-be-processed information corresponding to the information input operation through the interaction entrance.

6. The method according to claim 1, characterized in that The construction phase also includes: building a user feedback mechanism; The interaction stage also includes: obtaining user feedback information, and optimizing the large model according to the feedback information.

7. The method according to claim 1, characterized in that The determining the information type of the information to be processed by using the big model in the big model service platform includes: The large model is used to extract request parameters in the information to be processed, and the correctness of the request parameters is determined.

8. An intelligent interactive device, characterized in that: The device comprises: a construction module and an interaction module; The construction module is used to: in response to the portal setting operation, set an interactive portal in the original communication tool of the target enterprise; In response to the application service building operation, integrating the selected plug-ins and knowledge data sets corresponding to the original multiple independent applications in the target enterprise to form multiple application services; In response to the application service configuration operation, the application service is connected to a preset large model service platform, wherein the large model service platform is also connected to a plurality of interfaces to be selected; The interaction module is used to: in response to an information input operation, obtain the to-be-processed information corresponding to the information input operation through the interaction entrance; Determining the information type of the information to be processed by using the big model in the big model service platform; Based on the information type of the information to be processed, the information to be processed is processed using a corresponding processing method to obtain a processing result; wherein the processing method includes large model processing, application service processing and information supplementation.

9. An electronic device, characterized in that: include: A processor, a memory and a bus, wherein the memory stores machine-readable instructions executable by the processor, and when the electronic device is running, the processor and the memory communicate via the bus, and when the machine-readable instructions are executed by the processor, the steps of the intelligent interaction method as described in any one of claims 1 to 7 are performed.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the intelligent interaction method as described in any one of claims 1 to 7 are executed.

Images