Super employee artificial intelligence system

Abstract

The invention discloses a super employee artificial intelligence system which covers a question answering system based on an expert knowledge base large model, can fuse multi-modal knowledge, distributes permissions according to roles and realizes intelligent question answering and recommendation. A knowledge base management and update mechanism is provided, and uploading of files in various formats and workflow arrangement are supported; a user side and a background management side are arranged, so that different function requirements are met; and a module for realizing commercial value is also provided. The control method comprises user permission, question and answer process and dialogue process control. The data processing method relates to natural language processing, multi-modal content processing and user data analysis. The system effectively improves the knowledge acquisition efficiency, meets individual requirements, optimizes knowledge management, realizes commercial value, enhances enterprise competitiveness, and is suitable for the fields of human resource management and the like.

Description

Technical Field

[0001] The present invention relates to the technical field of the integration of human resource management and artificial intelligence technology, and particularly to a super employee artificial intelligence system. Background Art

[0002] In the current era of rapid digital and intelligent development, the field of human resource management is undergoing profound changes, and the demand for efficient, accurate, and personalized services is becoming increasingly prominent. However, the traditional human resource service model has many limitations in meeting the needs of different levels of HRs, which are specifically manifested as follows:

[0003] For junior small business HRs, limited by the enterprise scale and resources, they face severe career development dilemmas. On the one hand, lacking systematic and customized learning resources, it is difficult to specifically improve professional skills. For example, in key areas such as salary system design and employee performance evaluation, they can only rely on scattered online materials or limited industry exchanges to obtain knowledge, and the systematicness and accuracy of the knowledge are difficult to guarantee. On the other hand, when encountering problems in actual work, lacking accurate problem-solving channels, they often can only rely on their own limited experience or consult peers, with low efficiency and the professionalism of the answers being doubtful, which greatly hinders the improvement of their professional levels.

[0004] Senior medium and large-sized HRs, although having relatively rich resources, also have inconveniences in solving enterprise management problems and obtaining experience. As the enterprise scale expands and the management complexity increases, they need to handle complex problems such as organizational structure optimization and cross-cultural team management. However, most of the existing knowledge acquisition channels are general-purpose, lacking professional guidance for the specific enterprise situation and their own management styles. At the same time, lacking an efficient communication platform and in-depth analysis tools for obtaining industry front-line experience and best practices, they often feel powerless when dealing with complex management challenges.

[0005] As industry senior experts, 40 - 50-year-old expert consultants, although having rich practical experience, have deficiencies in transforming experience into reusable methodologies and tools. When providing consulting services to enterprises, they lack a set of standardized and systematic tools to support their consulting processes, and more rely on personal experience and subjective judgment. This not only affects the efficiency and quality of consulting services but also restricts the inheritance and innovation of industry knowledge.

[0006] In summary, the deficiencies of traditional human resource services in meeting the needs of different levels of HRs are becoming increasingly prominent, and there is an urgent need for an intelligent and personalized HRAI platform to integrate high-quality resources, provide accurate services, and promote the innovative development of the human resource management industry. Summary of the Invention

[0007] The objective of the present invention is to provide a super employee artificial intelligence system, aiming to solve the following problems:

[0008] Improve the efficiency and accuracy of knowledge acquisition: Meet the professional knowledge needs of different levels of HR (beginner, advanced, expert), and change the current situation of low knowledge acquisition efficiency.

[0009] Realize personalized knowledge recommendation and Q&A: Provide accurate knowledge recommendation and Q&A services according to user roles and needs, and meet personalized requirements.

[0010] Optimize knowledge base management and update: Overcome the problems of untimely update and fragmented knowledge in traditional knowledge bases, and promote efficient knowledge sharing within the enterprise.

[0011] Support multi-modal knowledge presentation: Present knowledge in various forms such as text, video, PPT, etc., to meet the diverse learning needs of users.

[0012] Enhance user experience and interactivity: Improve the defects in interactivity and interface friendliness of the existing system, and provide a smooth user experience.

[0013] Realize the commercial value of the knowledge base: Integrate the expert knowledge base and realize the commercial application of knowledge, such as paid Q&A, point redemption, etc.

[0014] The present invention is achieved through the following technical solutions:

[0015] A super employee artificial intelligence system, including:

[0016] A Q&A system based on the expert knowledge base large model, including:

[0017] A multi-modal knowledge fusion module, which integrates text, video, and PPT knowledge forms to provide rich knowledge presentation methods for users to learn in different scenarios

[0018] A role and permission management module, which assigns knowledge base access permissions according to the roles of beginner HR, advanced HR, and expert consultants to realize personalized knowledge recommendation and Q&A services.

[0019] An intelligent Q&A and recommendation module, which uses natural language processing technology to realize intelligent Q&A and keyword recommendation, and quickly matches and recommends knowledge content related to the user's input questions.

[0020] A questionnaire survey and intelligent recommendation module, which evaluates the user's knowledge level through questionnaire surveys and recommends courses and knowledge modules to improve learning effects.

[0021] A knowledge base management and update mechanism, including:

[0022] Knowledge Base Creation and Upload Module, which supports the upload of knowledge base files in txt, docx, and pdf file formats, facilitating enterprises to quickly build and update knowledge bases.

[0023] Question Answering Workflow Orchestration and Debugging Module, which provides the function of AI question answering workflow orchestration. Administrators can adjust the question answering logic according to needs to optimize the question answering effect of the knowledge base.

[0024] User Side and Backend Management Side, including:

[0025] User Side, which is equipped with an intelligent assistant homepage, knowledge base center, digital human creation, and personal center modules, providing intelligent question answering, knowledge base management, and integral exchange functions.

[0026] Backend Management Side, which provides knowledge base management and account management functions, facilitating enterprise administrators to flexibly configure knowledge base content and user permissions.

[0027] Commercial Value Realization Module, including:

[0028] Paid Question Answering and Integral Exchange Module, which sets up a trial and paid question answering mechanism with token limit and times limit, and establishes a paid question answering mechanism to realize the commercial value of the knowledge base. At the same time, it motivates users to learn through the integral exchange function.

[0029] Course Recommendation and Video Preview Function Module, which provides a video preview function when recommending courses, enhancing the user experience and promoting knowledge consumption.

[0030] Further as an improvement of the technical solution of the present invention, the system supports docking and integration with the enterprise's existing office software OA system and CRM system to achieve seamless connection of data interaction and business processes.

[0031] Further as an improvement of the technical solution of the present invention, the system includes a knowledge graph construction module, which constructs a knowledge graph in the field of human resources by analyzing and mining the content of the knowledge base, and displays the knowledge association relationship in a visual form to assist users in understanding and retrieving knowledge.

[0032] Further as an improvement of the technical solution of the present invention, the system also includes a data backup and recovery module, which regularly backs up the knowledge base data, user data, and operation logs in the system, and can quickly recover the data when the data is lost or damaged.

[0033] Further as an improvement of the technical solution of the present invention, a control method for a super employee artificial intelligence system includes:

[0034] User Permission Control, including:

[0035] Role permission allocation steps, allocate roles of employees, enterprise administrators, and general administrators of the rice warehouse according to the enterprise organizational structure and limit module access permissions;

[0036] Integral system control steps, formulate integral deduction logic, and build the enterprise-side package purchase and personal-side integral recharge processes;

[0037] Q&A process control, including:

[0038] Dynamic knowledge base matching steps, use algorithms to analyze user role information and match knowledge base content;

[0039] Multi-modal answer generation steps, preferentially call content from the expert library, secondly select the global knowledge base, and finally rely on GPT public domain data and mark the source;

[0040] Free trial and unlocking logic steps, provide chapter-by-chapter free trials for videos and PPTs, preview the table part, and trigger the payment process when the integral is insufficient;

[0041] Dialogue process control, including:

[0042] Third-party payment interface integration steps, integrate with mainstream third-party payment platforms, support enterprise package purchases and personal integral recharges, return to the integral exchange page after payment, and encrypt the payment process;

[0043] Negative review feedback mechanism steps, require mandatory feedback for 1-2 star negative reviews and trigger customer service follow-up visits, and optional for reviews above 3 stars.

[0044] Further as an improvement of the technical solution of the present invention, in the user permission control, there is a dynamic permission adjustment mechanism, which can regularly or real-time adjust the role permissions of users according to the user's behavior data and work performance factors.

[0045] Further as an improvement of the technical solution of the present invention, the Q&A process control further includes a credibility evaluation step for Q&A results, which scores the generated answers according to factors such as answer sources and matching degrees and displays them for users to refer to.

[0046] Further as an improvement of the technical solution of the present invention, a data processing method for a super employee artificial intelligence system includes:

[0047] Natural language processing NLP, including:

[0048] Intention recognition and Q&A matching steps, use deep learning technology to analyze the semantics of user questions, retrieve answers in combination with role permissions, and sort and screen;

[0049] Keyword recommendation steps, use data mining algorithms to generate high-frequency question keywords based on user roles and historical behavior data;

[0050] Multimodal content processing, including:

[0051] File format support step, supporting the upload and storage of txt, docx, and pdf file formats;

[0052] Video / PPT generation step, constructing a digital human creation process, supporting local video upload, and using speech synthesis and image processing technologies to generate short videos with page-by-page voiceovers for PPT and combining them with green screen materials for synthesis;

[0053] User data analysis, including:

[0054] Evaluation data collection step, collecting user ratings and feedback, establishing an evaluation database, and analyzing and optimizing the recommendation algorithm;

[0055] Questionnaire survey analysis step, designing a questionnaire to collect user information through multiple-choice questions, converting it into an ability radar chart, and recommending customized learning courses.

[0056] Furthermore, as an improvement to the technical solution of the present invention, in natural language processing NLP, there are semantic expansion and disambiguation processing steps for fuzzy problems, converting fuzzy problems into clear semantic expressions before retrieving answers.

[0057] Furthermore, as an improvement to the technical solution of the present invention, user data analysis also includes an analysis step of the user's learning path. According to the user's learning trajectory in the system, a personalized learning progression route is planned for the user.

[0058] In summary, the present invention has the following beneficial effects:

[0059] Improve knowledge acquisition efficiency: With the help of intelligent Q&A and personalized recommendations, users can quickly obtain the knowledge they need and improve work efficiency.

[0060] Enhance user experience: Support multimodal knowledge presentation, provide rich and diverse learning resources, meet the diverse needs of users, and improve user satisfaction.

[0061] Optimize knowledge management: The knowledge base management function can achieve rapid updates and flexible configuration, ensuring the timeliness and accuracy of the knowledge base content, and promoting internal knowledge sharing within the enterprise.

[0062] Realize the commercial value of knowledge: Through mechanisms such as paid Q&A and points redemption, convert the knowledge base into commercial resources and create economic value for the enterprise.

[0063] Enhance enterprise competitiveness: Provide professional knowledge bases and intelligent Q&A services for enterprises, help enterprises improve their human resource management capabilities, and enhance their market competitiveness.

[0064] Support multi - terminal applications: In the future, it will support multiple terminals such as the PC terminal and the mobile terminal (such as the DingTalk ecosystem), which is convenient for users to use anytime and anywhere, expanding the application scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] By reading the following detailed description of the non - restrictive embodiments with reference to the accompanying drawings, other features, objects, and advantages of the present invention will become more apparent:

[0066] Figure 1 It is a schematic structural diagram of a super - employee artificial intelligence system according to an embodiment of the present invention;

[0067] Figure 2 It is a business flow chart of a super - employee artificial intelligence system according to an embodiment of the present invention;

[0068] Figure 2-1 It is one of the business flow charts of a super - employee artificial intelligence system according to an embodiment of the present invention;

[0069] Figure 2-2 It is another business flow chart of a super - employee artificial intelligence system according to an embodiment of the present invention;

[0070] Figure 2-3 It is the third business flow chart of a super - employee artificial intelligence system according to an embodiment of the present invention;

[0071] Figure 2-4 It is the fourth business flow chart of a super - employee artificial intelligence system according to an embodiment of the present invention;

[0072] Figure 2-5 It is the fifth business flow chart of a super - employee artificial intelligence system according to an embodiment of the present invention;

[0073] Figure 3 It is a schematic diagram of a control method of a super - employee artificial intelligence system according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0074] To make the objectives, features, and advantages of the present application more obvious and understandable, the technical solutions in the embodiments of the present application are clearly and completely described. Obviously, the embodiments described below are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of the present application.

[0075] Referring to Figure 1 and Figure 2 , a super - employee artificial intelligence system includes:

[0076] A question - answering system based on an expert knowledge base large model, including:

[0077] Multimodal Knowledge Fusion Module, which integrates text, video, and PPT knowledge forms to provide users with rich knowledge presentation methods for learning in different scenarios.

[0078] Role and Permission Management Module, which assigns knowledge base access permissions according to the roles of junior HR, senior HR, and expert consulting advisors, and realizes personalized knowledge recommendation and Q&A services.

[0079] Intelligent Q&A and Recommendation Module, which uses natural language processing technology to realize intelligent Q&A and keyword recommendation, and quickly matches and recommends knowledge content related to the user's input questions.

[0080] Questionnaire Survey and Intelligent Recommendation Module, which evaluates the user's knowledge level through questionnaire surveys and recommends courses and knowledge modules to improve learning effects.

[0081] Knowledge Base Management and Update Mechanism, including:

[0082] Knowledge Base Creation and Upload Module, which supports the upload of knowledge base files in txt, docx, and pdf file formats, facilitating enterprises to quickly build and update knowledge bases.

[0083] Q&A Workflow Orchestration and Debugging Module, which provides AI Q&A workflow orchestration functions. Administrators can adjust Q&A logic according to needs to optimize the Q&A effect of the knowledge base.

[0084] User Side and Back-end Management Side, including:

[0085] User Side, which is equipped with an intelligent assistant homepage, knowledge base center, digital human creation, and personal center modules, providing intelligent Q&A, knowledge base management, and integral exchange functions;

[0086] Back-end Management Side, which provides knowledge base management and account management functions, facilitating enterprise administrators to flexibly configure knowledge base content and user permissions.

[0087] Business Value Realization Module, including:

[0088] Paid Q&A and Integral Exchange Module, which sets up trial and paid Q&A mechanisms limited by tokens and by times, and establishes a paid Q&A mechanism to realize the business value of the knowledge base. At the same time, it motivates users to learn through the integral exchange function.

[0089] Course Recommendation and Video Preview Function Module, which provides video preview functions when recommending courses, enhances the user experience, and promotes knowledge consumption.

[0090] Specifically, in the solution of this embodiment, the system supports docking and integration with the enterprise's existing office software OA system and CRM system to achieve seamless data interaction and business process connection.

[0091] Specifically, in the solution of this embodiment, the system includes a knowledge graph construction module, which constructs a knowledge graph in the human resources field by analyzing and mining the content of the knowledge base, displays the knowledge association relationship in a visual form, and assists users in understanding and retrieving knowledge.

[0092] Specifically, in the solution of this embodiment, the system further includes a data backup and recovery module, which regularly backs up the knowledge base data, user data, and operation logs in the system, and can quickly recover the data when the data is lost or damaged.

[0093] Refer to Figure 3 , specifically, in the solution of this embodiment, a control method for a super employee artificial intelligence system includes:

[0094] User permission control, including:

[0095] Role permission assignment step, which assigns roles of employees, enterprise administrators, and rice warehouse general administrators according to the enterprise organizational structure and limits module access permissions;

[0096] Integral system control step, formulating integral deduction logic and building enterprise-side package purchase and personal-side integral recharge processes;

[0097] Question and answer process control, including:

[0098] Dynamic knowledge base matching step, using algorithms to analyze user role information and match the knowledge base content;

[0099] Multi-modal answer generation step, preferentially calling the content of the expert library, secondly selecting the global knowledge base, and finally leveraging GPT public domain data and marking the source;

[0100] Trial view and unlocking logic step, providing chapter-by-chapter trial views for videos and PPTs, previewing table parts, and triggering the payment process when the integral is insufficient;

[0101] Conversation process control, including:

[0102] Third-party payment interface integration step, integrating with mainstream third-party payment platforms, supporting enterprise package purchases and personal integral recharges, returning to the integral exchange page after payment and encrypting the payment process;

[0103] Negative review feedback mechanism step, requiring mandatory feedback and triggering customer service follow-up for 1-2 star negative reviews, and optional filling for reviews above 3 stars.

[0104] Specifically, in the user permission control of this embodiment, there is a mechanism for dynamically adjusting permissions, which can regularly or real-time adjust the role permissions of users according to their behavior data and work performance factors.

[0105] Specifically, in the solution of this embodiment, the Q&A process control further includes a step of evaluating the credibility of the Q&A result, scoring the generated answer according to the answer source and matching degree factors, and presenting it to the user for reference.

[0106] Specifically, in the solution of this embodiment, a data processing method for a super employee artificial intelligence system includes:

[0107] Natural language processing NLP, including:

[0108] An intention recognition and Q&A matching step, which uses deep learning technology to analyze the semantics of the user's question, retrieves answers in combination with role permissions, and sorts and filters them;

[0109] A keyword entry recommendation step, which generates high-frequency question entries based on the user's role and historical behavior data using data mining algorithms;

[0110] Multimodal content processing, including:

[0111] A file format support step, which supports the upload and storage of txt, docx, and pdf file formats;

[0112] A video / PPT generation step, which constructs a digital human creation process, supports local video upload, and uses speech synthesis and image processing technologies to generate a short video with page-by-page voiceover for PPT and combines it with green screen materials for synthesis;

[0113] User data analysis, including:

[0114] An evaluation data collection step, which collects user ratings and feedback, establishes an evaluation database, and analyzes and optimizes the recommendation algorithm;

[0115] A questionnaire survey analysis step, which designs a questionnaire to collect user information through multiple-choice questions, converts it into an ability radar chart, and recommends customized learning courses.

[0116] Specifically, in the natural language processing NLP of this embodiment, for fuzzy questions, there are semantic expansion and disambiguation processing steps, which convert the fuzzy questions into clear semantic expressions before retrieving answers.

[0117] Specifically, in this embodiment's solution, user data analysis further includes an analysis step of the user's learning path. According to the user's learning trajectory in the system, a personalized learning progression route is planned for the user.

[0118] Embodiment:

[0119] User permission control

[0120] Role Permission Allocation: Deeply analyze the enterprise organizational structure and precisely divide different roles such as employees, enterprise administrators, and general administrators of the knowledge repository. For different roles, based on the principle of least privilege, strictly limit their access rights to each module of the system. For example, HR can only access the HR module knowledge base related to human resource management to ensure the security and privacy of system data, prevent information leakage and unauthorized operations.

[0121] Integral System Control: Develop a rigorous integral deduction logic. Clearly stipulate that after a user views specific content on a trial basis and needs to unlock the full content, the corresponding integral should be deducted. At the same time, build a convenient and secure recharge process, and design different recharge methods for the enterprise side and the personal side respectively. The enterprise side can obtain a large number of integral at one time through package purchases to meet the overall knowledge consumption needs of the enterprise; the personal side provides flexible integral recharge channels and supports multiple payment methods to facilitate individual users to recharge according to their own needs.

[0122] Q&A Process Control

[0123] Dynamic Knowledge Repository Matching: Use advanced algorithms to analyze user role information in real time. According to the specific role of the user, such as a compensation specialist, accurately match the corresponding knowledge repository content. Ensure that when a user asks a question, the system can quickly retrieve relevant knowledge from the expert library of the compensation module and provide answers with strong pertinence and high professionalism, avoiding interference from irrelevant information.

[0124] Multi-modal Answer Generation: Establish a priority strategy. When a user asks a question, the system first calls multi-modal knowledge content such as videos, PPTs, and tool sheets from the expert library for answering. If there is no matching content in the expert library, it will secondarily search the global knowledge repository. When the global knowledge repository cannot provide a satisfactory answer, use GPT public domain data for supplementation, but the data source will be clearly marked to ensure the reliability and traceability of the answer.

[0125] Trial Viewing and Unlock Logic: For video and PPT knowledge resources, provide a trial viewing function in a chapter-by-chapter display method so that users can have a preliminary understanding of the content before paying to unlock it. For tabular content, partial preview is supported to display key information. When a user's integral is insufficient, the system automatically triggers the payment process to guide the user to obtain the full content by recharging integral or purchasing paid Q&A services.

[0126] Dialogue Process Control

[0127] Third-party payment interface integration: Deep integration with mainstream third-party payment platforms, supporting two business models: enterprise package purchase (To B) and personal points recharge (To C). After the user completes the payment, the system automatically returns the user to the points redemption page, facilitating the user to promptly convert the payment amount into points for subsequent knowledge consumption. Meanwhile, the entire payment process is encrypted to ensure the security of the user's payment information.

[0128] Negative review feedback mechanism: Establish a strict evaluation and feedback system. For one- or two-star negative reviews given by users, the system compulsorily requires the user to fill in the feedback content, elaborating on the reasons for dissatisfaction. Once such negative reviews are received, the customer service follow-up process is immediately triggered, arranging professional customer service personnel to communicate with the user, understand the specific situation, and solve the problem. For reviews above three stars, it is set as an optional item, encouraging users to share their usage experiences and suggestions for continuous optimization of the system service.

[0129] Data processing methods

[0130] Natural Language Processing (NLP)

[0131] Intent recognition and question-answer matching: Adopt deep learning technology to perform semantic analysis and intent recognition on the questions input by users. Combining with the user's role permissions, conduct accurate retrieval from the enterprise knowledge base or expert database, and quickly match the most relevant answers. Meanwhile, sort and screen the retrieval results to ensure that the answers provided to users are accurate and useful.

[0132] Keyword entry recommendation: Based on the user's role and historical behavior data, use data mining algorithms to analyze the user's interest preferences and knowledge needs. Generate high-frequency question entries, such as a common question bank for HRBPs, to facilitate users to quickly find the answers to relevant questions and improve the efficiency of knowledge acquisition.

[0133] Multimodal content processing

[0134] File format support: Currently, the knowledge base supports the upload and storage of multiple common file formats such as.txt,.docx,.pdf, etc., meeting the diverse knowledge entry needs of enterprises. In the future, it is planned to introduce advanced Optical Character Recognition (OCR) technology to realize the recognition and processing of knowledge content in picture format, further enriching the content form of the knowledge base.

[0135] Video / PPT generation: Build a digital human creation process to support users to upload local videos. Meanwhile, utilize voice synthesis and image processing technologies to generate short videos with voiceovers for each page of the PPT and combine them with green screen materials for synthesis, providing users with diverse ways of knowledge creation and display.

[0136] User data analysis

[0137] Evaluation data collection: Collect users' ratings (1 - 5 stars) and feedback on knowledge content and services in real time, and establish a user evaluation database. Use data analysis algorithms to deeply mine the evaluation data, analyze users' satisfaction and pain points of needs. According to the analysis results, optimize the recommendation algorithm, and give priority to displaying the TOP10 content with a high positive review rate to improve users' trust in the system and usage frequency.

[0138] Questionnaire survey analysis: Design a scientific and reasonable questionnaire to collect information such as users' knowledge levels and hobbies through multiple-choice questions. Use data analysis tools to convert the collected data into a user ability radar chart to visually display users' knowledge structures and ability short - boards. Based on the analysis results of the radar chart, recommend customized learning courses for users. The course forms include video explanations and knowledge cards to meet users' personalized learning needs.

[0139] Actual cases:

[0140] Role - permission allocation: Take a large - scale manufacturing enterprise as an example. The enterprise has multiple departments such as production, sales, R & D, and human resources. After introducing the Super Employee AI system, the system deeply analyzes the enterprise's organizational structure and accurately divides roles such as ordinary employees, department heads, enterprise administrators, and Mica Warehouse general administrators. Among them, HR in the human resources department can only access the knowledge bases of HR modules related to human resources management, such as recruitment, training, salary and benefits, etc. For example, the recruitment specialist, Xiao Li, can only view relevant knowledge such as recruitment processes, interview skills, and talent market analysis, and cannot access the core content of salary design. This effectively prevents information leakage and unauthorized operations, and ensures the security and privacy of system data.

[0141] Integral system control: When an Internet startup company uses the system, it adopts an integral system to manage knowledge consumption. The company stipulates that after an employee views the first two chapters of the advanced data analysis course, if they want to continue learning the subsequent chapters, 50 points need to be deducted. For the enterprise side, the company purchases an integral package worth 5000 yuan at one time and obtains 5000 points. These points can be used by all employees of the company within a certain period to meet the overall knowledge learning needs of the enterprise. For individual users, such as employee Wang, he can flexibly recharge points through various payment methods such as Alipay and WeChat to meet his learning needs for specific knowledge.

[0142] Question - answering process control

[0143] Dynamic Knowledge Base Matching: Suppose Xiao Zhang, a compensation specialist in a financial enterprise, asks a question about the calculation method of year-end bonuses. The system uses advanced algorithms to analyze Xiao Zhang's compensation specialist role information in real time and accurately matches it to the expert library of the compensation module. The system quickly retrieves relevant knowledge from the expert library, including year-end bonus calculation rules, different position coefficient settings, etc., and quickly gives a highly targeted and professional answer, avoiding interference from other irrelevant information.

[0144] Multi-modal Answer Generation: When a manager of an education and training institution asks how to optimize the training course system, the system first calls relevant training course optimization case videos, PPTs, and corresponding tool sheets from the expert library to answer. If there is no matching content in the expert library, the system secondly searches the global knowledge base. For example, it finds the successful experiences and lessons learned from the course optimization of other educational institutions in the global knowledge base. If the global knowledge base also cannot provide a satisfactory answer, the system supplements it with GPT public domain data and clearly marks the data source. For example, it marks "This part of the content is from GPT public domain data for reference only" to ensure the reliability and traceability of the answer.

[0145] Preview and Unlock Logic: For an online vocational skills training platform, the video courses on the platform adopt a chapter-by-chapter preview method. After a user previews the first two chapters of a project management course and wants to unlock the subsequent chapters, when the user's points are insufficient, the system automatically pops up a payment process page. The user can choose to recharge points or directly purchase a paid Q&A service to obtain the complete course content. For the table-type project progress management template, the user can preview the table header and some key data to understand the general structure and main content of the template. If the user wants to obtain the complete template, it needs to be unlocked with points.

[0146] Dialogue Process Control

[0147] Third-party Payment Interface Integration: When a chain catering enterprise uses the system, by integrating with mainstream third-party payment platforms (such as WeChat Pay and Alipay), it supports corporate package purchases (To B) and individual point recharges (To C). The enterprise side purchases an annual package worth 10,000 yuan at one time for employees' training and learning and knowledge acquisition. After the payment is completed, the system automatically returns the enterprise administrator to the point exchange page, facilitating the administrator to exchange the purchased amount for points and allocate them to employees for use. At the same time, the system encrypts the entire payment process to ensure the security of the enterprise's payment information.

[0148] Negative review feedback mechanism: After using the system, an online education platform established a strict evaluation and feedback system. When user Li gave a one-star negative review to the purchased human resources course, the system forced Li to fill in the feedback content. Li reported that the course content was outdated, with cases from years ago and out of touch with the current market. After receiving the feedback, the platform customer service immediately called Li back to understand the details and reported the problem to the course production team. The course production team updated and optimized the course, re-recorded the videos, and added the latest industry cases to improve the course quality. For reviews above three stars, users can choose to fill in their usage experiences and suggestions. For example, user Wang evaluated that the course content was rich and the explanations were clear, and suggested adding some practical operation cases. The platform continuously optimized the course content and services based on the users' suggestions.

[0149] Data processing method

[0150] Natural Language Processing (NLP)

[0151] Intent recognition and question-answer matching: An employee of an e-commerce company asked in the system "How to improve the store conversion rate". The system used deep learning technology to perform semantic analysis and intent recognition on the question. Combining with the employee's e-commerce operation role permissions, the system retrieved accurately from the enterprise knowledge base and the e-commerce domain expert database. It quickly matched relevant answers including optimizing product detail pages, adjusting promotion strategies, and improving customer service quality, and sorted and filtered them according to relevance and accuracy to provide the most useful answers to the employee.

[0152] Keyword entry recommendation: Taking Xiao Zhao, an HRBP of a technology company, as an example, the system analyzed based on Xiao Zhao's role and historical behavior data using data mining algorithms and found that Xiao Zhao often checked knowledge in aspects such as recruitment and employee relations management. The system generated high-frequency question entries such as "How to recruit high-end technical talents" and "How to handle employee resignation disputes" to facilitate Xiao Zhao to quickly find answers to relevant questions and improve the efficiency of knowledge acquisition.

[0153] Multimodal content processing

[0154] File format support: When building a knowledge base, a multinational company uploaded a large number of files in.txt format such as employee manuals,.docx format such as project reports, and.pdf format such as industry research reports, meeting the diverse knowledge entry needs of the enterprise. In the future, the enterprise plans to introduce OCR technology to realize the recognition and processing of knowledge content in picture formats such as product instruction pictures and contract pictures, further enriching the content forms of the knowledge base. For example, converting product instruction pictures into editable text for convenient retrieval and use by employees.

[0155] Video / PPT Generation: Zhang, an employee of an advertising creative company, uploaded the advertising creative video materials shot locally through the systematic digital human creation process. At the same time, using the system's speech synthesis and image processing technologies, he generated short videos by dubbing each page of the PPT and combined them with green screen materials to produce a creative advertising video for the company's project display and customer promotion, providing users with diverse ways of knowledge creation and display.

[0156] User Data Analysis

[0157] Evaluation Data Collection: An online learning platform collects the ratings (1-5 stars) and feedback from users on course content and services in real time. Through data analysis algorithms, it is found that users' satisfaction with the Python programming course is relatively low, and the main pain points are too few practical cases and overly boring theoretical explanations. Based on the analysis results, the platform optimizes the recommendation algorithm, reduces the recommendation weight of this course, and at the same time optimizes the course by adding more practical cases and interactive sessions. It also gives priority to displaying the top 10 courses with high positive review rates, such as the data analysis practical course, improving users' trust and usage frequency of the system.

[0158] Questionnaire Survey Analysis: An enterprise management consulting company designed a scientific and reasonable questionnaire to collect information such as users' knowledge levels and hobbies through multiple-choice questions. Using data analysis tools, the collected data is converted into a user ability radar chart. For example, it is found that users are relatively weak in strategic planning and financial management. Based on the analysis results of the radar chart, customized learning courses are recommended for users, including video explanations of strategic planning and knowledge cards on financial management, meeting users' personalized learning needs.

[0159] In summary, compared with the prior art, the present invention has the following beneficial effects:

[0160] Improve Knowledge Acquisition Efficiency: With the help of intelligent Q&A and personalized recommendations, users can quickly obtain the knowledge they need and improve work efficiency.

[0161] Enhance User Experience: Support multi-modal knowledge presentation, provide rich and diverse learning resources, meet users' diverse needs, and improve user satisfaction.

[0162] Optimize Knowledge Management: The knowledge base management function can achieve rapid updates and flexible configurations, ensuring the timeliness and accuracy of the knowledge base content and promoting internal knowledge sharing within the enterprise.

[0163] Realize the Commercial Value of Knowledge: Through mechanisms such as paid Q&A and points redemption, convert the knowledge base into commercial resources and create economic value for the enterprise.

[0164] Enhance enterprise competitiveness: Provide professional knowledge bases and intelligent Q&A services for enterprises, helping enterprises improve their human resource management capabilities and enhance their market competitiveness.

[0165] Support multi-terminal applications: In the future, it will support multiple terminals such as the PC side and the mobile side (such as the DingTalk ecosystem), facilitating users to use it anytime and anywhere and expanding the application scenarios.

[0166] The above has introduced in detail the technical solutions provided by the embodiments of the present invention. Specific examples are used herein to elaborate on the principles and implementation manners of the embodiments of the present invention. The descriptions of the above embodiments are only applicable to helping understand the principles of the embodiments of the present invention; at the same time, for those of ordinary skill in the art, according to the embodiments of the present invention, there will be changes in the specific implementation manners and application scopes. In summary, the content of this specification should not be construed as a limitation on the present invention.

Claims

1. A super employee artificial intelligence system, characterized in that: include: The question-answering system based on the expert knowledge base model includes: Multimodal knowledge fusion module, integrating text, video, and PPT knowledge forms; The role and permission management module allocates knowledge base access rights based on the roles of entry-level HR, senior HR, and expert consultant; Intelligent question-answering and recommendation module, which uses natural language processing technology to realize intelligent question-answering and keyword recommendation; questionnaire survey and intelligent recommendation module, which evaluates user knowledge level through questionnaire survey and recommends courses and knowledge modules; Knowledge base management and update mechanism, including: The knowledge base creation and upload module supports uploading knowledge base files in txt, docx, and pdf file formats; the question-and-answer workflow orchestration and debugging module provides AI question-and-answer workflow orchestration functions; The user end and the backend management end include: The user end has the smart assistant homepage, knowledge base center, digital human creation, and personal center modules, providing smart question and answer, database management, and points redemption functions; The backend management terminal provides knowledge base management and account management functions; Business value realization module, including: Paid Q&A and points redemption module, setting up token restrictions, trial restrictions by number of times, and paid Q&A mechanisms; The course recommendation and trial viewing function module provides a video trial viewing function when recommending courses.

2. A super employee artificial intelligence system according to claim 1, characterized in that: The system supports integration with the company's existing office software OA system and CRM system to achieve seamless connection between data interaction and business processes.

3. A super employee artificial intelligence system according to claim 1, characterized in that: The system includes a knowledge graph construction module, which constructs a knowledge graph in the field of human resources by analyzing and mining the content of the knowledge base, displays knowledge associations in a visual form, and assists users in understanding and retrieving knowledge.

4. A super employee artificial intelligence system according to claim 1, characterized in that: The system also includes a data backup and recovery module, which regularly backs up the knowledge base data, user data, and operation logs in the system, and can quickly restore data when the data is lost or damaged.

5. A control method for the super employee artificial intelligence system according to any one of claims 1 to 4, characterized in that: include: User permission control, including: Steps for allocating roles and permissions: assign employee, enterprise administrator, and rice warehouse administrator roles according to the enterprise organizational structure and limit module access permissions; Control steps of the points system, formulate logic for points deduction, and build the process of enterprise package purchase and personal points recharge; Question and answer process control, including: Dynamic knowledge base matching step, using algorithms to analyze user role information to match knowledge base content; In the step of generating multimodal answers, the expert database is used first, followed by the global knowledge base, and finally GPT public domain data is used and the source is marked; Preview and unlock logic steps, preview videos and PPT chapters, preview tables, and trigger payment process when points are insufficient; Dialogue flow control, including: The third-party payment interface integration steps are integrated with mainstream third-party payment platforms to support enterprise package purchases and personal points recharge. After payment, the points redemption page is returned and the payment process is encrypted; The steps of the negative review feedback mechanism are that it is mandatory to fill in the feedback for 1-2 star negative reviews and trigger a customer service return visit, while it is optional for reviews above 3 stars.

6. The control method of a super employee artificial intelligence system according to claim 5 is characterized in that: In the user authority control, there is a dynamic authority adjustment mechanism, which can adjust the user's role authority regularly or in real time according to the user's behavior data and work performance factors.

7. The control method of a super employee artificial intelligence system according to claim 5 is characterized in that: The question-and-answer process control also includes a credibility assessment step for the question-and-answer results, where the credibility of the generated answers is scored based on the answer source and matching factors, and displayed to the user for reference.

8. A data processing method for a super employee artificial intelligence system, characterized in that: include: Natural Language Processing (NLP), including: In the step of intent recognition and question-answer matching, deep learning technology is used to analyze the semantics of user questions and retrieve answers and sort and filter them in combination with role permissions; Keyword recommendation step: Generate high-frequency question terms using data mining algorithms based on user roles and historical behavior data; Multimodal content processing, including: File format support steps: support txt, docx, pdf file formats for uploading and storage; Video / PPT generation steps, building a digital human creation process, supporting local video uploads, using speech synthesis and image processing technology to achieve PPT page dubbing to generate short videos and combine them with green screen material synthesis; User data analysis, including: Evaluation data recovery step, collecting user ratings and feedback, establishing an evaluation database and analyzing and optimizing the recommendation algorithm; Questionnaire survey analysis steps: design a questionnaire to collect user information through multiple-choice questions, convert it into a capability radar chart and recommend customized learning courses.

9. The data processing method of a super employee artificial intelligence system according to claim 8 is characterized in that: In natural language processing (NLP), there are semantic expansion and disambiguation steps for fuzzy questions, which convert fuzzy questions into clear semantic expressions before answer retrieval.

10. The control method of a super employee artificial intelligence system according to claim 8, characterized in that: User data analysis also includes the analysis of user learning paths, and plans personalized learning advancement routes for users based on their learning trajectories in the system.

Images