An intelligent office integration system and method supporting cross-device collaboration

Abstract

This invention discloses an integrated intelligent office system and method supporting cross-device collaboration. The method balances security and convenience by maintaining session-level face verification status locally on the mobile terminal and dynamically controlling the verification frequency based on configurable thresholds. It provides integrated functions for document signing, task management, meeting management, AI Q&A, and PPT generation. It enhances the accuracy of intelligent Q&A by employing a two-stage retrieval enhancement generation technology combined with document metadata filtering. Furthermore, it utilizes digital signatures and intelligent conflict resolution strategies to ensure the consistency and reliability of cross-device data synchronization. The system is deployed collaboratively on mobile terminals, servers, and PCs, enabling real-time sharing of operational data and collaborative work, effectively solving the problems of fragmented functions, difficulties in collaboration, and insufficient intelligence in traditional office systems.

Description

Technical Field

[0001] This invention relates to the technical field of smart office, and in particular to the technical field of an integrated smart office system and method that supports cross-device collaboration. Background Technology

[0002] With the rapid development of mobile internet and artificial intelligence technologies, smart office has become an important direction for enterprise digital transformation. Traditional office systems suffer from the following problems: (1) Difficulty in cross-device collaboration: After a user operates on the PC, the mobile terminal cannot view the latest data in real time and needs to log in again and refresh the page. Data synchronization is delayed, resulting in inconsistency between PC and mobile terminal data. Existing technologies lack a unified data storage and synchronization mechanism, making it impossible to achieve true cross-device collaborative work. (2) Dispersed functions: Different office functions are scattered in different applications, and users need to switch between multiple applications, which is cumbersome and cannot form a unified office ecosystem; existing technologies cannot organically integrate functions such as document signing, task management, meeting management, and AI intelligence to form an integrated office platform; (3) Low level of intelligence: Most office systems lack AI intelligent functions and cannot provide intelligent question answering, intelligent generation and other auxiliary office capabilities; even if some systems provide AI question answering functions, they only use simple semantic search and are not optimized for office scenarios, resulting in low search accuracy and failing to meet the needs of enterprise office work. (4) Inconvenient identity verification: When logging in to a mobile terminal, users need to repeatedly enter their account and password, making the verification process cumbersome. When performing sensitive operations (such as signing documents), existing technologies either require users to verify their identity every time they operate (leading to cumbersome operations) or do not verify for a long time (leading to security risks), lacking a verification mechanism that balances security and convenience. (5) Data synchronization delay: After cross-device operation, data synchronization is not timely, resulting in inconsistency between PC and mobile terminal data; existing technology lacks intelligent data conflict detection and resolution mechanism, and when PC and mobile terminal modify the same data item at the same time, data overwriting or loss problems are likely to occur. (6) Poor user experience: There is a lack of connection between functional modules, making it impossible to form a complete office loop; existing technologies cannot achieve synergy between functional modules, and cannot produce a technical effect of "1+1>2".

[0003] Currently, some solutions exist for specific needs, such as those supporting only document signing, task management only for mobile devices, or standalone AI question-answering tools. However, these solutions are often limited in functionality, failing to achieve efficient cross-device collaboration and lacking deep integration with the entire office workflow, thus hindering the construction of a complete integrated smart office ecosystem. Therefore, existing technologies have not yet been able to integrate comprehensive functions such as identity verification, document processing, task and meeting management, and AI assistance while achieving real-time cross-device collaboration, failing to meet the growing demand for intelligent and integrated office solutions from enterprises. Summary of the Invention

[0004] The purpose of this invention is to solve the problems in the prior art and propose an integrated intelligent office system and method that supports cross-device collaboration. It can realize real-time data collaboration between PC and mobile terminals and provide complete intelligent office functions such as document signing, task management, meeting management, problem feedback, AI Q&A, and PPT generation.

[0005] To achieve the above objectives, this invention proposes an integrated intelligent office system and method supporting cross-device collaboration, comprising the following steps: Face recognition verification and session establishment: User identity is verified via face recognition on a mobile terminal, and a user session is established upon successful verification; State maintenance: A session state object is created and maintained locally on the mobile terminal, recording the timestamp of the user's last successful face recognition verification in this session; Verification decision and execution: When an office operation requiring identity verification is triggered on the mobile terminal, the validity of the verification status is determined based on the timestamp; if invalid, face recognition verification is triggered, and the timestamp is updated upon successful verification; Operation execution and data synchronization: The office operation is executed, and the operation data is synchronized to the server; Cross-device collaboration: The PC accesses the server to obtain and display the operation data synchronized with the mobile terminal.

[0006] Preferably, the "determining whether the verification status is valid" step in the verification decision and execution process specifically includes: calculating the difference between the current time and the timestamp; determining whether the difference exceeds a preset threshold; wherein the preset threshold is adjustable through system configuration. This preferred solution allows system administrators to flexibly set the verification validity period according to the security level of different operations (such as signing important documents and approving routine tasks), achieving a dynamic balance between security and user operation convenience, and avoiding frequent verification or security risks caused by fixed thresholds.

[0007] Preferably, the office operation includes intelligent question-and-answer operation, which includes: receiving a question-and-answer request input by the user; obtaining an answer based on a selected knowledge base using retrieval enhancement generation technology, wherein the retrieval process includes: a first stage: retrieving candidate document fragments from the knowledge base based on semantic similarity; a second stage: filtering and reordering the candidate document fragments based on document metadata associated with the question-and-answer request and / or user identity; and generating an answer based on the filtered document fragments. This preferred solution, by introducing metadata filtering and reordering that combines user identity and question context, can significantly improve the document retrieval accuracy in specific office scenarios, reduce interference from irrelevant information, and thus provide more accurate and relevant intelligent question-and-answer results.

[0008] Preferably, the document metadata includes at least one of department, project number, and creator. This preferred solution can utilize document attributes naturally present in an office setting for precise filtering, such as prioritizing the return of documents from the user's department or documents matching a specific project number, making AI question answering more aligned with the company's actual business structure.

[0009] Preferably, the operation execution and synchronization steps, specifically synchronizing the operation data to the server, include: digitally signing the operation data; uploading the signed data to the server; having the server verify the signature and detect any concurrent modification conflicts for the same data item; and handling any detected conflicts according to a preset conflict resolution strategy. This preferred solution, by digitally signing data packets, ensures the integrity and non-repudiation of data during transmission; and through the conflict detection mechanism, it can promptly identify data inconsistencies that may arise from concurrent operations by multiple devices, thus guaranteeing data consistency.

[0010] Preferably, the conflict resolution strategy is one of the following: a last-write priority strategy based on the timestamps of each modification operation; or a strategy that provides conflict information to the user and receives the user's selection instruction. This preferred solution provides both automatic and manual resolution paths. It can efficiently and automatically resolve conflicts in most cases through the "last-write priority" strategy, and also allows the user to make a judgment in cases of critical data conflicts, thus balancing processing efficiency and operational accuracy.

[0011] Preferably, the office operations include at least one of the following: document signing based on handwritten signatures or template signatures, task approval, meeting scheduling, and PPT generation. This preferred solution integrates multiple core and intelligent office functions into one, allowing users to complete the entire workflow from document processing and task collaboration to intelligent assistance within a single platform without switching between different applications, greatly improving the integration and convenience of office work.

[0012] To achieve the above objectives, this invention also proposes an integrated intelligent office system that supports cross-device collaboration, comprising a mobile terminal, a server, and a PC: The mobile terminal is equipped with: an identity verification module for verifying user identity via facial recognition; a session-level facial verification status maintenance module, connected to the identity verification module, for creating and maintaining a session status object locally that records the timestamp of the last successful facial verification, and determining the verification status based on the timestamp when an office operation requiring verification is triggered, and triggering re-verification if invalid; an intelligent office function module for providing the operation interface and logic for document signing, task management, meeting management, AI Q&A, and PPT generation functions; and a first data synchronization module for synchronously uploading the operation data generated by the intelligent office function module to the server. The server-side deployment includes: a second data synchronization module for verifying data received from the mobile terminal, performing conflict detection and resolution, and storing data; a task management module for managing task processes; an identity authentication module for verifying facial features; an AI service module for providing intelligent question answering and PPT generation services; and a data storage module for storing user data and office data. The PC client is equipped with: a PC office module for accessing the server's data storage module to obtain and manipulate office data synchronized by the mobile terminal; and a third data synchronization module for synchronizing data between the PC client and the server. In this preferred embodiment, the first, second, and third data synchronization modules work collaboratively to achieve data sharing and synchronization between the mobile terminal and the PC. This optimized solution, by deploying collaborative data synchronization modules across the three terminals, constructs a unified and reliable data flow channel, ensuring real-time consistency of operational status between the mobile terminal and the PC. This forms the technological foundation for truly seamless cross-device collaborative work.

[0013] Preferably, the intelligent office function module includes an AI question-and-answer submodule; the AI ​​service module is used to perform the intelligent question-and-answer operation, and its retrieval process includes: a first stage, retrieving candidate document fragments based on semantic similarity; and a second stage, filtering and reordering the candidate document fragments based on document metadata associated with the question-and-answer request and / or user identity. This preferred solution deploys the optimized two-stage retrieval enhancement generation technology on the server side, providing a unified intelligent service for mobile terminals and PCs, ensuring the efficiency and accuracy of AI question-and-answer capabilities, and supporting cross-device sharing of dialogue history.

[0014] Preferably, the second data synchronization module on the server side implements a conflict resolution strategy that prioritizes the last write operation based on its timestamp, or a strategy that provides conflict information to the user and receives the user's selection instruction. This preferred solution clarifies the data conflict resolution logic at the system level, enabling the server to automatically or interactively handle concurrent modifications, fundamentally ensuring the consistency of core data in a collaborative work environment with multiple users and devices.

[0015] The beneficial effects of this invention are: 1. This invention achieves a dynamic balance between security and ease of operation through a session-level face verification state maintenance mechanism, avoiding the drawbacks of frequent verification or long periods without verification.

[0016] 2. This invention significantly improves the retrieval accuracy and answer relevance of intelligent question answering in office scenarios by employing a two-stage retrieval enhancement generation technology that combines metadata filtering.

[0017] 3. This invention ensures the integrity, non-repudiation, and eventual consistency of cross-device data synchronization through digital signatures and intelligent conflict resolution strategies.

[0018] 4. This invention deeply integrates functions such as document signing, task management, meeting management, AI Q&A, and PPT generation into a unified platform, and relies on the above mechanisms to achieve real-time cross-device collaboration, thus building an integrated intelligent office ecosystem and significantly improving office efficiency and user experience.

[0019] The features and advantages of the present invention will be described in detail through embodiments and in conjunction with the accompanying drawings. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the overall system architecture and workflow of an intelligent office integrated system and method that supports cross-device collaboration according to the present invention.

[0021] In the diagram: 1- Identity Verification Module: Deployed on both the mobile terminal and the server, responsible for identity verification via account password or facial recognition; 2- Task Receiving Module: Deployed on the mobile terminal, responsible for receiving notifications such as pending tasks and meeting reminders; 3- Smart Office Module: Deployed on the mobile terminal, responsible for providing core office functions such as document signing, task management, meeting management, and problem feedback; 4- AI Smart Module: Deployed on the mobile terminal, responsible for providing intelligent functions such as AI Q&A and PPT generation; 5- Data Synchronization Module: Deployed on both the mobile terminal (first data synchronization module) and the server (second data synchronization module), responsible for synchronizing the operation data of the mobile terminal. To the server side; 6-PC Office Module: Deployed on the PC side, responsible for providing office functions such as document signing, task management, and meeting management; 7-Task Management Module: Deployed on the server side, responsible for managing tasks, meetings, and signing processes; 8-Identity Authentication Module: Deployed on the server side, responsible for facial feature comparison and session management; 9-Document Processing Module: Deployed on the server side, responsible for PDF processing, signature merging, and electronic signatures; 10-AI Service Module: Deployed on the server side, responsible for knowledge base retrieval and PPT generation services; 11-Data Storage Module: Deployed on the server side, responsible for storing tasks, signing records, user data, and knowledge base data. Detailed Implementation

[0022] The following will be combined with the appendix Figure 1 The technical solutions in the embodiments of the present invention will be clearly and completely described.

[0023] The physical foundation of the system includes a server cluster interconnected via a network, storage devices, and user terminals. The server runs the program modules of this invention. The storage devices store user-registered facial feature data, task records, signed documents, meeting appointments, knowledge base documents, and other data. The user terminals include PCs (web browsers) and mobile terminals (Android applications) for identity verification, intelligent office operations, and data access.

[0024] Authentication Phase: Authentication module 1 includes a mobile terminal login submodule and a server-side authentication submodule. The mobile terminal provides two login methods: account / password login and face recognition login. Account / Password Login: The user enters their employee ID and password. The mobile terminal sends the login information to the server via HTTPS. The server verifies the correctness of the account and password. Upon successful verification, an access token (JWT format, containing user ID, validity period, etc.) is generated. The mobile terminal saves the access token for subsequent API requests. Face Recognition Login: The mobile terminal uses the device's front-facing camera and the FaceCapture component to capture facial images. During the capture process, the system performs real-time image preprocessing: First, it uses MTCNN or RetinaFace algorithms for face detection to ensure the face is fully visible in the image. Then, it performs image quality assessment, detecting image sharpness, lighting conditions, and face angle. When image quality is found to be unsatisfactory (e.g., blur > 0.3, insufficient lighting, face angle > 15 degrees), the user is prompted to adjust their position or lighting. Finally, it performs lighting and angle correction using histogram equalization or adaptive lighting correction algorithms to improve image quality. After acquiring a qualified facial image, the mobile terminal compresses the image into JPEG format (85% quality, maximum size 1920×1080) and uploads it to the server via HTTPS. The server-side authentication submodule (identity authentication module 8) receives the uploaded facial image and extracts 128-dimensional or 512-dimensional facial feature vectors using deep learning models such as FaceNet or ArcFace. The server queries the facial feature vectors of all registered users (obtained from data storage module 11 and stored using AES-256 encryption), performs a 1:N comparison between the extracted feature vectors and all registered features, calculates similarity using cosine similarity or Euclidean distance, and searches for a matching user among all users. When a user with a similarity greater than a preset threshold (0.7) is found, the verification is considered successful, the server generates an access token and returns user information, and the mobile terminal saves the access token and user information to complete the login. The server stores session information in a Redis in-memory database, sets the session validity period to 24 hours, and supports sharing sessions across devices.

[0025] Intelligent Office Function Phase: Intelligent Office Module 3 provides various mobile office functions. Document Signing Function: When a user opens the document signing interface, the system loads a list of documents to be signed from the server. After the user selects a file, the system loads the PDF file, supporting optimized delayed loading of multi-page PDFs (loading only the currently visible page and the three pages before and after it). The system provides multiple signing modes: Handwritten signature mode uses the Canvas component to draw the signature. When the user touches the screen, the system records the coordinate sequence of the touch point and uses a cubic Bézier curve algorithm to smooth the coordinate sequence, generating a smooth signature path. The system dynamically adjusts the line thickness according to the pen speed: calculating the distance and time difference between two adjacent touch points to obtain the pen speed (pixels / milliseconds). When the pen speed > 2 pixels / milliseconds, the line width is set to 4 pixels; when the pen speed < 0.1 pixels / milliseconds, the line width is set to 10 pixels; otherwise, linear interpolation is used. During signature drawing, the system automatically locks the screen to landscape mode to improve the signing experience. Template Signature Mode loads preset signature templates from the user's signature library, including checkmark templates, agreement templates ("Agree" text templates), and user-uploaded signature images. The signature location is determined using a PDF coordinate transformation algorithm: The original dimensions of the PDF file (e.g., 595×842 dots, corresponding to A4 paper) are obtained, along with the display size and scaling factor of the PDF page on the mobile terminal screen. A mapping relationship between the PDF coordinate system and the screen coordinate system is established, and the coordinates of the touch point on the screen are converted to precise coordinates in the PDF file. Before performing operations requiring authentication (such as placing a signature, completing signing, task approval, etc.), the system checks the session-level face verification status: The mobile terminal maintains a session state object (SessionState) locally, stored in local storage (LocalStorage), containing the following fields: last_face_verify_timestamp (timestamp of the last successful face verification, in Unix timestamp format), user_id (user ID), and session_id (session ID). The system calculates the difference between the current time (current_time) and the last_face_verify_timestamp (time_diff = current_time - last_face_verify_timestamp). A preset threshold is configurable: the system supports setting the validity period threshold for the verification status through a configuration file or management interface. The default value is 3600 seconds (i.e., 1 hour). Administrators can adjust this threshold according to enterprise security policies and business needs. For example, for high-security operations (such as signing important documents), a shorter time (e.g., 30 minutes, i.e., 1800 seconds) can be set; for routine operations (such as task approval), a longer time (e.g., 2 hours, i.e., 7200 seconds) can be set.The main technical consideration in setting this threshold is to balance security and convenience: a threshold that is too short will lead to frequent verification, affecting user experience and operational efficiency; a threshold that is too long will increase security risks and may lead to identity theft. Through a configurable threshold mechanism, the system can flexibly adjust security policies in different scenarios to achieve a dynamic balance between security and convenience. If `time_diff` exceeds the preset threshold, the verification status is deemed insufficient, triggering the face recognition verification process. After successful verification, the system updates `last_face_verify_timestamp` in the session state object to the current timestamp and allows further operation. The session state object is initialized when the user logs in and cleared when the user logs out. This session-level face verification state maintenance mechanism works in conjunction with the cross-device data synchronization mechanism: after the user completes verification on the mobile terminal, the PC can continue operating based on the session state without repeated verification, improving the smoothness of cross-device collaboration; at the same time, this mechanism ensures operational security while avoiding the cumbersome operation caused by frequent verification, achieving a balance between security and convenience. Meeting management function: When the user opens the meeting management interface, the system loads a list of meeting rooms and available time slots from the server. The meeting room list displays available meeting room information, including capacity, equipment, location, and available time slots, and supports filtering by date. The quick booking function allows users to quickly create meeting bookings by clicking on a time slot, supporting adding participants (searching for users), setting the meeting type (internal meeting, external meeting, etc.), and entering a meeting description. After adding participants, the system automatically sends a meeting invitation notification to all participants, including meeting time, location, and topic information. The "My Bookings" function displays a personal meeting booking list, supporting cancellation, modification of bookings, and viewing booking details. The meeting reminder function sends push notifications to remind users of upcoming meetings, supporting sound and vibration alerts, with reminder times set to 15 minutes, 30 minutes, or 1 hour before the meeting. Participants who receive the meeting invitation notification will also receive a reminder notification before the meeting starts, ensuring they are informed of meeting information in a timely manner. The task management function: When users open the task management interface, the system loads a task list from the server. The task list supports filtering by status (pending submission, pending approval, completed), by type (single person, multiple people, department, global), and keyword search, and supports pull-down refresh and pull-up to load more. The task details section displays complete task information, including task title, description, deadline, creator, and approver. It supports tasks such as submitting, approving, forwarding, and adding comments. The task dashboard provides statistics on pending tasks, tasks awaiting approval, completed tasks, and the total number of tasks, visually displaying work progress in card format, including completion rate calculations. The task calendar displays tasks in a calendar view, allowing users to view task assignments by date. Tasks in different statuses are identified by different colors.Issue Feedback Function: The mobile terminal also provides an issue feedback function, including issue creation (supports text description, image upload, and selection of issue type), issue forwarding (forwarding issues to other users or departments, supporting the addition of forwarding instructions), issue receipt (confirming the receipt of processed issues and marking the issue processing status), and issue tracking (viewing the issue processing progress and receiving issue processing notifications).

[0026] AI Intelligent Function Phase: AI Intelligent Module 4 provides AI Q&A and PPT generation functions. AI Q&A Function: When a user opens the AI ​​Q&A interface, the system loads a list of available knowledge bases from the server. The user can select different knowledge bases from the list (such as "Project Program Knowledge Base", "Safety Specification Knowledge Base", etc.), and each knowledge base corresponds to a different set of documents. After the user inputs a question, the system sends it to the AI ​​service module 10. The server uses RAG (Retrieval Augmentation) technology for intelligent question answering, which is optimized for office scenarios. First, it retrieves relevant document fragments from a selected knowledge base. The retrieval process is divided into two stages: The first stage uses vector similarity search (e.g., using Faiss or Milvus vector databases) to calculate the cosine similarity between the semantic vectors of the document content and the semantic vectors of the question, finding the top K document fragments with the highest semantic similarity (K is usually 10-20). The second stage performs metadata filtering, further filtering based on the document's metadata information (including department, project ID, creator, document type, and creation time). For example, if the user belongs to the "Engineering Department," documents with that department are returned first; if the question contains a project ID, documents matching the project ID are returned first; if the user is the creator of the document, the document's priority is increased. Metadata filtering significantly improves retrieval accuracy and reduces interference from irrelevant documents. The retrieved document fragments are then input into a large language model (such as GPT, Claude, etc.) along with the user's question to generate an answer; finally, the answer and a list of references are returned. After receiving the answer, the mobile terminal uses a rich-text component to render Markdown format content, supporting rich text display such as code blocks, lists, images, and tables. The system saves the dialogue history and supports viewing historical question-and-answer records. The dialogue history is stored in data storage module 11, and PCs and mobile terminals can share and view the dialogue history. This metadata filtering RAG technology works synergistically with the dialogue history sharing function: the user's question-and-answer records on any device can provide context for subsequent questions and answers, further improving the intelligence level of AI question answering; at the same time, this technology works synergistically with the cross-device data synchronization mechanism, enabling PCs and mobile terminals to share the dialogue history, realizing true cross-device intelligent collaboration. PPT generation function: The user opens the PPT generation interface, inputs PPT requirements (such as "Give me a PPT outline for personnel qualification management"), and the system sends the requirements to AI service module 10. The server calls the PPT generation service to generate a PPT outline through dialogue. The system supports multi-round dialogue to modify and improve the outline. Users can provide modification suggestions for the generated outline (such as "add a security management chapter"), and the system will adjust the outline based on the modification suggestions.The mobile device provides a real-time preview of the generated PPT outline, displaying its status (generating, generated, modified). The outline is presented in a tree structure, including hierarchical relationships such as titles, chapters, and sub-chaps. Once the user confirms the outline, clicking the "Generate PPT" button sends the final outline to the server. The server then uses a PPT generation engine (such as python-pptx or a template-based tool). This engine has a built-in enterprise template style library with various preset templates: title page templates (including company logo, title font, and color scheme), content page templates (including title styles, body text styles, list styles, and chart styles), and ending page templates (including standard closings such as "Thank you for watching"). The PPT generation engine automatically matches the corresponding style template based on the outline's theme: for example, if the outline theme contains keywords such as "safety" or "standards," it matches the "safety theme" template (blue color scheme, bold font); if the outline theme contains keywords such as "project" or "progress," it matches the "project theme" template (orange color scheme, Microsoft YaHei font). The generation engine also automatically adjusts the font size and color of headings based on chapter levels to ensure the generated PPT conforms to corporate visual guidelines. Once generated, the server returns a download link for the PPT file, which can be downloaded on mobile devices. The PPT file is stored in data storage module 11 and can be accessed and downloaded by both PCs and mobile devices.

[0027] Data synchronization phase: Data synchronization module 5 includes a mobile terminal upload submodule and a server-side storage submodule. The mobile terminal upload submodule packages operation data (signature data, meeting reservation data, task approval data, AI dialogue data, PPT generation data, etc.), file information, timestamps, and user identifiers into a data packet in JSON format. The system digitally signs the data packet to ensure data integrity and non-repudiation: the system uses the RSA-SHA256 algorithm for digital signature. Specifically, the mobile terminal uses the user's private key to RSA-SHA256 sign the data packet, generating a digital signature. The key management mechanism adopts hierarchical key management: the user's private key is stored in the secure storage area of ​​the mobile terminal (such as Android's KeyStore), protected by hardware security module (HSM) or software encryption. The private key is not directly transmitted and is only used for local signing. The server stores the user's public key for verifying digital signatures. Key generation and distribution... The process is completed during user registration. The server generates an RSA key pair (2048 or 4096 bits in length). The private key is encrypted using AES-256 and transmitted to the mobile terminal via HTTPS. The mobile terminal decrypts the private key and stores it in a secure storage area, while the public key is stored in the server's database. The system supports a key rotation mechanism, periodically (e.g., every 6 months) updating the key pair to improve security. During key rotation, the system generates a new key pair, updates the server with the new public key, encrypts the new private key, and transmits it to the mobile terminal. The old key pair is retained for a period (e.g., 30 days) to verify the signature of old data packets, after which it is automatically deleted. The mobile terminal uploads the data packet and digital signature to the server via HTTPS. The server verifies the digital signature: using the stored user public key, the server performs RSA-SHA256 verification on the digital signature to verify the integrity and authenticity of the data packet. If verification fails, the data packet is rejected and an error message is returned. After verifying the digital signature on the server side, a data conflict detection is performed: The server maintains a data modification log table (DataModificationLog) to record detailed information for each data modification, including the data item ID (data_id), modification timestamp (modify_timestamp), modifying user ID (user_id), modification device type (device_type, PC or Mobile), and modification content (modify_content). When the server receives a new data modification request, it queries the DataModificationLog table to check if any other device has modified the same data item within the same time window (e.g., within 5 seconds).If a conflict is detected, the server adopts one of the following conflict resolution strategies: Strategy 1: A timestamp-based "last write first" strategy. It compares the timestamps (modify_timestamp) of the two modification operations, retains the modification with the newer timestamp, marks the modification with the older timestamp as "overwritten," and records the conflict information. Strategy 2: A user-manual resolution strategy. When a conflict is detected, the server does not immediately save the new data. Instead, it returns the conflict information (including the modification content of the two versions, the modification time, and the user who made the modification) to the client. The client displays a conflict prompt interface, requiring the user to manually choose which version to keep or to perform a merge operation. After the user makes a choice, the client sends the result back to the server, and the server saves the data according to the user's choice. The conflict resolution strategy can be selected through system configuration; Strategy 1 is used by default. After the conflict is resolved, the server stores the data in the database (data storage module 11) and updates the data version number (using a timestamp or an incrementing sequence number). The server-side storage submodule stores data by type: task data is stored in the task table, signature records are stored in the signature table, meeting appointments are stored in the meeting table, AI dialogue records are stored in the dialogue table, and PPT files are stored in the file system. The system sets a version number for each data record, and the version number automatically increments when the data is updated.

[0028] Cross-device data sharing phase: The PC-based office module 6 and the mobile terminal share data by accessing the server. Both the PC and mobile terminal obtain data updates via polling or WebSocket: the PC sends a data version query request to the server every 5 seconds, and the mobile terminal sends one every 3 seconds. The server returns the current data version number. When a change in the data version number is detected, the client automatically refreshes the display and retrieves the latest data from the server. For scenarios with high real-time requirements (such as task status changes, meeting appointment confirmations, etc.), the server proactively pushes data update notifications via WebSocket, and the client immediately refreshes the data upon receiving the notification. Data shared between the PC and mobile terminal includes: task lists and details, meeting appointment information, signature records, AI dialogue history, PPT files, etc. After a user completes an operation on the mobile terminal, the data is immediately synchronized to the server. When a PC user opens the corresponding function, they can see the operation results from the mobile terminal in real time, enabling cross-device collaborative work.

[0029] Example 1: An employee opens the document signing function on a mobile device, views the list of documents to be signed, and selects a procurement contract to sign. The system loads the PDF file, the employee selects the handwritten signature mode, the system automatically locks the screen in landscape mode, the employee clicks on the specified location on the contract, and the system checks the session-level face verification status: it reads the session state object in local storage, obtains the last_face_verify_timestamp field, calculates the difference between the current time and the timestamp, and finds that the difference exceeds the preset threshold (1 hour), determining that the verification status is insufficient and triggering face recognition verification. The employee faces the camera as prompted, the system captures the face image and uploads it to the server, the server performs a 1:1 comparison verification, and after successful verification, updates the timestamp in the session state object and allows subsequent signing operations to be completed. The system uploads the signing result to the server, the server accurately places the signature image into the specified location in the PDF file according to the coordinate information, generates an electronic seal, and completes the signing process. Subsequently, the employee opens the document signing function on the PC and can immediately see the signed contract status, realizing cross-device data synchronization. This embodiment demonstrates the advantages of a session-level face verification state maintenance mechanism: it ensures operational security (triggers re-verification when the verification state expires) while avoiding the cumbersome operation caused by frequent verification (no need for repeated verification within the verification validity period), achieving a balance between security and convenience. Simultaneously, this mechanism works in conjunction with a cross-device data synchronization mechanism, allowing users to continue operations on the PC based on the session state after completing verification on the mobile terminal, without needing to repeat verification, thus improving the smoothness of cross-device collaboration.

[0030] Example 2: A project manager opens the task management function on a mobile device, views the list of tasks awaiting approval, and finds an urgent task requiring approval. The project manager clicks on the task details, views the task information, selects "Approved," adds approval comments, and the system checks the session-level face verification status. Once verification is successful, the approval operation is allowed, and the system uploads the approval result to the server. The server performs a data conflict detection, finds no conflict, updates the task status, and triggers a data version number update. The project manager then opens the task management function on a PC. The system detects the data version number change, automatically refreshes the display, and the project manager can immediately see that the task status has been updated to "Approved." Simultaneously, the project manager opens the AI ​​question-and-answer function, selects "Safety Standards Knowledge Base," and enters the question "What are the safety protection measures for construction sites?" The system uses RAG technology to search relevant knowledge base documents: first, it searches for relevant document fragments based on semantic similarity, then filters them based on document metadata. Since the project manager belongs to the "Engineering Department," the system prioritizes returning documents from departments listed as "Engineering Department," improving search accuracy, generating detailed answers, and displaying reference materials. The dialogue history is saved to the server. When the project manager opens the AI ​​Q&A function on the PC, they can view the complete dialogue history and continue asking questions based on the previous dialogue context, further improving the intelligence level of AI Q&A. This embodiment demonstrates the synergistic effect of technical means: the session-level face verification state maintenance mechanism ensures operational security; the intelligent conflict resolution strategy ensures data consistency across devices; and the RAG technology for metadata filtering, in conjunction with the dialogue history sharing function, significantly improves the retrieval accuracy and intelligence level of AI Q&A. The combination of these technical means produces technical effects that cannot be achieved by a single technical means, i.e., a synergistic effect of "1+1>2".

[0031] Example 3: A department manager opens the meeting management function on a mobile device, views the list of available meeting rooms, selects a meeting room that can accommodate 20 people, and clicks the time slot "Tomorrow morning 9:00-11:00" to quickly create a meeting reservation. The system automatically fills in the meeting information. The department manager adds participants (selecting department members), sets the meeting type to "Internal Meeting," and enters the meeting topic "Project Review Meeting." After confirmation, the system uploads the meeting reservation data to the server. The server performs a data conflict check, finds no conflicts, saves the reservation information, and sends a meeting notification to all participants. The department manager then opens the meeting management function on a PC. The system detects the data version number change, automatically refreshes the display, and immediately displays the newly created meeting reservation, supporting viewing, modification, or cancellation. Simultaneously, the department manager opens the PPT generation function on the mobile device, inputting the requirement "Generate a PPT outline for a project progress report, including four parts: project overview, current progress, existing problems, and next steps." The system generates the PPT outline through a dialog. After reviewing the outline, the department manager suggests modifications, such as "Add risk analysis to the 'existing problems' section." The system adjusts the outline based on these suggestions. After confirming the outline, the department manager clicks the "Generate PPT" button. The system sends the outline to the server, which then calls the PPT generation engine. Since the outline theme contains keywords such as "project" and "progress," the engine automatically matches a "project theme" template (orange color scheme, Microsoft YaHei font), automatically adjusts the title font size and color according to the chapter hierarchy, generates a PPT file that conforms to the company's visual specifications, and saves it to the server. The department manager can then open the PPT generation function on their PC to download the generated PPT file for use in meetings. This example demonstrates the synergistic effect of these technologies: the intelligent conflict resolution strategy ensures cross-device data consistency, enabling real-time synchronization of meeting appointment data; the cross-device data sharing mechanism allows PPT files to be accessed on both PCs and mobile devices, achieving true cross-device collaboration; and the automatic matching function of the enterprise template style library improves the quality of the generated PPT. This combination of technologies not only achieves functional integration but also technological synergy, producing technical effects that cannot be achieved by a single technology.

[0032] The above embodiments are illustrative of the present invention and are not intended to limit the present invention. Any simple modifications to the present invention are within the scope of protection of the present invention.

Claims

1. A method for integrated smart office work supporting cross-device collaboration, characterized in that, Includes the following steps: Facial recognition verification and session establishment steps: Verify the user's identity via facial recognition on the mobile terminal, and establish a user session after successful verification; State maintenance steps: Create and maintain a session state object locally on the mobile terminal. The session state object records the timestamp of the last time the user successfully passed face recognition verification in this session. Verification decision and execution steps: When an office operation requiring identity verification is triggered on the mobile terminal, the validity of the verification status is determined based on the timestamp; if invalid, facial recognition verification is triggered and the timestamp is updated after successful verification. Operation execution and data synchronization steps: Perform the office operation and synchronize the operation data to the server; Cross-device collaboration steps: The PC accesses the server to obtain and display the operation data synchronized with the mobile terminal.

2. The method according to claim 1, characterized in that, The "determining whether the verification status is valid" step in the verification decision and execution process specifically includes: Calculate the difference between the current time and the timestamp; Determine whether the difference exceeds a preset threshold; The preset threshold is adjustable through system configuration.

3. The method according to claim 1, characterized in that, The office operations include intelligent question-and-answer operations, which include: Receive user-inputted question-and-answer requests; Based on the selected knowledge base, retrieval enhancement generation technology is used to obtain the answer. The retrieval process includes: Phase 1: Retrieve candidate document fragments from the knowledge base based on semantic similarity; The second stage involves filtering and reordering the candidate document fragments based on document metadata associated with the question-and-answer request and / or user identity. Answers are generated based on the filtered document fragments.

4. The method according to claim 3, characterized in that, The document metadata includes at least one of the following: department, project number, and creator.

5. The method according to claim 1, characterized in that, The operation execution and synchronization steps specifically include synchronizing the operation data to the server: The operation data is digitally signed; Upload the signed data to the server. The server verifies the signature and checks for concurrent modification conflicts targeting the same data item. If a conflict is detected, it will be handled according to the preset conflict resolution strategy.

6. The method according to claim 5, characterized in that, The conflict resolution strategy is any one of the following: a last-write priority strategy based on the timestamps of each modification operation; or a strategy that provides conflict information to the user and receives the user's selection instruction.

7. The method according to claim 1, characterized in that, The office operations include at least one of the following: document signing based on handwritten signature or template signature, task approval, meeting reservation, and PPT generation.

8. A smart office integrated system supporting cross-device collaboration, characterized in that, Including mobile terminals, server-side, and PC-side: The mobile terminal is deployed with: The identity verification module is used to verify the user's identity through facial recognition; The session-level face verification status maintenance module is connected to the identity verification module. It is used to create and maintain a session status object that records the timestamp of the last successful face verification locally. When an office operation that requires verification is triggered, the verification status is determined according to the timestamp. If it is invalid, a re-verification is triggered. The intelligent office function module provides the user interface and logic for document signing, task management, meeting management, AI Q&A, and PPT generation. The first data synchronization module is used to synchronize and upload the operation data generated by the intelligent office function module to the server. The server-side deployment includes: The second data synchronization module is used to verify the data received from the mobile terminal, perform conflict detection and resolution, and store the data. The task management module is used to manage task processes; The identity authentication module is used to verify facial features; The AI ​​service module is used to provide intelligent question answering and PPT generation services; The data storage module is used to store user data and office data; The PC terminal is equipped with: The PC-side office module is used to access the data storage module on the server side to obtain and operate the office data synchronized by the mobile terminal. The third data synchronization module is used to realize data synchronization between the PC and the server. The first data synchronization module, the second data synchronization module, and the third data synchronization module work together to achieve data sharing and synchronization between the mobile terminal and the PC.

9. The system according to claim 8, characterized in that, The intelligent office function module includes an AI question-and-answer sub-module; The AI ​​service module is used to perform the intelligent question-answering operation. Its retrieval process includes: a first stage, retrieving candidate document fragments based on semantic similarity; and a second stage, filtering and reordering the candidate document fragments based on document metadata associated with the question-answering request and / or user identity.

10. The system according to claim 8 or 9, characterized in that, The second data synchronization module on the server side implements the following conflict resolution strategies: either a last-write priority strategy based on the timestamp of each modification operation, or a strategy that provides conflict information to the user and receives the user's selection instruction.

Images