A SMART SYSTEM FOR MESSAGING-BASED WORK GROUPS

Abstract

This invention relates to a system (1) that enables contextual task extraction from the chat content of users in messaging-based work groups, assigning these tasks to relevant people, setting deadlines, integrating into the calendar and sending reminders via push notification.

Description

1 TARIFF A SMART APPROACH TO MESSAGING-BASED WORK GROUPS. SYSTEM Technical Area This invention allows users to extract information from chat content in messaging-based work groups. contextual task inference and assigning these tasks to the relevant individuals. assignment, deadline setting, calendar integration and push notifications It relates to a system that provides reminders via notification. Previous Technique In the known state of the art, specific keywords are identified and extracted from the message. Team messaging and task 15 that enables the suggestion of creating tasks. enabling the integration of tracking and manual assignment of tasks from messages. enabling the creation of AI-powered automation workflows into chat interfaces integration, thereby simplifying NLP-based intent detection. Various solutions are encountered that enable the triggering of actions. In current techniques, the NLP model is run on a keyword-based approach, focusing on in-depth learning. Semantic analysis (intent + entity recognition) is not being performed; the same name User disambiguation is not present, calendar synchronization, file content matching, and AI-powered task prioritization. Modules such as (AI Scoring Engine) are not defined, 25 during task assignment. Interactive user approval process (UI / UX-based task validation) Not offered. 2 Also, in the known state of the technology, a message sent to a group chat. performing an intent analysis on the content, thereby determining the purpose of the message. determining whether or not it contains a task in the message content in the analysis in question If a highly probable structure is detected, confirmation will be sent to the user / person who sent the message. a window is sent asking "Should this message be saved as a task?" enabling the sending of messages to multiple users with the same name in the messaging group. If a user is found, their IDs are analyzed to identify the correct person. a solution that ensures the task is completed and delivered to the correct user It has not been encountered. United States Law No. US2024144197, regarding the known state of the art. Task-based user scheduling / scheduling in the patent document an invention where changes are certified in a software user's virtual calendar It is explained that the software is automated against the user's virtual calendar. Calendar rules specific to the user for approval, meetings related to the user, and 15 It is generated by modeling trends and priority data for tasks. The software user obtains a series of pending tasks. Then, the virtual Pending tasks are managed according to the calendar by applying user-specific calendar rules. The priority level of an existing task in the cluster is determined by the schedule in the virtual calendar. It is determined that the priority level is higher than that of an existing meeting. 20 Based on this determination, an adjustment will be made to the virtual calendar for the current meeting. The existing task is being removed and a new activity corresponding to the current task is being added. The priority level of the processing tool included in the invention is the task set with pending tasks. determining the order of priority and prioritizing meetings on the virtual calendar It determines the priority levels. The priority level for a task or a meeting is 25. or more user input, batch sorting information, metadata factors, based on the prioritization of relevant tasks and meetings This is determined by manual user input, for example, software user or other. a software user (for example, a software user team within the same organization) 30 indicating the importance of a task or meeting by the leader or another superior 3 It can represent an expression. For example, in a meeting or task tag (e.g., words like "important", "urgent", "help", "immediately" used in the title or description Permutations of similar words, as manual user input. can be processed However, in the document numbered US2024144197, if there is more than one identical item in the group... If a user named [username] exists, the Entity Disambiguation module can be used to identify the users. By evaluating their IDs, the correct person can be identified and assigned to the relevant task. from assigning the task to the correct user, a task-related file chat from performing semantic similarity operations when shared in the environment, 10 In this way, if the similarity ratio is above the determined threshold value (e.g., 0.85), There is no mention of the file being automatically associated with the relevant task. Brief Description of the Invention This invention focuses on the intent behind the content of messages sent in a group chat. performing an analysis, thereby determining whether the message contains a task or not. the identification of structure 20, which has a high probability of being a task in the message content, in the analysis in question. If detected, a confirmation window will be sent to the user / person who sent the message. Sending a message that says, "Should this message be saved as a task?" this is provided if there are multiple users with the same name in the messaging group. By analyzing user IDs, the correct person is identified and the relevant a task-related file chat 25 that ensures the task is delivered to the correct user when shared in the environment, semantic similarity operations are performed, this In this way, if the similarity ratio is above the determined threshold value (e.g., 0.85), messaging that enables the file to be automatically associated with the relevant task. It implements an intelligent system related to workgroup-based structures. 4 Detailed Description of the Invention The purpose of this invention is to achieve "Messaging-Based Business". "An Intelligent System Related to Groups" is shown in the figure, and this figure is 5. Figure 1: The subject of the invention is an intelligent system related to messaging-based work groups. This is a schematic view. The parts shown in the figure are individually numbered, and the corresponding numbers represent 10. It is given below. 1. System 2. Electronic device 3. Calendar app 15 4. Chat application 5. Server In messaging-based work groups, users can take contextual information from chat content. the task of identifying and assigning these tasks to the relevant individuals, and the deadline specifying, integrating into the calendar, and sending reminders via push notifications. The invention enables the creation of smart messaging-based work groups. system (1); - enables the user to communicate with other people and connect with global networks. 25 at least one electronic device configured to provide (2), - at least one calendar application (3) run on the electronic device (2) enabling communication, allowing the user to send instant messages to multiple people. at least one chat application configured to provide (4) and - connecting with the chat application (4), calendar application (3), inside the most at least a data preprocessing service, an embedding service, a buffer layer, AI scoring engine and learning cycle (feedback pipeline) module / service to run, preprocess the message content in the chat application (4) to ensure that the text in the message content is cleaned and normalized. 5 to ensure that, then the text content is analyzed with at least a intent detection model. by processing the message content, it is determined whether or not there is a task involved. to ensure that the user of the electronic device (2) in question has a message If a structure with a high probability of being a task is detected, a confirmation window (4) will be opened in the chat application. This message is forwarded to the relevant user with the question, "Should this message be saved as a task?" to ensure the question is asked, the user from the chat application (4) to himself If the user answers yes to the question, it will be processed using the NER (Named Entity Identification) model. To ensure that task information is extracted from the relevant message content, the user If there are multiple people with the same name in the chat application (4), the entity Thanks to the Entity Disambiguation module, the task content is 15 in the chat application in which the user of the electronic device (2) is also included (4) to ensure that the task created is forwarded to the correct person in the group chat. (3) in the calendar application (2) on the user’s electronic device as an event To ensure the addition of the handover date information when the handover date is identified in the task, the handover date To ensure that the user is warned about this via the chat application (4), 20 A file related to the task identified on the chat application (4) was shared. time, retrieving file metadata, converting file content to vectors To enable the calculation of similarity scores using cosine similarity management, If the similarity score is higher than a certain threshold, the file will be... To enable task association, so that tasks are automatically assigned to chat 25 the assignment of the persons in the application (4) and the calendar of duties to ensure that their processing is carried out in their applications (3) It includes at least one configured server (5). 6 The equipment located on the electronic device (2) in the system that is the subject of the invention (1) to enable the user to run the chat application (4), the user's other The mobile phone is configured to allow group chats with registered users. It is a smart device in the form of a phone, tablet, or computer. Electronic device (2) Communication between the hardware on it and the user’s chat application (4) 5 to enable it to run at least one calendar application (3) that can install, calendar application (3) to enable the user to track time through the interface It is being structured. Communication between the server (5) and the chat application (4) in the system (1) is the subject of the invention 10 to set up, task, relevant chat messages in the group chat the user is in the logs are cleaned, tokenized and processed using a data preprocessing service, and It is configured to ensure normalization. Server (5) data pre- The task, message, and text included in the chat messages normalized in the processing service. High-dimensional vector representations of file contents using BERT or SBERT 15 It is structured to enable generation using algorithms based on algorithms. By communicating with the buffer layer inside the server (5) via the embedding service thanks to this buffer layer, real-time task queries The embedding results that enable acceleration include 20 in temporary memory. It is structured to ensure that it is retained. By communicating with the server (5) through the chat application (4), the user’s group The content of the messages in the conversation can be analyzed using NLP-based intent detection models and NER. By enabling the processing of the conversation in real time with its models, the conversation in question 25 to enable the extraction of mission information from the content of the messages is being configured. Server (5) chat application (4) user messages The content is analyzed using NLP-based intent detection models and NER models. by ensuring timely processing, if the text task statement “prepare”, If it includes information such as "complete," "do," or "finish," then the task intent is determined. 30 7 when a task intent is detected, the user is asked, "Is this a task?" a confirmation window is shown in the chat application (4) to provide, the user is shown the approval from the chat application (4) When the user answers "Yes" to the window, the system switches to task mode and displays the message. 5 users in the mention field on the interface in the chat application (4) It is configured to enable display. Server (5) chat NLP-based intent detection of user message content in (4) application By enabling real-time processing with models and NER models, The message marked as a task contains the person (entity), action, and datetime. It is structured to enable the extraction of information. 10 In one arrangement of the invention, the server (5) sends the chat application (4) to the user The message sent read, "Elif, complete the budget table by Friday." content analysis using NLP-based intent detection models and NER models The assigned person is: Elif, task: complete the budget table, deadline: 15 to ensure that information in the form of "Friday" is extracted is being configured. Server (5) is present on chat application (4) The process is carried out using the Entity Disambiguation module. to provide, so that there is more than one group in the relevant chat application (4) If a user with the same name already exists, the mention-based user ID will be 20. Verification is required; the NLP engine detects "Elif" as a name within the text. When this happens, if there is more than one "Elif" in the group at that moment, this situation is called an "ambiguous entity". to mark as mentioned, then mention the user via the chat application (4) the presentation of a selection window, the user being presented with the options 25 to prevent the creation of the task without selecting someone. is being configured. Server (5) is present on chat application (4) The process is carried out using the Entity Disambiguation module. to ensure that the IDs of the (4) users in the chat application during this process using this method to identify the right person and assign the task to the identified correct person. Enabling assignment to a user or person, with its built-in push notification service, 30 8 To ensure that instant notifications are sent to the correct assigned person. It is being structured. Server (5) task related to the user in the chat application (4) When created, the deadline information for the relevant task is also determined. 5 to ensure that the aforementioned deadline information is automatically generated (2) to ensure that the calendar application (3) is added to the electronic device, the word The subject of the addition process is at least one of the electronic devices (2) that are operated within it. The server is configured to allow this to be done via the calendar API. (5) by communicating with the electronic device (2) user (3) through the calendar application 10 (3) “Business Group Chat – [Task Name]” in the calendar application for each relevant task. The event description should include the title, task details, and chat link. It is configured to ensure that the server (5) calendar is found. by communicating with the application (3) when the time for the task approaches The TaskReminderScheduler module should be activated, allowing you to extend the deadline by 1-15 days. days before the deadline, 1 hour before the deadline, notification according to the time remaining until the deadline. To ensure planning, the chat application (4) interface “Reminder: Budget To ensure that a notification is displayed stating, "There is 1 hour left until the painting submission." It is being structured. Server (5) connects with the chat application (4), a task-related file When shared in the chat environment (3) of the chat application, both the message Ensuring that both the content and the file metadata are retrieved, Embedding on it Its service enables the conversion of text and file contents into vectors, and the converted files... storing the vectors in a distributed vector database and Cosine 25 Calculating the similarity score of the task using the similarity method. It is structured accordingly. The server (5) similarity ratio is determined by the threshold value. If it is above (e.g., 0.85), the file is automatically assigned to the relevant task. associating the task (4) in the chat application with the user 9 to enable the system to automatically display task-related documents in the details screen It is being structured. The server (5) communicates with the chat application (4) and the calendar application (3), This way, users get 5 when they complete tasks or change deadlines. or the learning cycle of those behaviors when responding to notifications (feedback pipeline) ensures that it is collected in the learning cycle. AI scoring engine on the relevant data of the (feedback pipeline) module to enable the transfer of information, thereby allowing the processing of task inference from similar messages. making it more accurate, optimizing reminder times on a user-by-user basis. (e.g., earlier notifications for users who frequently procrastinate), task to enable the improvement of prioritization estimates It is being structured. Server (5) selects tasks, files or 15 from millions of messages in corporate environments When decision moments need to be identified, the history in the chat application (4) processing of chat messages through a data preprocessing service to provide, the vector is sent to the embedding service on the processed data. conversion to a representation and similarity analysis via a distributed vector database. to enable the search to be performed, then the search results are scored by an AI scoring engine 20 contextual scoring by, prioritizing the most relevant content its ordering, with the buffer layer on top, in the frequently used chat application (4) storing results for queries and providing instant access to the user. It is structured to ensure that it is provided. By connecting with the server (5) chat application (4) and calendar application (3) The user says "I have completed the budget table" or within the chat application (4). When they send messages like, "Let's move the delivery date to Monday," their intentions... to ensure that it is processed by the detection module, then to find the relevant task ID, Update the task management service within the relevant task and these 30 Changes are made through the Synchronization Layer in both the chat application (4) and to ensure that the task status is reflected in the calendar application (3), Information such as file link and date details are shared among all users. to remain consistent, and these task updates also contribute to the learning cycle. The behavioral learning of the model is completed by transferring it to the (feedback pipeline) module. It is structured to ensure its sustainability. Industrial application of the invention In the system that is the subject of the invention (1) the user is a chat with many people It is located in a group chat in the application (4). Also, the user Calendar application (3) is also run on the electronic device (2). Server (5) messages sent to the chat application (4) with AI-supported modules inside analyzing the content, the job descriptions found in the message content, delivery 15 to determine their times, the appropriate person is identified on the chat application (4) screen It sends the assigned task description as a message notification. Then the user He sends the content of a message about the task given to him to the chat application (4) sending, new message content sent to the chat application (4) server (5) The message content was analyzed by [the relevant authority] and an update regarding the task was provided on 20th. It is determined whether or not it contains any information. If an update related to the task is detected... server (5) information about the task in the user's calendar application (3) It updates automatically. The content of the message sent to a group chat with the system (1) that is the subject of the invention Performing an intent analysis on the message, thereby determining whether the message contains a mission statement. Determining that it does not contain a function probability in the message content in the analysis in question. If a tall building is detected, a confirmation window will appear for the user / person who sent the message. a message is sent asking "Should this message be saved as a task?" 11 sent if there are multiple users with the same name in the messaging group By analyzing user IDs, the correct person is identified and the relevant The task is delivered to the correct user, a file related to the task is shared in a chat environment. performing semantic similarity operations when shared, thus If the similarity ratio is above the specified threshold (e.g., 0.85), the file will be deleted. It is automatically associated with the relevant task. The system in question (1) is within the scope of the Personal Data Protection Law (KVKK). It operates. The invention concerns a smart system (1) for messaging-based work groups. It is possible to develop various applications of the invention described here. It cannot be limited to examples; it is essentially as stated in the requests.

Claims

12 REQUESTS 1. In messaging-based work groups, contextual information can be extracted from users' chat content. the task of identifying and assigning these tasks to the relevant individuals, and the deadline 5. (This section appears to be a list of features, a calendar integration, and push notification reminders.) enabling it to be done; - enables the user to communicate with other people and connect with global networks. containing at least one electronic device (2) configured to provide, - at least one calendar application (3) run on the electronic device (2) 10 that enables communication, allowing the user to send instant messages to multiple people. at least one chat application configured to provide (4) and - connecting with the chat application (4), calendar application (3), inside the most at least a data preprocessing service, an embedding service, a buffer layer, AI scoring engine and learning cycle (feedback pipeline) module / service to run, preprocess the message content in the chat application (4) 15 to ensure that the text in the message content is cleaned and normalized. to ensure that, then the text content is analyzed with at least a intent detection model. by processing the message content, it is determined whether or not there is a task involved. to ensure that the user of the electronic device (2) in question has a message If a structure with a high probability of being a task is detected, a confirmation window (4) will appear in the chat application. This message is forwarded to the relevant user with the question, "Should this message be saved as a task?" to ensure the question is asked, the user from the chat application (4) to himself If the user answers yes to the question, it will be processed using the NER (Named Entity Identification) model. To ensure that task information is extracted from the relevant message content, the user If there is more than one person with the same name in the chat application (4), entity 25 The Entity Disambiguation module helps to clarify the task content. in the chat application in which the user of the electronic device (2) is also included (4) to ensure that the task created is forwarded to the correct person in the group chat. (3) in the calendar application (2) on the user’s electronic device as an event To ensure its addition, the delivery date will be 30 when the on-duty delivery date information is identified. 13 To ensure that the user is warned about this via the chat application (4), A file related to the task identified on the chat application (4) was shared. time, retrieving file metadata, converting file content to vectors To enable the calculation of similarity scores using cosine similarity management, If the similarity score is higher than a certain threshold, the file will be linked to 5. Enabling task association, so that tasks are automatically linked in chat. the assignment of the persons in (4) and the calendar of duties to ensure that their processing is carried out in their applications (3) messaging-based characterized by having at least one server (5) configured An intelligent system for work groups (1). 10 2. With the hardware on it, the user's chat application (4) enabling it, group chat with other registered users mobile phones, tablets, and computers configured to enable this Claim 15 is characterized by an electronic device (2), which is a smart device. An intelligent system for messaging-based work groups, as in 1 (1).

3. Communication between the hardware on it and the user's chat application (4) to enable it to run at least one calendar application (3) that can install, calendar application (3) interface to enable the user to track time 20 as in Claim 1 or 2 characterized by the configured electronic device (2). An intelligent system for messaging-based work groups (1).

4. Communicate with the chat application (4), the group the user is in Data preprocessing 25 in the task logs of the relevant chat messages in the chat. to ensure it is cleaned, tokenized and normalized through the service from the above requests characterized by the server (5) configured for A smart system for messaging-based work groups like any other (1). 14 5. The task involved in normalizing chat messages in the data preprocessing service, BERT or high-dimensional vector representations of mesa and file contents Server configured to enable generation using SBERT-based algorithms. as in any of the above claims characterized by (5). An intelligent system for messaging-based work groups (1). 5 6. This is achieved by the buffer layer inside communicating with the embedding service. the subject is real-time task queries thanks to the buffer layer The results of embedding, which enables acceleration, are in temporary memory. 10 characterized by the server (5) configured to ensure its retention with messaging-based workgroups as in any of the above requests a related intelligent system (1).

7. By communicating with the chat application (4), the user's group chat The content of the messages can be analyzed using NLP-based intent detection models and NER models. 15 by enabling the processing of these chat messages in real time. configured to extract task information from their contents as in any of the above requests characterized by the server (5) An intelligent system for messaging-based work groups (1).

8. NLP-based intent analysis of the content of user messages (4) in the chat application. processing in real time with detection models and NER models by providing, if the text task statement is “prepare”, “complete”, “do”, “finish” If it contains information in this form, it is necessary to determine the mission intent, mission intent When detected, a confirmation window appears to the user asking, "Is this message a task?" to ensure that the chat application (4) is displayed within the user chat He answered “Yes” to the confirmation window shown to him by the application (4). the timer mode is activated and the mention area in the message interface changes to chat. server configured to ensure that (4) the application is shown to the user as in any of the above claims characterized by (5). An intelligent system for messaging-based work groups (1).

9. NLP-based intent analysis of user message content in the chat application (4) 5. Processing in real time with detection models and NER models By providing the person (entity), action, and time from the message marked as a task, Server configured to extract (datetime) information (5) messaging as in any of the requests characterized above. an intelligent system related to work groups based on (1).

10. “Elif, send the budget table on Friday” sent to user (4) via chat application. NLP-based intent detection of message content such as "complete by [date]". The analysis is done using models and NER models, and the assigned person is: Elif. Task: Complete the budget table, deadline: Friday. 15 characterized by the server (5) configured to enable its extraction with messaging-based workgroups as in any of the above requests a related intelligent system (1).

11. Eliminating entity ambiguity on the chat application (4) (Entity Enabling the operation with the Disambiguation module, so that the relevant chat 20 In the application (4), there is more than one user with the same name in the group. In this case, mention-based user authentication should be implemented, NLP When the name "Elif" is detected as a person's name in the text by the engine, at that moment, suddenly in the group If there are too many "Elif"s, this situation should be marked as an "ambiguous entity," then... 25 in the mention selection window via the chat application (4) to the user presented, without the user choosing one of the options presented to them Server configured to prevent the creation of the task (5) messaging as in any of the above requests characterized by an intelligent system related to work groups based on (1). 16 12. Eliminating entity ambiguity on the chat application (4) (Entity Enabling the process to be carried out using the Disambiguation module, and chatting during this process. Using the IDs of the (4) users in the application, the correct person was identified. and assigning the task to the correct user or person identified. to provide instant notifications to the correct person assigned via its push notification service 5 characterized by the server (5) configured to enable sending with messaging-based workgroups as in any of the above requests a related intelligent system (1).

13. A task was created for the user in the chat application (4) 10 the time, including the deadline for the relevant task. to provide the aforementioned deadline information automatically electronically. to ensure that the calendar application (3) is added to the (2) calendar application on the device, the said addition the process is run within the electronic device (2) via at least a calendar API 15 characterized by the server (5) configured to enable it to be done with with messaging-based workgroups as in any of the above requests a related intelligent system (1).

14. Communicating with the electronic device (2) user through the calendar application (3). In the calendar application for each relevant task (3) “Business Group Chat – [Task Name]” 20 The event description should include the title, task details, and chat link. characterized by the server (5) configured to ensure its location with messaging-based workgroups as in any of the above requests a related intelligent system (1).

15. Communicate with the calendar application (3) when the task time is approaching this activates the TaskReminderScheduler module on it, thereby 1 day before the deadline, 1 hour before the deadline, time remaining until the final deadline To enable notification scheduling, in the chat application (4) interface "Reminder: There is 1 hour left until the budget statement submission." This is the notification on the 30th. 17 characterized by the server (5) configured to enable display with messaging-based workgroups as in any of the above requests a related intelligent system (1).

16. Connect with the chat application (4), chat a task-related file 5 When shared in the chat environment in the application (3), both the message content and also to enable the retrieval of file metadata, Embedding on it Its service enables the conversion of text and file contents into vectors, and the converted files... storing the vectors in a distributed vector database within Cosine Calculating the similarity score of the task using the similarity method is also possible. 10 from the above requests characterized by the server (5) configured for A smart system for messaging-based work groups like any other (1).

17. If the similarity ratio is above the specified threshold (e.g., 0.85), the file will be deleted. automatically associating it with the relevant task, thus allowing the user to chat 15 In the task details screen (4) of the application, the documents related to the task are automatically characterized by the server (5) configured to enable it to see as messaging-based work as in any of the above requests an intelligent system related to groups (1).

18. Communicate with the chat application (4) and the calendar application (3), this way, users can see when they complete tasks or change deadlines. or the learning cycle of those behaviors when responding to notifications (feedback pipeline) ensures that it is collected in the learning cycle. 25 on the AI ​​scoring engine of the relevant data from the (feedback pipeline) module to enable the transfer of information, thereby allowing the processing of task inference from similar messages. making reminder times more accurate and optimizing them on a user-by-user basis (e.g., earlier notifications for users who frequently procrastinate), task structured to enable the improvement of prioritization estimates 18 as in any of the above requests characterized by the server (5) An intelligent system for messaging-based work groups (1).

19. In corporate environments, sifting through millions of messages to find tasks, files, or decisions. When the moments need to be identified, the past 5 (4) in the chat application processing of chat messages through a data preprocessing service to provide, the vector is sent to the embedding service on the processed data. conversion to a representation and similarity analysis via a distributed vector database. to enable the search, then to score the search results using an AI scoring engine Scoring contextually, the most relevant content is prioritized into 10 categories. its ordering, with the buffer layer on top, in the frequently used chat application (4) storing results for queries and providing instant access to the user. characterized by the server (5) configured to provide the delivery with messaging-based workgroups as in any of the above requests a related intelligent system (1). 15 20. By connecting with the chat application (4) and the calendar application (3) The user says "I have completed the budget table" or within the chat application (4). When they send messages like, "Let's move the delivery date to Monday," their intentions... to ensure processing with the detection module, then finding the relevant task ID, 20 Update the task management service within the relevant task and this Changes are made through the Synchronization Layer in both the chat application (4) and to ensure that the task status is reflected in the calendar application (3), Information such as file link and date details are shared among all users. to remain consistent, and these task updates are also part of learning cycle 25 The behavioral learning of the model is transferred to the (feedback pipeline) module. characterized by the server (5) configured to ensure its continuation with messaging-based workgroups as in any of the above requests a related intelligent system (1).

Images