Artificial intelligence-based to-do recommendation device and method

Abstract

To provide an artificial intelligence-based to-do recommendation device and method for supporting effective fulfillment of a project and work.SOLUTION: An artificial intelligence-based to-do recommendation device includes a memory and a processor that is electrically connected to the memory. The processor is configured to: receive a user request for generation or retrieval of meeting records from a user terminal; generate the meeting records or a meeting record list including the meeting records according to the user request; generate to-do recommendations related to conference contents of the meeting records through an artificial intelligence model when a to-do recommendation request related to the meeting records is received from the user terminal; and provide a to-do recommendation list including selection options for each to-do recommendation to the user terminal.SELECTED DRAWING: Figure 3

Description

[Technical field]

[0001] The present invention relates to a technology for recommending tasks (hereinafter referred to as "tasks"), and more specifically, to a technology for automatically recommending tasks related to minutes in the process of generating or referencing minutes that record the contents of meetings between business participants during the progress of a project, and for supporting task generation to support the effective execution of projects and business. [Background technology]

[0002] Generally, an Internet messenger corresponds to an application that transmits messages including text or graphics between users, and is realized as a chat room in which multiple users participate. In one embodiment, the Internet messenger includes a mobile messenger executed in a mobile environment (e.g., a mobile phone), such as KakaoTalk (registered trademark), LINE (registered trademark), WeChat (registered trademark), Facebook Messenger, etc. There is a tendency that such Internet messengers are increasingly being used in a variety of ways in the management and progress of business.

[0003] In particular, as the scale of a project gradually increases and the structure of the project becomes more complex, the number of chat rooms in which business participants simultaneously participate within a single project also increases, and business participants may find it difficult to easily grasp the flow of related business operations in the process of communication that is distributed and proceeding in a large number of chat rooms.

[0004] For this reason, various tools have been developed to support communication between the various participants in a project. For example, meeting minutes are used to record the contents of meetings between participants, and can be shared between users to easily share the progress of a project.

[0005] Furthermore, various tasks related to the contents of a meeting are generated and shared among participants, and a technique is required for improving work efficiency in generating and managing tasks. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] Korean Patent No. 10-1182535 Summary of the Invention [Problem to be solved by the invention]

[0007] The present invention has been made in consideration of the above-mentioned conventional technology, and an object of the present invention is to provide an artificial intelligence-based task recommendation device and method that automatically recommends tasks related to minutes of meetings that record the contents of meetings between business participants during the progress of a project, and supports task generation in the process of generating or referencing minutes, thereby supporting effective execution of projects and tasks. [Means for solving the problem]

[0008] In order to achieve the above-mentioned object, according to one aspect of the present invention, an artificial intelligence based task recommendation device includes a memory and a processor electrically connected to the memory, wherein the processor receives a user request for generating or querying minutes from a user terminal, generates the minutes according to the user request or generates a minutes list including the minutes, and, when a task recommendation request for the minutes is received from the user terminal, generates recommended tasks related to the meeting content of the minutes via an artificial intelligence model, and provides the user terminal with a recommended task list including selection options for each recommended task.

[0009] The processor can receive from the user terminal a dialogue recording file that records a dialogue between business participants regarding the content of the meeting along with a user request for generating minutes, convert the dialogue recording file into text to generate dialogue text regarding the content of the meeting, and generate minutes including a summary of the dialogue text and selection options for each recommended task.

[0010] The processor can transmit the dialogue recording file to an external STT (Speech-To-Text) server and receive the dialogue text from the STT server.

[0011] The processor can generate the minutes list according to a request from any one of the business participants on a chat room in which the business participants are participating or on a conference map for managing minutes.

[0012] If the minutes correspond to general minutes, the processor can add the recommended duties list to the minutes and convert the minutes into duties minutes.

[0013] The processor may remove a particular recommended task from the recommended tasks list when a task is generated for the particular recommended task included in the task minutes.

[0014] The processor may convert the task minutes to the general task minutes if all the recommended tasks are removed from the recommended task list.

[0015] The processor may generate a task related to a recommendation role when a user selection is input via the recommendation role-specific selection option.

[0016] The processor may set the user who inputs the user selection as a task instructor for the task, and set the corresponding recommendation role as task content for the task.

[0017] The processor may, if multiple selection options are selected by the user who input the user selection, sequentially generate tasks for each of the corresponding recommendation duties.

[0018] The processor can receive a user's audio file associated with the conference content from the user terminal, recognize the speech in the audio file to generate a script and a summary message converted into text, and display the summary message as a dialogue message in a chat room selected by the user.

[0019] The processor can receive the user's voice in real time, inputted in a streaming manner from the user terminal, and generate the voice file.

[0020] The processor may input the audio file into a previously constructed speech recognition model to generate the script and the summary message, respectively.

[0021] The processor may generate and associate a tag associated with the summary message with the summary message, and provide a query function for the summary message via the tag.

[0022] The processor may assign a favorite function to each of the conversation messages, store the favorite function, and provide a list of the conversation messages to which the favorite function is assigned via a favorite page.

[0023] In order to achieve the above-mentioned object, one aspect of the present invention provides an artificial intelligence based task recommendation method, which is executed in a task recommendation device including a memory and a processor electrically connected to the memory, and the method is executed by the processor and includes the steps of receiving a user request for generating or querying minutes from a user terminal, generating the minutes or generating a minutes list including the minutes in response to the user request, and, when a task recommendation request related to the minutes is received from the user terminal, generating recommended tasks related to the meeting content of the minutes via an artificial intelligence model and providing the user terminal with a recommended task list including selection options for each recommended task. Effect of the Invention

[0024] The present invention has the following effects. However, this does not mean that a specific embodiment should include all of the following effects or should include only the following effects, and the technical scope of the present invention should not be understood to be limited thereby.

[0025] The artificial intelligence-based task recommendation device and method according to the present invention can support the effective execution of projects and tasks by automatically recommending tasks related to minutes or supporting task generation in the process of generating or querying minutes that record the contents of meetings between business participants during the progress of a project. [Brief description of the drawings]

[0026] [Figure 1] FIG. 1 is a diagram illustrating a task recommendation system according to the present invention. [Diagram 2] FIG. 2 is a diagram illustrating a system configuration of the task recommendation device of FIG. [Diagram 3] 3 is a diagram illustrating a functional configuration of the processor in FIG. 2. [Figure 4] 2 is a flow chart illustrating an artificial intelligence based job recommendation method according to the present invention. [Diagram 5] FIG. 1 is a diagram illustrating an embodiment of minutes according to the present invention. [Figure 6] FIG. 2 is a diagram illustrating an embodiment of a process for generating minutes according to the present invention. [Figure 7] 2 is a diagram illustrating an embodiment of a dialogue message generation process according to the present invention; FIG. [Figure 8] 1 is a diagram illustrating an embodiment of a process for providing related functions in a chat room according to the present invention; [Figure 9] FIG. 2 is a diagram illustrating an embodiment of a task generation process according to the present invention. [Figure 10] FIG. 1 is a diagram illustrating an embodiment of a task modification process according to the present invention. [Figure 11] 1 is a diagram illustrating general minutes and task minutes according to the present invention. FIG. [Figure 12] 1 is a diagram illustrating an embodiment of a recommended task list update process according to the present invention; [Figure 13] 4 is a diagram illustrating a process of sharing a dialogue message related to a user's voice in a chat room according to the present invention; [Figure 14] FIG. 13 illustrates one embodiment of a detailed view page of an interaction message in accordance with the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0027] The description of the present invention is merely an embodiment for structural or functional description, and therefore the technical scope of the present invention should not be construed as being limited by the embodiment described in the present specification. In other words, since the embodiment can be modified in various ways and can have various forms, the technical scope of the present invention should be understood to include equivalents that can realize the technical idea. In addition, the objectives or effects presented in the present invention do not mean that a specific embodiment should include all of them or should include only such effects, and therefore the technical scope of the present invention should not be understood as being limited thereby.

[0028] Meanwhile, the meanings of the terms used in this specification should be understood as follows.

[0029] Terms such as "first" and "second" are used to distinguish one component from another component, and should not be used to limit the scope of the technology. For example, a first component may be named a second component, and similarly, a second component may be named a first component.

[0030] When an element is said to be "connected" to another element, it should be understood that it is directly connected to the other element, or that there may be other elements in between. In contrast, when an element is said to be "directly connected" to another element, it should be understood that there are no other elements in between. Meanwhile, other expressions describing the relationship between elements, such as "between" and "immediately between," or "adjacent to" and "directly adjacent to," should be interpreted in the same way.

[0031] Singular expressions shall be understood to include plural expressions unless the context clearly indicates otherwise, and terms such as "comprise" or "have" are intended to specify the presence of embodied features, numbers, steps, operations, components, parts, or combinations thereof, and do not preclude the presence or additional possibility of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0032] The identifiers (e.g., a, b, c, etc.) in each step are used for convenience of description, and the identifiers do not describe the order of each step, and each step may occur in a different order than stated unless the context clearly dictates a particular order, i.e., each step may occur in the same order as stated, may occur substantially simultaneously, or may occur in the opposite order.

[0033] The present invention is embodied as computer readable codes in a computer readable recording medium, which includes any kind of recording device in which data that can be read by a computer system is stored. Examples of computer readable recording media include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage device, etc. Also, the computer readable recording media may be distributed among computer systems connected by a network, so that the computer readable codes may be stored and executed in a distributed manner.

[0034] All terms used herein, unless otherwise defined, have the same meaning as commonly understood by a person of ordinary skill in the art to which the present invention pertains. Terms defined in commonly used dictionaries should be interpreted to be consistent with the meaning they have in the context of the relevant art, and should not be interpreted as having an ideal or overly formal meaning unless expressly defined herein.

[0035] FIG. 1 is a diagram illustrating a task recommendation system according to the present invention.

[0036] Referring to FIG. 1, a task recommendation system 100 includes a plurality of user terminals 110 , a task recommendation device 130 , and a database 150 .

[0037] The user terminal 110 corresponds to a computing device operated by a user. For example, the user terminal 110 may be realized as a desktop PC, a notebook PC, a tablet PC, or a smartphone, but is not limited thereto, and may be realized as various devices.

[0038] There may be one or more user terminals 110, and in this case, the user terminal may be any one or more of the first user terminal 110a, the second user terminal 110b, and the third user terminal 110c. For convenience, the user terminal 110 used by the first user is referred to as the first user terminal 110a, the user terminal 110 used by the second user is referred to as the second user terminal 110b, the user terminal 110 used by the third user is referred to as the third user terminal 110c, etc.

[0039] In an embodiment of the present invention, multiple users are included in one or more user groups, which are referred to as a first user group, a second user group, a third user group, etc. Meanwhile, one user may be included in one or more user groups at the same time.

[0040] Furthermore, the multiple users correspond to business participants who participate in a joint project or business. For example, business participants include a business director who instructs the business, a business executor who executes the business, and business associates related to the business.

[0041] In this case, there is one whole project, which can include multiple business projects that proceed independently. In addition, plans, cards, notes, tasks, etc. are created and stored in relation to the whole project or the business projects.

[0042] Here, a plan corresponds to a business plan established to achieve a specific goal, a card corresponds to a management card for a series of tasks, a note corresponds to a business record that stores content related to the tasks, and a task corresponds to various unit tasks that are created and processed according to a plan, card, or note.

[0043] Furthermore, in the process of processing a plan, a card, a note, or a task, objects such as messages, files, and photos (or images) can be shared between users, and a chat room can be provided for users to have a conversation and share objects. In this case, messages shared through the chat room can include conversation messages related to daily content and business messages related to business content.

[0044] In one embodiment, at least one of the user terminals 110 is a mobile terminal and is connected to the task recommendation device 130 via cellular or Wi-Fi communication. In another embodiment, at least one of the user terminals 110 is a desktop PC and is connected to the task recommendation device 130 via the Internet.

[0045] The task recommendation device 130 corresponds to a computing device connected via a network to at least one user terminal 110. In one embodiment, the task recommendation device 130 can manage at least one user group in which other users associated with one user are included as members, i.e., business participants.

[0046] In one embodiment, the task recommendation device 130 is connected to the user terminal 110 through a dedicated agent installed in the user terminal 110. Here, the dedicated agent corresponds to an agent program, which is software that can link the user terminal 110 and the task recommendation device 130 with each other under the approval of the user terminal 110 when installed in the user terminal 110.

[0047] Meanwhile, the connection and coupling between the task recommendation device 130 and the user terminal 110 described here corresponds to one embodiment, and may be modified and applied in various forms within normal limits according to various operating and implementation environments.

[0048] The database 150 corresponds to a storage device that stores various information required in the operation process of the task recommendation device 130. For example, the database 150 may store, but is not limited to, a dialogue recording file that records the contents of a meeting, a voice file that records the voice of a user, or text information converted through voice recognition. The task recommendation device 130 may store information collected or processed in various forms in the process of performing the AI-based task recommendation method according to the present invention.

[0049] Furthermore, although in FIG. 1 the database 150 is shown as a logical storage device included in the task recommendation device 130, it is not necessarily limited thereto and may be implemented as a device independent of the task recommendation device 130.

[0050] FIG. 2 is a diagram illustrating a system configuration of the task recommendation device of FIG.

[0051] Referring to FIG. 2, the task recommendation device 130 includes a processor 210 , a memory 230 , a user input / output unit 250 , and a network input / output unit 270 .

[0052] The processor 210 executes the artificial intelligence based task recommendation procedure of the present invention, manages the memory 230 that is read or written in such process, and schedules synchronization times between the volatile and non-volatile memories in the memory 230.

[0053] The processor 210 controls the overall operation of the task recommendation device 130, and is electrically connected to the memory 230, the user input / output unit 250, and the network input / output unit 270, and controls the data flow between them. The processor 210 is realized by a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit) of the task recommendation device 130.

[0054] The memory 230 includes a secondary storage device implemented as a non-volatile memory such as a solid state disk (SSD) or a hard disk drive (HDD) used to store all data required for the task recommendation device 130, and a primary storage device implemented as a volatile memory such as a random access memory (RAM). Thus, the memory 230 may be implemented as a volatile and non-volatile memory, and if implemented as a non-volatile memory, it may be implemented to be connected via a hyperlink.

[0055] The user input / output unit 250 includes an environment for receiving user input and an environment for outputting specific information to a user, including an input device including or connected to an adapter such as a mouse, trackball, touchpad, graphic tablet, scanner, touch screen, keyboard, or pointing device, and an output device including an adapter such as a monitor or touch screen. In one embodiment, the user input / output unit 250 corresponds to a computing device connected via a remote connection, in which case the task recommendation device 130 acts as a server.

[0056] The network input / output unit 270 provides a communication environment for connecting to the user terminal 110 via a network, and includes, for example, adapters for communication such as a LAN (Local Area Network), a MAN (Metropolitan Area Network), a WAN (Wide Area Network), and a VAN (Value Added Network). In addition, the network input / output unit 270 is realized to provide a short-range communication function such as WiFi (registered trademark) or Bluetooth (registered trademark) for wireless data transmission, or a wireless communication function of 4G or higher.

[0057] FIG. 3 is a diagram illustrating the functional configuration of the processor in FIG.

[0058] 3, the task recommendation device 130 executes the AI-based task recommendation method according to the present invention through the processor 210. To this end, the processor 210 includes a user request receiving unit 310, a user request processing unit 330, a task recommendation unit 350, a task management unit 370, and a control unit (not shown in FIG. 3).

[0059] In this case, the embodiment of the present invention does not include all of the above functional configurations at the same time, but may be realized by omitting some of the above configurations or selectively including some or all of the above configurations according to each embodiment. Furthermore, one embodiment of the present invention may be realized as independent modules selectively including some of the above configurations, and the AI-based job recommendation method according to the present invention may be performed through the interworking of each module. The operation of each configuration will now be described in detail.

[0060] The user request receiving unit 310 receives a user request for creating or referencing minutes from the user terminal 110. Here, minutes correspond to a business record that records the contents of a meeting discussed during a meeting held between business participants, and corresponds to a type of note generated in relation to an entire project or a business project. Also, a user request corresponds to a command or information that a user requests the system to execute in relation to minutes, and is generated by a user (e.g., a business participant) on the user terminal 110. That is, user requests include a request to create minutes, a request to refer to minutes, a request to delete minutes, a request to correct minutes, and the like. The user request receiving unit 310 receives and interprets a user request input from the user terminal 110, and then supports a series of operations related to the request to be executed. The user request receiving unit 310 analyzes the user request to determine the type of user request, the target minutes, the requested task, and the like.

[0061] In addition, the user request receiving unit 310 may operate in conjunction with an independent module to process a user request. For example, in the case of a request to create minutes, the user request receiving unit 310 calls a minutes creation module to start a minutes creation operation. In the case of a request to query minutes, the user request receiving unit 310 calls a minutes query module to start a minutes query operation. The user request receiving unit 310 provides the user terminal 110 with an interface through which a request for creating or querying minutes can be input, and the interface is realized through an application operating on the user terminal 110.

[0062] In one embodiment, the user request receiving unit 310 may receive from the user terminal 110 a dialogue recording file in which dialogue between business participants regarding the contents of the conference is recorded together with a user request for generating minutes. In particular, the user request receiving unit 310 may receive from the user terminal 110 a dialogue recording file in which dialogue between business participants regarding the contents of the conference is recorded in conjunction with the user terminal 110. At this time, the dialogue recording file corresponds to a file in which dialogue voices between business participants who participated in the conference are recorded, and is recorded using a recording function included in the user terminal 110. The user request receiving unit 310 directly receives the dialogue recording file from the user terminal 110, or generates the dialogue recording file by recording dialogue contents transmitted from the user terminal 110 in real time.

[0063] In one embodiment, the user request receiving unit 310 can receive a user's voice file related to the contents of the conference from the user terminal 110. Here, the voice file corresponds to a file in which the user's voice is recorded, and is recorded using a recording function included in the user terminal 110. The user can directly record voice through the user terminal 110 to generate a voice file, and can select and transmit any one of various voice files stored in the user terminal 110.

[0064] Also, the user request receiving unit 310 can receive the user's voice inputted from the user terminal 110 in a streaming manner in real time to directly generate a voice file. That is, the user directly records his / her voice on the user terminal 110 equipped with a voice recording function such as a smartphone, tablet, or PC, and the user request receiving unit 310 receives the user's voice recorded in real time in a streaming manner in conjunction with the user terminal 110 to generate a file in a voice data format such as WAV or MP3. At this time, the user request receiving unit 310 compresses the voice data received in real time to save storage space for the voice file, and determines the compression method in consideration of the network condition, etc.

[0065] The user request processing unit 330 generates minutes in response to a user request, or generates a minutes list including the minutes. Specifically, the user request processing unit 330 can generate minutes that record the contents of a meeting in response to a request to generate minutes. At this time, the minutes are manually generated by business participants who attended the meeting, or are automatically generated by artificial intelligence based on a file that records the contents of the meeting, and the generated minutes are stored and managed in the database 150. In addition, the user request processing unit 330 can inquire about the minutes in response to a request to inquire about the minutes, and generate a list related to one or more minutes as an inquiry result. That is, the minutes list corresponds to a list of minutes generated as an inquiry result.

[0066] More specifically, when the user request processor 330 receives a dialogue recording file in response to a request to generate minutes, it converts the dialogue recording file into text to generate dialogue text related to the contents of the meeting, and then generates minutes including a summary of the dialogue text and a selection option for each recommended task. The user request processor 330 can recognize the voice of the business participants from the dialogue recording file using a STT (Speech-To-Text) algorithm or a voice recognition model, and generates dialogue text by identifying the dialogue of each business participant from the voice. At this time, the dialogue text corresponds to a text file generated by converting the dialogue contents recorded during the meeting between the business participants into text, and is generated via an external STT (Speech-To-Text) server as necessary.

[0067] Furthermore, the dialogue text can be used as input data for the text summarization model, and the text summarization model can generate a summary sentence summarizing the dialogue text as an output. Here, the text summarization model corresponds to a language model that receives a text or a text file as an input and generates a summary sentence summarizing the corresponding text or text file. Furthermore, the text summarization model can be designed to generate a summary sentence regarding the conference content while also generating a recommendation task related to the conference content.

[0068] For example, a text summarization model is realized as ChatGPT. ChatGPT is a conversational AI chatbot that is a super-large language model. In particular, GPT is an abbreviation for "Generative Pre-trained Transformer," and corresponds to generative AI that pre-trains a huge amount of data through machine learning and generates sentences. In this case, the text summarization model can selectively generate at least one of a summary sentence and a recommendation by setting the input and output data format and structure.

[0069] In one embodiment, the user request processor 330 may transmit a dialogue recording file to an external STT (Speech-To-Text) server and receive a dialogue text from the STT server. For example, the external STT server may include a CLOVA Speech server, a Whisper server supporting multilingual voice recognition, etc. That is, the user request processor 330 may generate a dialogue text regarding the contents of a meeting between business participants via an external STT server providing an Automatic Speech Recognition (ASR) service.

[0070] In one embodiment, the user request processor 330 may insert the minutes into a chat room in which the business participant is participating or into a conference map for managing the minutes, depending on the location of the minutes generation request. Here, the conference map corresponds to a dedicated management tool for managing the minutes. Also, when the minutes are generated and inserted into the chat room in response to the minutes generation request, a message regarding the minutes generation is shared through the chat room, and the business participant can directly select the message to access the minutes or download the minutes file. Also, when the minutes are generated and inserted into the conference map in response to the minutes generation request, various functions such as inquiry, modification, and deletion of the minutes are performed through a dedicated interface. Meanwhile, the minutes inserted into the conference map or chat room are stored and managed through an independent memory area associated with the conference map or chat room.

[0071] In one embodiment, the user request processor 330 may generate a minutes list according to a request for inquiry from any one of the business participants on a chat room in which the business participants participate or on a conference map for managing minutes. The user request processor 330 provides an interface for inputting a minutes inquiry request for each location where minutes can be inquired, and the interface is realized through an application running on the user terminal 110. For example, a business participant may activate a minutes inquiry function through a menu in a chat room, or may activate the minutes inquiry function through a dedicated interface provided on the conference map.

[0072] In addition, when minutes are viewed in a chat room in response to a request to view minutes, a minutes list generated as a result of the view is shared through the chat room, and business participants can directly select a particular minute from the minutes list to access the minute or download the minute file. In this case, the minutes list is displayed in the form of a message in the chat room and is displayed independently through separate interfaces on the conference map. In the minutes list provided in list form, the minutes are sorted and displayed according to the time when they were created, and are sorted according to various criteria other than the viewpoint.

[0073] In addition, when minutes are viewed on the conference map in response to a request to view minutes, various functions such as confirmation, modification, and deletion of each minute can be performed along with the minutes list through a dedicated interface. Meanwhile, the minutes list viewed in the conference map or chat room is temporarily generated, shared, and then deleted, and is selectively stored and managed through an independent memory area associated with the conference map or chat room as necessary.

[0074] In one embodiment, the user request processor 330 may generate independent query results for each type of minutes. For example, the user request processor 330 may generate query results by classifying general minutes and task minutes in response to a minutes query request, in which case, a minutes list for general minutes and task minutes is generated independently. Here, the general minutes correspond to minutes that record regular meeting contents, and the task minutes correspond to minutes that are generated by including one or more recommended tasks in addition to regular meeting contents.

[0075] In one embodiment, when the user request processor 330 receives a user's voice file related to the contents of the conference from the user terminal 110, it recognizes the voice of the voice file, converts it into text, generates a script and a summary message, and displays the summary message as a dialogue message in a chat room selected by the user. That is, the user request processor 330 can share the summary message generated by summarizing the script recognized from the user's voice as a dialogue message in a chat room in which the user is participating. Here, the dialogue message is defined as a voice memo, which is a message shared in a chat room.

[0076] At this time, other users participating in the chat room can recognize the conversation message as a message entered by the user. Thus, the user can easily convey the voice-recorded content to other users in the chat room through the conversation message (or voice memo) without having to enter a separate message in the chat room. The user request processing unit 330 can basically convert the text of the summary message into a conversation message format to share in the chat room, and display an image or video related to the content of the summary message together with the conversation message as necessary.

[0077] More specifically, the user request processor 330 performs an operation of analyzing voice data stored in a voice file using various voice recognition algorithms and then converting the voice data into text. For example, the voice recognition process includes a voice analysis process for decomposing a voice signal into elements such as frequency and intensity, a voice analysis process for analyzing a voice signal using a learned model, and a voice conversion process for converting the voice extracted through the model into text. Here, the script corresponds to the result of converting the user's voice content stored in a voice file into text, and the summary message corresponds to the result of briefly summarizing the main contents of the script.

[0078] Therefore, the script and summary message are represented as text data and are stored and managed in association with the corresponding user or voice file. For example, a voice file associated with a specific user is selectively referenced, or a script or summary message associated with a specific voice file is selectively referenced.

[0079] In one embodiment, the user request processor 330 may input the voice file into a pre-constructed voice recognition model to generate the script and summary message, respectively. The user request processor 330 may generate both the script and summary message from the voice file using only the voice recognition model. For this purpose, the voice recognition model is constructed by learning to receive the voice file as input data and generate output data including the script and summary message. In this case, the voice recognition model may be realized as one model, or may be realized including a plurality of sub-models as necessary. For example, the voice recognition model may be realized including a first model that generates a script from the voice file and a second model that generates a summary message from the script, and may be realized in a structure in which the output of the first model is connected to the input of the second model.

[0080] In one embodiment, when the user request processor 330 receives a selection of a conversation message from a chat room participant in a chat room, it displays an interface related to the conversation message. That is, each user participating in the chat room can directly select a conversation message to be shared, and when the user request processor 330 detects a selection of each user for each conversation message, it can provide a function related to the conversation message through a separate interface. The interface basically includes basic functions such as copying, deleting, and sharing a message.

[0081] In one embodiment, the user request processor 330 stores and manages summary messages or conversation messages in the database 150. The user request processor 330 assigns a unique identifier to identify each message, and stores the date and time each message was created, audio file and script information related to each message, and the content of the message (i.e., text) in the database 150 in association with each other. The user request processor 330 also stores and manages additional information such as tags and categories created in association with each message. The user request processor 330 provides functions such as search, delete, modify, sort, and filter as management functions for managing each message.

[0082] In one embodiment, the user request processor 330 generates tags related to the summary message, binds them to the summary message, and provides a function of browsing the summary message through the tags. The user request processor 330 performs an operation of analyzing the content of each message to extract key keywords, etc., to generate tags, and binds them to the message based on the extracted information. In this case, the types of tags include subject tags indicating the subject of each message, entity tags indicating people, places, and things appearing in each message, and description tags briefly explaining the content of each message. In addition, the user request processor 330 can search for tagged messages by tags, and performs an operation of filtering search results using a specific tag, recommending messages based on tags that the user is interested in, etc.

[0083] In one embodiment, the user request processing unit 330 blocks other users from viewing the corresponding audio file, script, or summary message according to a privacy setting set by the user in the process of providing at least one of an audio file, a script, and a summary message related to the message through a detailed view page of the conversation message. The user request processing unit 330 provides a detailed view function for each conversation message generated in association with an audio file and shared in a chat room, and provides a detailed view page for displaying detailed content of the conversation message in the chat room.

[0084] For example, user a can share a text-format conversation message through a voice file in which his / her voice is recorded, via a chat room. In addition, the user can set whether to make public the voice file, script, summary message, etc. associated with the conversation message as the setting information of the conversation message. In other words, if user a sets the voice file associated with the conversation message m to private, user a can view all of the voice file, script, and summary message, while other user b can view only the script and summary message excluding the voice file set to private.

[0085] In one embodiment, the user request processor 330 selectively provides a message translation function in the process of providing detailed contents of a message through a detail view page of a conversation message. The user request processor 330 provides a separate and independent interface for the detail view page. That is, the user request processor 330 provides a message translation function as one of various functions within the interface for the detail view page. To this end, the interface is implemented to operate in conjunction with a translation engine. A user can receive translation contents of a conversation message by selecting the message translation function. The user request processor 330 can provide each conversation message together with its translation through the interface and replace a specific conversation message with the translation and display it. Meanwhile, the user request processor 330 can provide the same function for summary messages.

[0086] In one embodiment, the user request processor 330 may assign a favorite function to each conversation message, store the favorite function, and provide a list of the conversation messages to which the favorite function has been assigned via a favorite page. Here, the favorite page corresponds to a dedicated interface that provides detailed operations and functions related to the favorite function. The favorite page may be provided within a chat room or provided via an independent interface from outside the chat room. For example, the favorite page may provide a list of conversation messages to which the favorite function has been assigned for each chat room or for all chat rooms.

[0087] Meanwhile, the favorite function corresponds to a function that supports a user to easily view and manage important or frequently used conversation messages by adding them to a separate list. That is, a user can add the favorite function to conversation messages created by the user as well as conversation messages created by other users shared through a chat room.

[0088] For example, the user request processor 330 may provide a favorites menu from a detail view page for each conversation message, and the user may add the conversation message to a favorites list by selecting the favorites menu. The user request processor 330 stores the favorites list in association with a user account. The user request processor 330 may provide the favorites list to the user terminal 110 through various interfaces.

[0089] In addition, the user request processing unit 330 can sort the favorites list according to various criteria such as most recently added, by title, or by date, and selectively remove conversation messages desired by the user from the favorites list. The user request processing unit 330 manages important conversation messages through separate favorites lists for each chat room or for the entire chat room. The user request processing unit 330 provides a function for sharing the favorites list with other users, a tag function for the favorites list, and a notification to the user when a conversation message added to the favorites list is updated.

[0090] When the task recommendation unit 350 receives a task recommendation request for the minutes from the user terminal 110, it generates recommended tasks for the meeting contents of the minutes through an artificial intelligence model, and can provide a recommended task list including selection options for each recommended task to the user terminal 110. The task recommendation unit 350 receives a signal regarding the task recommendation request from the user terminal 110, and starts a response operation in response to the reception of the task recommendation request.

[0091] In this case, the AI ​​model for recommending tasks is implemented as a deep learning model that receives text related to the contents of a meeting as an input and generates recommended tasks as an output, but is not limited thereto. Also, the AI ​​model is constructed in conjunction with a task table that stores recommended tasks, and in this case, the AI ​​model generates index information of the table as recommended task information. That is, the task recommendation unit 350 queries the task table based on the index information output by the AI ​​model to generate recommended tasks.

[0092] The task recommendation unit 350 may also generate a recommended task list including selection options for each recommended task and provide it to the user terminal 110. Here, the selection options correspond to functions that are activated according to a user's selection for each of one or more recommended tasks. For example, the selection options are realized by checkboxes, etc. In other words, the recommended task list corresponds to a recommended task list and is represented as a set of pairs between recommended tasks and selection options. The user can individually select a recommended task through each selection option on the recommended task list. The task recommendation unit 350 generates a recommended task list in response to a task recommendation request and provides it to the user terminal 110, and the recommended task list is displayed through a dedicated interface implemented on the user terminal 110.

[0093] In one embodiment, when the minutes correspond to general minutes, the task recommendation unit 350 may add a recommended task list to the minutes to convert the minutes to task minutes. Here, the general minutes correspond to minutes that do not include a recommended task, and the task minutes correspond to minutes that include a recommended task. That is, the minutes are basically generated in a state of being general minutes, and when a recommended task is generated in response to a task recommendation request from a user (e.g., a business participant), the recommended task is generated and managed as an object separate from the minutes. At this time, the task recommendation unit 350 adds the generated recommended task list to the minutes as necessary, and as a result of adding the recommended task list to the minutes, the general minutes are converted to task minutes. The task recommendation unit 350 performs an operation of converting the general minutes to task minutes when a user's request is received or when a preset condition is met.

[0094] In one embodiment, the task recommendation unit 350 removes a specific recommended task from the recommended task list when a task is generated for a specific recommended task included in the task minutes. For example, when the minutes A includes recommended tasks a and b, if a task for recommended task b is generated by a user (e.g., a business participant), recommended task b is removed from the minutes A. Thereafter, when the minutes A is referenced, only recommended task a is referenced together with the minutes A. Meanwhile, the task recommendation unit 350 can maintain the recommended task as it is even when a task related to the task minutes is generated. That is, in the above example, even when a task for recommended task b is generated, the minutes A is still maintained in a state in which recommended tasks a and b are included, and recommended tasks a and b are also referenced when the minutes A is referenced.

[0095] In one embodiment, the task recommendation unit 350 converts the task minutes into general minutes when all recommended tasks are removed from the recommended task list. For example, when recommended tasks a and b are included in the minutes A, when a task for recommended task b is generated by a user (e.g., a business participant), recommended task b is removed from the minutes A. Thereafter, when a task for recommended task a is generated by the user, recommended task a is removed from the minutes A. As a result, when there is no recommended task included in the minutes A, the minutes A is automatically converted into general minutes.

[0096] In one embodiment, when a new recommended task list is generated in response to a task recommendation request while a specific meeting is a task minutes, the task recommendation unit 350 updates the existing recommended task list included in the specific meeting minutes to the new recommended task list. The recommended tasks associated with a specific meeting minutes may change at the time of a task recommendation request due to changes in conditions such as the content of the meeting being modified or time passing, and when a new recommended task list is generated, the task recommendation unit 350 updates the existing recommended task list included in the task minutes to the new list. That is, even if a recommended task already exists for a specific meeting minutes, a user (e.g., a business participant) can request additional task recommendation, and the task recommendation unit 350 generates a new recommended task for the meeting minutes in which a recommended task exists in response to the additional request.

[0097] When a user selection is input through a selection option for each recommendation task, the task management unit 370 generates a task related to the recommendation task. The user selects an activated selection option on a recommendation task list provided through an interface to request generation of a task related to the recommendation task. The task management unit 370 starts a task generation operation related to the recommendation task for which the user selection is input through the selection option.

[0098] In one embodiment, the task management unit 370 sets the user who inputs the user selection for the recommended task selection option as the task instructor of the corresponding task, and sets the recommended task as the task content of the task. A task can be generated including various work-related items. For example, a task is generated including the task instructor who generated the task, the task content, etc. The task management unit 370 automatically sets the user who requested the generation of the task as the task instructor, and automatically sets the recommended task as the task content.

[0099] In one embodiment, when a task is created, the task management unit 370 grants the user who requested the task creation the right to modify the task content. In other words, the task may be updated after the task is created by the task creator by inputting additional items. For example, the task creator may input a task performer who will handle the work related to the task after the task is created, or may input a deadline by which the task should be performed.

[0100] In one embodiment, when multiple selection options are selected by the user, the task management unit 370 sequentially generates tasks related to each of the corresponding recommendation tasks. When multiple recommendation tasks exist in one minutes, the task generator selects a selection option for each recommendation task to generate a task for each recommendation task. In this case, the corresponding tasks are generated sequentially according to the selection order of the task generator for the selection options. In addition to the recommendation tasks, the task management unit 370 receives input of task contents directly from the task generator to generate tasks, and thus multiple tasks related to one minutes are generated sequentially.

[0101] In one embodiment, the task management unit 370, when a task related to the corresponding recommended task is generated, adds the remaining recommended tasks in the recommended task list to the specific minutes, and converts the specific minutes into task minutes. Basically, when a recommended task list is provided in response to a task recommendation request, the business participant selects the recommended task and generates a task, and the remaining tasks for which no task is generated are removed without being stored. In contrast, the task management unit 370, in cooperation with the task recommendation unit 350, can add the remaining recommended tasks in the recommended task list to the specific minutes, and converts the specific minutes into task minutes, when a task related to a specific recommended task is generated. That is, when the task management unit 370 completes task generation, it transmits a task generation completion signal to the task recommendation unit 350, and the task recommendation unit 350 performs an operation of adding the recommended tasks in the recommended task list to the minutes and converting the general minutes into task minutes.

[0102] The control unit (not shown in FIG. 3) controls the overall operation of the task recommendation device 130 and manages the control flow or data flow between the user request receiving unit 310, the user request processing unit 330, the task recommendation unit 350, and the task management unit 370.

[0103] FIG. 4 is a flow chart illustrating the artificial intelligence based task recommendation method according to the present invention.

[0104] 4, the task recommendation device 130 receives a user request for generating or querying minutes from the user terminal 110 via the processor 210 (step S410). The task recommendation device 130 generates minutes or generates a minutes list including the minutes in response to the user request via the processor 210 (step S430).

[0105] In addition, when the task recommendation device 130 receives a task recommendation request for minutes from the user terminal 110 via the processor 210, it generates recommended tasks related to the meeting content of the minutes through an artificial intelligence model, and provides the user terminal 110 with a recommended task list including selection options for each recommended task (steps S450 and S470).

[0106] In one embodiment, the task recommendation device 130 generates a task related to the recommended task when a user selection is input through the recommended task selection option. At this time, the task generation operation is performed via the task management unit 370 of the processor 210.

[0107] FIG. 5 is a diagram for explaining an embodiment of the minutes of a meeting according to the present invention.

[0108] 5, the task recommendation device 130 queries and provides minutes 510 that record meeting contents 550 between business participants in response to a minutes query request. At this time, the minutes 510 are manually generated by the business participants or automatically generated by an artificial intelligence model.

[0109] In one embodiment, the minutes 510 may include various items, such as meeting contents 550 automatically summarized by the AI ​​model and recommended tasks 560 recommended by the AI ​​model, as shown in Fig. 5. For example, the minutes 510 may include a creator 520, a meeting date and time 530, a storage location 540, meeting contents 550, and recommended tasks 560.

[0110] The creator 520 corresponds to the user who created the minutes, and corresponds to any one of the business participants who attended the meeting. The creator information is automatically determined in response to a request to create minutes. The meeting date and time 530 includes the date, day of the week, time, etc., when the meeting was held, and is automatically determined based on the time of the request to create minutes.

[0111] Also, the storage location 540 corresponds to a location where the minutes 510 are generated and stored. For example, the storage location 540 corresponds to information on a space (e.g., a folder name, etc.) in which the minutes are actually stored in the user terminal 110 or the task recommendation device 130, and in some cases, is represented by link information that allows access to the minutes 510.

[0112] In addition, the meeting contents 550 corresponds to a summary of the meeting contents automatically summarized by the AI ​​model, and corresponds to text information generated through voice recognition from a dialogue recording file that records the meeting contents. The recommended tasks 560 corresponds to task information automatically recommended by the AI ​​model, and includes business information related to the meeting contents. The recommended tasks 560 are generated including one or more detailed tasks 570 and selection options 580. The detailed tasks 570 are expressed in a simple text format, such as the title of the task and the purpose of the task. The selection options 580 are generated independently for each detailed task 570, and are realized in various ways, such as a checkbox or a radio button.

[0113] In addition, the minutes 510 may include the name of the meeting, information on the meeting participants, etc., and may include various minutes menus. For example, the minutes menu may include a task generation reservation menu, a shared link generation menu, and a file, image, image upload menu, etc. In this case, the minutes menu is provided together with the minutes when the minutes are referenced after the minutes are generated.

[0114] The task creation reservation menu corresponds to a function that allows a business participant with access rights to reserve task creation based on a specific point in time after the minutes are generated. In other words, when task creation is reserved for time A, a notification or reminder for task creation is provided to the business participant who made the reservation at time A. The shared link creation menu corresponds to a function that generates a shared link that allows direct access to the minutes from outside as a link for sharing the minutes. The upload menu corresponds to a function that allows files and images related to the meeting contents to be added to the minutes after they are generated.

[0115] FIG. 6 is a diagram illustrating an embodiment of a process for generating minutes according to the present invention.

[0116] 6, the task recommendation device 130 generates a summary and a recommended task regarding the meeting contents through an artificial intelligence model 630. The task recommendation device 130 receives a dialogue recording file of the meeting contents between the business participants together with a user request for generating minutes, and converts the dialogue recording file into text to generate a dialogue text 610 regarding the meeting contents.

[0117] The task recommendation device 130 provides the dialogue text 610 as an input to the artificial intelligence model 630, and generates a summary sentence and a recommended task from the output of the artificial intelligence model 630. At this time, the artificial intelligence model is pre-constructed based on a language model that receives an input related to the text and generates an output related to the text. The output of the artificial intelligence model is generated in the form of an output vector 650 having a dimension of a specific size, and the task recommendation device 130 generates a summary sentence and one or more tasks using each component data of the output vector 650. Furthermore, the input of the artificial intelligence model may also be converted into vector-shaped data extracted from the dialogue text 610.

[0118] FIG. 7 is a diagram illustrating an embodiment of a process for generating a dialogue message according to the present invention.

[0119] Referring to FIG. 7, the task recommendation device 130 receives a voice file 710 of a user A recorded on the user terminal 110. The task recommendation device 130 generates a script and a summary message that are converted into text by recognizing the voice of the voice file 710. At this time, the summary message is generated based on the script and is generated through a pre-constructed voice recognition model 730. That is, the voice recognition model 730 is designed to receive a script as an input and output a summary message that summarizes the content of the corresponding text. The task recommendation device 130 displays the summary message as a dialogue message in a chat room 750 related to the user's selection.

[0120] In addition, the task recommendation device 130 stores and manages the script, summary message, and dialogue message associated with the voice file 710 of the user A in the database 150. When a dialogue message is selected in the chat room 750, the task recommendation device 130 queries the database 150 and provides the voice file, script, summary message, etc. associated with the corresponding dialogue message.

[0121] FIG. 8 is a diagram illustrating an embodiment of a process for providing related functions in a chat room according to the present invention.

[0122] Referring to Fig. 8, when the task recommendation device 130 receives a selection of a conversation message from a chat room participant of a chat room, it displays an interface related to the conversation message. In this case, the interface includes functions related to the conversation message. For example, functions such as (1) listening to an audio file, (2) script view, (3) message view, and (4) task creation are provided through the interface, and the user can select a specific function to access the corresponding content. The task recommendation device 130 accesses the database 150 to inquire about data related to the user's selection, and then provides it through the user terminal 110.

[0123] In addition, the task recommendation device 130 restricts the provision of an interface when a user who is participating in a chat room does not have the access right to the dialogue message. That is, when a dialogue message is selected by a user who does not have the access right, the task recommendation device 130 does not perform an interface providing operation according to the message selection. For this reason, when the task recommendation device 130 displays a dialogue message related to a summary message through a chat room, it sets an access right to the dialogue message.

[0124] That is, when the task recommendation device 130 displays a summary message extracted from a user's voice file as a conversation message in a chat room, the task recommendation device 130 grants the access right preset by the user. For example, when a user records his / her voice and shares it as a conversation message through a chat room, the access right can be set to restrict access to other users in the chat room, and the task recommendation device 130 can release the access right of other users while displaying the conversation message in the chat room.

[0125] FIG. 9 is a diagram illustrating one embodiment of a task generation process according to the present invention.

[0126] 9, the task recommendation device 130 queries general minutes or task minutes 910 according to whether the minutes contain recommended tasks, and generates a query result. Unlike general minutes, task minutes 910 include a list of recommended tasks 970 related to the contents of the meetings in the minutes. That is, a business participant can select a specific minutes on the minutes list provided as a result of querying the minutes to check the details, and can select a part of the recommended tasks included in the minutes to generate a related task. Here, a task corresponds to the smallest unit of work related to a project.

[0127] Specifically, the detailed items of the minutes provided to the business participant include a summary of the meeting contents and recommended tasks related to the meeting contents, and the recommended tasks include individually selectable selection options 930. At this time, the task recommendation device 130 changes the activation state of each recommended task according to the business participant and provides it. For example, if the business participant has normal authority or the recommended task has a high priority, the task recommendation device 130 activates and provides the selection option 930. In other words, the activation state of the selection option 930 corresponds to a state selectable by the business participant.

[0128] In Fig. 9, when task #1 and task #2 are selected by a business participant (USER1), the corresponding selection option 930 is converted to a "selected" state, and the task recommendation device 130 generates tasks (Task 1 and Task 2) that are matched to the corresponding tasks. At this time, Task 2 corresponds to the task 950 generated corresponding to task #2 among the recommended tasks. The corresponding task 950 corresponds to a temporary task with some items being blank.

[0129] In this case, when the corresponding task 950 is generated, the task instructor is automatically set to "USER1" who requested the generation of the corresponding task, and the task content is automatically set to the content of "duty #2" selected by USER1. Also, the corresponding task 950 is generated and stored with information on the remaining items other than the task instructor and task content left blank (empty), and then the corresponding items are input by the business participant (USER1).

[0130] Meanwhile, the corresponding task 950 may be generated and stored with all items filled in through additional input from the business participant, or may be generated and stored with some items left blank depending on the selective input from the business participant.

[0131] FIG. 10 is a diagram illustrating an embodiment of a task modification process according to the present invention.

[0132] 10, the task recommendation device 130 may generate a task 950 according to a selection by a business participant for each recommended task included in the minutes. In this case, the task recommendation device 130 grants the task modification authority to the business participant who generated the task 950.

[0133] In the case of Fig. 10, the task 950 is generated in such a state that the task instructor is set to the task participant "USER1" who requested the task generation of the recommended "task #2" in Fig. 9, and the content of the selected "task #2" is set to the task content. After that, the task recommendation device 130 receives inputs for blank items from the task participant "USER1" who has the authority to modify via the input interfaces 1010 and 1030, and updates the corresponding task 950.

[0134] For example, the task performer is set to another task participant “USER4” selected by the task participant “USER1” on the first input interface 1010, and the task deadline is set to “00.05.05” input by the task participant “USER1” on the second input interface 1030. That is, the task recommendation device 130 updates the corresponding task 950 when the remaining information is input by the task participant.

[0135] FIG. 11 is a diagram for explaining the general minutes and the task minutes according to the present invention.

[0136] Referring to FIG. 11, the task recommendation device 130 may add a recommended task list 1130 to a general minutes 1110 to convert it into a task minutes 1150. The general minutes 1110 correspond to a minutes that do not include a recommended task, and the task minutes 1150 correspond to a minutes that include a recommended task. That is, the task recommendation device 130 basically generates and stores the general minutes 1110 that do not include a recommended task. The task recommendation device 130 converts the general minutes 1110 into a task minutes 1150 by adding a recommended task list 1130 to the general minutes 1110 when a business participant requests or meets a specific condition. In this case, the business participant can also check the recommended tasks related to the minutes through the inquiry result for the minutes, and can easily generate a task related to the recommended task by selecting the recommended task as necessary.

[0137] FIG. 12 is a diagram illustrating an embodiment of a recommended task list updating process according to the present invention.

[0138] 12, when a task is generated according to a selection by a business participant for a specific recommended task on the recommended task list 1210 included in the task minutes, the recommended task device 130 removes the specific recommended task from the recommended task list 1210. For example, in FIG. 12, when the recommended task list 1210 includes task #1, task #2, and task #3, the business participant "USER1" selects the selection option for task #1 to generate the first task 1231 for task #1. When the generation of the first task 1231 for task #1 is completed, the recommended task device 130 removes task #1 from the recommended task list 1210.

[0139] In addition, the business participant "USER1" selects the selection option for task #2 to generate a second task 1233 for task #2. When the generation of the second task 1233 for task #2 is completed, the recommended task device 130 removes task #2 from the recommended task list 1210. If all recommended tasks are removed from the recommended task list 1210 by sequential task generation of the remaining recommended tasks including task #3, the recommended task device 130 converts the task minutes into general minutes.

[0140] FIG. 13 is a diagram illustrating a process of sharing a dialogue message related to a user's voice in a chat room according to the present invention.

[0141] 13, the task recommendation device 130 shares a dialogue message 1310 generated from a user's voice file through a chat room 1300. In one embodiment, the dialogue message 1310 generated as a result of recognizing the user's voice is displayed as a "voice memo" in the chat room 1300. That is, the voice memo includes a result of converting the user's voice into a text form and summarizing it, and is displayed in the chat room 1300 in the form of a dialogue message 1310.

[0142] Furthermore, the task recommendation device 130 can provide a detailed view page of the interactive message 1310 while displaying the interactive message 1310 via the chat room 1300. To this end, the task recommendation device 130 provides a detailed view menu 1330 together with the interactive message 1310.

[0143] That is, a user can access a detail view page that provides details of the corresponding conversation message 1310 by selecting (e.g., clicking) the detail view menu 1330 of the conversation message 1310. For example, the detail view page of the conversation message 1310 provides at least one of an audio file, a script, and a summary message related to the corresponding conversation message 1310. At this time, the audio file, the script, or the summary message provided via the detail view page may be restricted from being accessed by other users according to a privacy setting set by the user who created the corresponding conversation message 1310.

[0144] FIG. 14 is a diagram illustrating one embodiment of a detailed view page of a conversation message according to the present invention.

[0145] 14, the task recommendation device 130 may provide a dialogue message detail view page 1400. In one embodiment, the dialogue message generated based on the user's voice is defined as a voice memo, and the following description focuses on the voice memo.

[0146] More specifically, the voice memo detail view page 1400 is accessed by a user selecting a dialogue message in a chat room. The detail view page 1400 provides details associated with the voice memo. For example, the detail view page 1400 displays the title of the voice memo (e.g., voice memo #1), the date and time of creation, the creator, etc. In addition, the detail view page 1400 displays the user's audio file 1410, summary message 1430, and script 1450 associated with the voice memo.

[0147] A user can select an audio file 1410 displayed on the detail view page 1400, download the file itself, and check the summary message 1430 and the entire contents of the script 1450. If the overall length of the script 1450 is long, the script 1450 is converted into a file format such as "voice memo recording.txt" and attached to the detail view page 1400. Furthermore, if the voice memo is set to private by the creator of the voice memo, other users who do not have access rights are restricted from downloading the audio file 1410 that is set to private, are restricted from viewing the summary message 1430 that is set to private, and are restricted from downloading the attached file related to the script 1450 that is set to private.

[0148] Although the present invention has been described above with reference to preferred embodiments, it will be understood that those skilled in the art can make various modifications to the present invention without departing from the spirit and technical scope of the present invention. [Explanation of symbols]

[0149] 100: Job recommendation system 110: User terminal 130: Duty recommendation device 150: Database 210: Processor 510:Minutes 610: Dialogue text 630: Artificial Intelligence Model 650: Output vector 730: Speech recognition model 750, 1300: Chat rooms 910: Minutes of Service 930:Selection options 950: Task 970: Recommended task list 1010: First input interface 1030: second input interface 1110: General minutes 1130, 1210: Recommended Works List 1150:Minutes of Service 1231: First Task 1233: Second Task

Claims

1. Memory, a processor electrically connected to the memory; The processor receiving a user request for generating or querying minutes from a user terminal; generating the minutes in response to the user request or generating a minutes list including the minutes; The artificial intelligence-based task recommendation device is configured to, when receiving a task recommendation request for the minutes from the user terminal, generate recommended tasks related to the meeting content of the minutes through an artificial intelligence model, and provide the user terminal with a recommended task list including selection options for each recommended task.

2. The processor, receiving, from the user terminal, a dialogue recording file that records a dialogue between business participants regarding the contents of the meeting, together with a user request for generating the minutes of the meeting; converting the dialogue recording file into text to generate a dialogue text relating to the contents of the conference; The AI-based task recommendation device according to claim 1 , further comprising a minutes generating unit that generates a summary of the dialogue text and a selection option for each of the recommended tasks.

3. The processor, The AI-based task recommendation device of claim 2, wherein the dialogue recording file is transmitted to an external STT (Speech-To-Text) server, and the dialogue text is received from the STT server.

4. The processor, The artificial intelligence-based task recommendation device according to claim 1, characterized in that the minutes list is generated according to a query request from any one of the business participants on a chat room in which the business participants are participating or on a conference map for managing minutes.

5. The processor, The artificial intelligence-based task recommendation device according to claim 1, characterized in that if the minutes correspond to general minutes, the recommended task list is added to the minutes, and the minutes are converted into task minutes.

6. The processor, The artificial intelligence based task recommendation device of claim 5, characterized in that if a task is generated for a particular recommended task included in the task minutes, the particular recommended task is removed from the recommended task list.

7. The processor, The artificial intelligence-based task recommendation device according to claim 6, characterized in that, when all the recommended tasks are removed from the recommended task list, the task minutes are converted into the general minutes.

8. The processor, The AI-based task recommendation device according to claim 1 , further comprising: generating a task related to the recommended task when a user selection is input via the recommended task-specific selection option.

9. The processor, The artificial intelligence-based task recommendation device according to claim 8, further comprising: setting a user who inputs the user selection as a task instructor for the task; and setting the recommended task as a task content for the task.

10. The processor, 9. The artificial intelligence based task recommendation device of claim 8, further comprising: if multiple selection options are selected by the user who inputs the user selection, generating tasks for each of the recommended tasks in sequence.

11. The processor, receiving a user's voice file associated with the conference content from the user terminal; Recognizing speech in the audio file to generate a converted text script and summary message; The AI-based task recommendation device according to claim 1, characterized in that the summary message is displayed as a dialogue message in a chat room selected by the user.

12. The processor, The AI-based task recommendation device according to claim 11, wherein the voice file is generated by receiving the user's voice in real time, the voice being input from the user terminal in a streaming manner.

13. The processor, The artificial intelligence based task recommendation device of claim 11, further comprising inputting the audio file into a pre-built speech recognition model to generate the script and the summary message, respectively.

14. The processor, generating and binding a tag associated with the summary message to the summary message; The artificial intelligence based task recommendation device according to claim 11, characterized in that it provides a query function for the summary message via the tag.

15. The processor, A favorite function is assigned to each conversation message and the message is stored. The artificial intelligence-based task recommendation device according to claim 11, further comprising: providing a list of dialogue messages to which the favorite function has been assigned via a favorites page.

16. Memory, A method implemented in a task recommendation device including a processor electrically connected to the memory, The method is executed by the processor, receiving a user request for creating or viewing minutes from a user terminal; generating the minutes in response to the user request or generating a minutes list including the minutes; When a task recommendation request for the minutes is received from the user terminal, generating recommended tasks related to the meeting content of the minutes through an artificial intelligence model, and providing a recommended task list including selection options for each recommended task to the user terminal.

Images