Chat systems, chat methods, chat devices, and programs

Abstract

We provide a chat system, chat method, chat device, and program that guides the user to the desired operation at the appropriate time according to the situation. [Solution] In the chat system 1000, the chat device 1, which is configured to communicate with the terminal device 2, comprises one or more control units. The control units obtain responses from the terminal device 2 to messages generated by the machine learning model in the language model device 3, input the responses to the machine learning model, cause the machine learning model to determine whether or not to include an operation element operated on the terminal device 2 in the message, request the machine learning model to generate a message according to the determination result, and cause the message generated by the machine learning model to be displayed on the terminal device 2.

Description

[Technical Field]

[0001] The present invention relates to a chat system, a chat method, a chat apparatus, and a program. [Background Art]

[0002] In recent years, systems that generate answers in response to user inputs and provide information in an interactive format have been widely used. In such systems, by repeating the transmission and reception of messages between the user and the system, information corresponding to the user's request is sequentially provided.

[0003] Related to this, an information processing apparatus that presents candidate question sentences to the user based on keywords input by the user and presents a text associated with the selected question sentence to the user is known (see, for example, Patent Document 1). According to such an information processing apparatus, the user can access desired information by selecting a question sentence close to his or her own intention. [Prior Art Documents] [Patent Documents]

[0004] [Patent Document 1] Japanese Unexamined Patent Application Publication No. 2024-140625 [Summary of the Invention] [Problems to be Solved by the Invention]

[0005] In the above-mentioned conventional apparatus, the information presented to the user is mainly text information, and only a response corresponding to the user's input is presented. For this reason, it has not been possible to present an element for prompting the user to perform a specific operation at an appropriate timing in accordance with the content of the dialogue, and there have been cases where it is difficult to guide the user to an operation corresponding to the situation.

[0006] The present invention has been made in view of the above problems, and an object of the present invention is to guide a user to a desired operation at an appropriate timing corresponding to a situation. [Means for solving the problem]

[0007] According to one embodiment, a chat system configured to communicate with a terminal comprises one or more control units, the control units obtain a response from the terminal to a message generated by a machine learning model, input the response to the machine learning model, determine whether or not to include an operation element operated on the terminal in the message, and if it determines to include it, generate a message that includes the operation element, and if it determines not to include it, request the machine learning model to generate a message that does not include the operation element, and cause the message generated by the machine learning model to be displayed on the terminal. [Effects of the Invention]

[0008] According to one embodiment, the user can be guided to the desired operation at an appropriate time depending on the situation. [Brief explanation of the drawing]

[0009] [Figure 1] This is a diagram showing an example of the configuration of chat system 1000. [Figure 2] This figure shows an example of the hardware configuration of the information processing device 100. [Figure 3] This diagram shows an example of the functional configuration of the chat device 1. [Figure 4] This figure shows an example of the functional configuration of terminal device 2. [Figure 5] This is a flowchart showing an example of a chat method in the first embodiment. [Figure 6] This is a diagram showing an example of a chat screen. [Figure 7] This figure shows the first example of a chat screen that includes interactive elements. [Figure 8] This figure shows a second example of a chat screen that includes interactive elements. [Figure 9] This is a flowchart showing an example of a chat method in the second embodiment. [Figure 10] This figure shows an example of a chat screen displaying a message that includes at least some electronic information. [Figure 11] This figure shows an example of a chat screen displaying a message containing multiple interaction elements. [Modes for carrying out the invention]

[0010] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. Figures 1 to 8 are diagrams relating to the first embodiment. Figure 9 is a diagram relating to the second embodiment. In the description and drawings of the embodiments, components having substantially the same functional configuration are denoted by the same reference numerals, and redundant explanations are omitted.

[0011] (First Embodiment) <System Configuration> First, an overview of the chat system 1000 according to this embodiment will be described. The chat system 1000 is configured to communicate with the terminal device 2 and is an information processing system for guiding the user of the terminal device 2 to a desired operation through chat. In the chat system 1000, a machine learning model displays a message on the terminal device 2 in response to a response received from the user.

[0012] The chat takes place, for example, between a terminal device 2 operated by a user and a chat device 1. The chat progresses through the exchange of responses and messages. A response refers to information entered from terminal device 2 into the chat system 1000. A response may be an answer to a message, a question about information displayed on terminal device 2, or information entered in response to an operation on an operation element displayed on terminal device 2. A response may consist of text, include an electronic file, or be a combination of these.

[0013] A message is information generated by a machine learning model based on responses entered by terminal device 2 and displayed on terminal device 2. A message includes one or more questions for the user of terminal device 2. A message may be generated based on the history of previous chats and may include images or audio in addition to text. A message may include an action element to guide the user to a desired action. The desired action may be, for example, a CTA (Call To Action), or it may be a request to enter information for sending materials, such as an email address or address.

[0014] The operation element is a user interface that is displayed on the terminal device 2 and guides the user to a predetermined operation. The operation element may be included in a message generated by a machine learning model, or may be generated based on information for specifying the operation element (hereinafter referred to as "operation element specifying information") included in an output of the machine learning model. The operation element specifying information may include one or more pieces of information for specifying the display mode, type, content of a message, transition destination, input item or processing content of the operation element. The operation element specifying information may be any information as long as it is information interpretable as a message including an operation element by the chat device 1 or the terminal device 2. It is determined whether to include the operation element in a message based on a user's response and chat history. Thereby, the operation element is included in the message at an appropriate timing according to the user's situation or interest. The operation element may include, for example, a link, an input form, or another interface that accepts a user operation. The link may be one that implements redirection to a calendar display screen or another website. The input form may be one for allowing input of an email address, an address, or other information. A plurality of operation elements may be included in one message. Further, the operation element may be set by an administrator of the chat system 1000 via an external information processing apparatus. The administrator may be any person who is involved in providing the chat system 1000 such as construction, operation or development of the chat system 1000. Settings related to the operation element may be included in advance in a prompt input to the machine learning model, or may be managed as setting information stored in the storage unit 12 of the chat device 1. Note that the mode and setting method of the operation element are not limited thereto.

[0015] A user is a user who uses the chat system 1000 via a terminal device. For example, a user may be a person who accesses the online chat system 1000 using a web browser or a smartphone application, but the present invention is not limited thereto. For example, the chat system 1000 can provide information to a user and prompt the user to execute a CTA or transition to a predetermined process in an online service such as a website or SNS (Social Networking Service), but the application of the chat system 1000 is not limited thereto.

[0016] FIG. 1 is a diagram illustrating an example of the configuration of the chat system 1000. As illustrated in FIG. 1, the chat system 1000 includes a chat device 1, a terminal device 2, and a language model device 3 that are communicably connected to each other via a network N. The network N is, for example, a wired LAN (Local Area Network), a wireless LAN, the Internet, a public switched telephone network, a mobile data communication network, or a combination thereof. In the example of FIG. 1, the chat system 1000 includes one chat device 1, one terminal device 2, and one language model device 3, but may include a plurality of each of these devices.

[0017] The chat device 1 is an information processing device that performs chat with the terminal device 2 by transmitting a message to the terminal device 2 and acquiring a response from the terminal device 2. The chat device 1 is, for example, a server device, a PC (Personal Computer), a smartphone, a tablet terminal, or a microcomputer, but is not limited thereto. In the example of FIG. 1, the chat device 1 is one information processing device, but may be implemented as a system including a plurality of information processing devices connected via the network N.

[0018] The terminal device 2 is an information processing device used by a user and utilized for performing chat with the chat device 1. The terminal device 2 is, for example, a PC, a smartphone, a tablet terminal, a headset, or a head-mounted display, but is not limited thereto.

[0019] Language model device 3 is an information processing device having a machine learning model that generates new data not included in the training data by learning the regularities and structure of a large amount of training data. The machine learning model has the function of generating messages based on user responses and chat history. The machine learning model may also function as a so-called AI agent. Language model device 3 is, for example, a server device, a PC, a smartphone, a tablet terminal, or a microcomputer, but is not limited to these. In the example in Figure 1, language model device 3 is a single information processing device, but it may be realized as a system consisting of multiple information processing devices connected via a network N. Language model device 3 may be constructed using a Large Language Model (LLM). Language model device 3 is connected to network N so that it can be accessed from chat device 1. Note that language model device 3 may be installed inside chat device 1, or it may be connected to chat device 1 via a communication path different from network N.

[0020] <Hardware configuration of the information processing device 100> Next, the hardware configuration of the information processing device 100 will be described. Figure 2 shows an example of the hardware configuration of the information processing device 100. As shown in Figure 2, the information processing device 100 comprises a processor 101, memory 102, storage 103, communication I / F 104, input device 105, output device 106, and drive device 107, all interconnected via bus B.

[0021] The processor 101 controls the various components of the information processing device 100 and realizes the functions of the information processing device 100 by loading various programs, including the OS (Operating System), stored in the storage 103, into the memory 102 and executing them. The processor 101 is, for example, a CPU (Central Processing Unit), an MPU (Micro Processing Unit), a GPU (Graphics Processing Unit), an ASIC (Application Specific Integrated Circuit), or a DSP (Digital Signal Processor), but is not limited to these.

[0022] Memory 102 is, for example, ROM (Read Only Memory), RAM (Random Access Memory), or a combination thereof. ROM is, for example, PROM (Programmable ROM), EPROM (Erasable Programmable ROM), EEPROM (Electrically Erasable Programmable ROM), or a combination thereof. RAM is, for example, DRAM (Dynamic RAM) or SRAM (Static RAM), but is not limited to these.

[0023] Storage 103 stores various programs and data, including the OS. Storage 103 is, for example, flash memory, HDD (Hard Disk Drive), SSD (Solid State Drive), or SCM (Storage Class Memories), but is not limited to these.

[0024] The communication interface 104 is an interface for connecting the information processing device 100 to an external device via the network N and controlling communication. The communication interface 104 is, for example, Bluetooth®, Wi-Fi®, ZigBee®, or Ethernet®.

[0025] The input device 105 is a device for inputting information into the information processing device 100. The input device 105 may be, but is not limited to, a mouse, keyboard, microphone, scanner, camera, various sensors, or operation buttons.

[0026] The output device 106 is a device for outputting information from the information processing device 100. The output device 106 is, for example, a display device, a projector, a printer, a speaker, or a vibrator, but is not limited to these. In this embodiment, the input device 105 and the output device 106 may be configured as a touch panel integrated with the display device.

[0027] The drive device 107 is a device for reading and writing data to the recording medium 108. The drive device 107 is, for example, a magnetic disk drive, an optical disk drive, a magneto-optical disk drive, or an SD card reader, but is not limited thereto. The recording medium 108 is, for example, a CD (Compact Disc), a DVD (Digital Versatile Disc), an FD (Floppy Disk), an MO (Magneto-Optical disk), a BD (Blu-ray® Disc), a USB® memory, or an SD card, but is not limited thereto.

[0028] In this embodiment, the program may be written to the memory 102 or storage 103 during the manufacturing stage of the information processing device 100, or it may be provided to the information processing device 100 via the network N, or it may be provided to the information processing device 100 via a non-temporary, computer-readable recording medium such as a recording medium 108.

[0029] <Functional Configuration of Chat Device 1> Next, the functional configuration of the chat device 1 will be described. Figure 3 is a diagram showing an example of the functional configuration of the chat device 1. As shown in Figure 3, the chat device 1 comprises a communication unit 11, a storage unit 12, and a control unit 13.

[0030] The communication unit 11 is implemented by the communication interface 104. The communication unit 11 transmits and receives information with the terminal device 2 or the language model device 3 via the network N.

[0031] The memory unit 12 is implemented by memory 102 and storage 103. The memory unit 12 stores electronic information 121.

[0032] Electronic information 121 is information displayed on the terminal device 2 and may include content. Electronic information 121 may also be content embedded and displayed in an application such as a web browser or a smartphone app. The content may include electronic files such as slide presentations, document presentations, or videos, and may have an interactive configuration in which the displayed content transitions or is updated in response to user operations. The content of electronic information 121 may include, but is not limited to, information about services provided to the user. Electronic information 121 may be uploaded to the chat device 1 by the administrator of the chat system 1000 from an external information processing device and recorded in the storage unit 12. Recorded electronic information 121 may be displayed on the terminal device 2 by link or embedded format.

[0033] The control unit 13 is realized by the processor 101 reading and executing a program from memory 102 and cooperating with other hardware components. The control unit 13 controls the overall operation of the chat device 1.

[0034] The functional configuration of chat device 1 is not limited to the example above. For example, chat device 1 may have some of the above functional configurations, with the remainder provided by terminal device 2 or language model device 3. Chat device 1 may also have functional configurations other than those described above. Furthermore, each functional configuration of chat device 1 may be implemented by software as described above, or by hardware such as IC chips, SoCs, LSIs, or microcomputers.

[0035] <Functional configuration of terminal device 2> Next, the functional configuration of terminal device 2 will be described. Figure 4 is a diagram showing an example of the functional configuration of terminal device 2. As shown in Figure 4, terminal device 2 comprises a communication unit 21, a storage unit 22, a control unit 23, a display 24, and an operation unit 25. Terminal device 2 is operated by a user. Terminal device 2 is a specific example of a terminal as defined in the claims.

[0036] The communication unit 21 is implemented by the communication interface 104. The communication unit 21 sends and receives information to and from the chat device 1 or the language model device 3 via the network N.

[0037] The memory unit 22 is realized by memory 102 and storage 103.

[0038] The control unit 23 is realized by the processor 101 reading and executing a program from memory 102 and cooperating with other hardware components. The control unit 23 controls the overall operation of terminal device 2. The control unit 23 is connected to the chat device 1 via network N to enable communication and execute the chat method in the chat system 1000. Specifically, the control unit 23 requests a program from the chat device 1 through the operation of the operation unit 25. Subsequently, the program obtained by the control unit 23 is loaded into the control unit 23 of terminal device 2. The control unit 23 executes the loaded program. Once the program is executed, the control unit 23 executes the chat method.

[0039] The display 24 is, for example, a display that outputs information as a screen for a user-operable graphical user interface (GUI). The display 24 may be included in the housing of the terminal device 2 or it may be an external device. Specifically, the display 24 can be implemented as a display device such as a CRT display, liquid crystal display, organic EL display, or plasma display. It is preferable that these display devices be used in accordance with the type of terminal device 2.

[0040] The operation unit 25 is a device that inputs information to the terminal device 2 in response to user operations. The operation unit 25 receives operation inputs made by the user. The operation inputs are transmitted to the control unit 23 as command signals. The operation unit 25 may be implemented as a touch panel integrated with the display 24. When the operation unit 25 is implemented as a touch panel, the user can input tap operations, swipe operations, etc. to the operation unit 25. Instead of a touch panel, the operation unit 25 can be a switch button, mouse, trackpad, QWERTY keyboard, etc. It is preferable to use these input devices depending on the type of terminal device 2.

[0041] The functional configuration of terminal device 2 is not limited to the example above. For example, terminal device 2 may have some of the above functional configurations, with the chat device 1 having the rest. Also, terminal device 2 may have functional configurations other than those described above. Furthermore, each functional configuration of terminal device 2 may be implemented by software as described above, or by hardware such as IC chips, SoCs, LSIs, microcomputers, etc.

[0042] <How to chat> The chat method performed by the chat system 1000 will be described. Figure 5 is a flowchart showing an example of the chat method in the first embodiment.

[0043] (Step S101) The control unit 23 of terminal device 2 accesses chat device 1. Specifically, the control unit 23 of terminal device 2 accesses a chattable link or URL (Uniform Resource Locator) to chat device 1 via a web browser or smartphone app, and requests a program from chat device 1. Next, terminal device 2 reads the program obtained by the control unit 23. When the control unit 23 executes the loaded program, the chat method is executed, and the chat screen is displayed on the display 24 of terminal device 2. Note that the link or URL may also be information used to display electronic information 121 recorded in the storage unit 12 along with the chat screen on the display 24.

[0044] (Step S102) The control unit 13 of the chat device 1 requests a message from the language model device 3. Specifically, the control unit 13 inputs instruction information into the machine learning model of the language model device 3, instructing it to generate a message to start a chat with the terminal device 2. The instruction information may include a predetermined prompt, information obtained through communication with the terminal device 2, or electronic information 121 recorded in the storage unit 12. The information obtained through communication with the terminal device 2 may, but is not limited to, the date and time, region, account information, or cookie information.

[0045] (Step S103) The language model device 3 generates messages. Specifically, the language model device 3 generates messages based on instruction information input to the machine learning model.

[0046] (Step S104) The language model device 3 sends the message generated by the machine learning model to the chat device 1.

[0047] (Step S105) The control unit 13 of the chat device 1 sends the message generated by the machine learning model to the terminal device 2. The control unit 13 displays the message generated by the machine learning model on the display 24 of the terminal device 2.

[0048] (Step S106) The display 24 of terminal device 2 displays the messages generated by the machine learning model as a chat screen. The display 24 may also display electronic information 121 along with the messages. In this case, the control unit 13 of chat device 1 may send the electronic information 121 along with the messages generated by the machine learning model to terminal device 2. The electronic information 121 may also be displayed on the display 24 in step S101 when terminal device 2 accesses a link or URL of chat device 1. (Step S107) The control unit 23 of the terminal device 2 receives responses to messages from the user via the operation unit 25.

[0049] Figure 6 shows an example of a chat screen. The chat screen 500 is displayed, for example, on the display 24 of the terminal device 2. The chat screen 500 displays a chat area 501 and an electronic information area 502.

[0050] The chat area 501 displays the chat that took place between terminal device 2 and chat device 1. In Figure 6, the chat area 501 includes message 511, response 521, message 514, and input area 530.

[0051] Message 511 is information generated by a machine learning model. For example, message 511 includes two questions: a first message 512 and a second message 513. The first message 512 is a question about the user's position on terminal device 2. The question may be answered with text or with buttons. If the question can be answered with buttons, buttons are displayed in the message. For example, the first message 512 further includes answer buttons 512a to 512c for answering the question. When the user of terminal device 2 selects one of the answer buttons 512a to 512c via the operation unit 25, the text displayed on the selected answer button is entered into the input area 530. Next, the user of terminal device 2 operates "Send" displayed in the input area 530 via the operation unit 25, and the control unit 23 accepts the entered text as a response to the first message 512. Alternatively, by selecting one of the answer buttons 512a to 512c, the text displayed on the answer button may be sent to chat device 1 as a response to the first message 512.

[0052] The second message 513 is a question to confirm what the user of terminal device 2 is having trouble with or wants to know. The second message 513 does not contain any buttons. In this case, the user of terminal device 2 can answer the question in text. For example, the user of terminal device 2 enters text into the input area 530 via the operation unit 25. Then, the user of terminal device 2 operates the "Send" button displayed in the input area 530 via the operation unit 25, and the control unit 23 receives the entered text as a response to the second message 513.

[0053] Response 521 is the user's response to message 511. Response 521 is text entered into the input area 530 via the operation unit 25. Response 521 is displayed when the "Send" button is operated after text has been entered into the input area 530. Response 521 may also be displayed if any of the response buttons 512a to 512c are selected. The content of response 521 is sent to the chat device 1. Response 521 sent to the chat device 1 is input into a machine learning model and used to generate the next message.

[0054] Message 514 is information generated by a machine learning model. Message 514 is generated by inputting response 521 into the machine learning model and may contain information based on the content of response 521. Message 514 may contain personalized key points, such as the user's problem inferred from the content of response 521 or a summary of response 521. In addition to the content of response 521, message 514 may also be generated by considering the previous chat history. Furthermore, message 514 may include questions to obtain information for including action elements in subsequent messages.

[0055] The electronic information area 502 is an area where at least a portion of the content contained in the electronic information 121 recorded in the storage unit 12 is displayed. For example, if the electronic information 121 is a slide-format document, the electronic information display area 502 may display one or more slides. For example, if the electronic information 121 is a video, the electronic information display area 502 may display the video.

[0056] (Step S108) The control unit 23 of terminal device 2 sends the received response to chat device 1. The control unit 13 of chat device 1 receives the response corresponding to the message from terminal device 2.

[0057] (Step S109) The control unit 13 of the chat device 1 requests the language model device 3 to generate a message. Specifically, based on the response obtained from the terminal device 2, the control unit 13 inputs instruction information for generating the next message to be sent to the terminal device 2 into the machine learning model of the language model device 3. The instruction information may use predetermined prompts, and may include the response obtained from the terminal device 2 and the chat history up to that point. The instruction information may also include information that causes the machine learning model to determine whether or not to include an operation element in the generated message. For example, the instruction information may include a request for the machine learning model to determine whether or not to include an operation element in the message and to generate a message according to the determination result. In this case, the control unit 13 requests the machine learning model to generate a message that includes the operation element if it determines that it should be included, and to generate a message that does not include the operation element if it determines that it should not be included. Note that the determination conditions may differ for each operation element. The determination may also be made for each determination condition. In this case, if even one of the determination conditions is met, a message is generated that includes the operation element for which that determination condition was met.

[0058] (Step S110) The machine learning model of the language model device 3 determines whether or not to include an operational element in the generated message based on the instruction information.

[0059] (Step S111) If the machine learning model determines that the message does not contain any manipulation elements (step S111: No), the process proceeds to step S103. In step S103, the machine learning model of the language model device 3 generates a message that does not contain any manipulation elements.

[0060] (Step S112) If the machine learning model determines that an interaction element should be included (step S111: Yes), the machine learning model generates a message that includes the interaction element. The message may contain one or more interaction elements. For example, a message may be generated that includes all of the interaction elements that satisfy the determination condition. In this way, the machine learning model can determine whether or not to include an interaction element in a message based on the user's response and chat history, and generate a message according to the determination result.

[0061] (Step S113) The language model device 3 sends a message containing manipulation elements to the chat device 1.

[0062] (Step S114) The control unit 13 of the chat device 1 sends a message containing operation elements to the terminal device 2. The control unit 13 displays the message containing operation elements on the display 24 of the terminal device 2.

[0063] (Step S115) The display 24 of terminal device 2 displays a message containing operation elements as a chat screen.

[0064] (Step S116) The control unit 23 of the terminal device 2 receives operations on the operation elements from the user via the operation unit 25.

[0065] Figure 7 shows a first example of a chat screen including operation elements. The chat screen 500a is displayed, for example, on the display 24 of the terminal device 2. The chat screen 500a displays a chat area 501a and an electronic information area 502. In Figure 7, explanations that overlap with the chat screen 500 shown in Figure 6 are omitted as appropriate.

[0066] The chat area 501a displays the chat history, which shows the history of chats conducted between terminal device 2 and chat device 1. In Figure 7, the chat area 501a includes response 522, message 515, and input area 530. Response 522 is the response from the user of terminal device 2 to the message. Message 515 is generated based on the content of response 522.

[0067] Message 515 is information generated by a machine learning model. Message 515 includes an operation element 515a. Operation element 515a is a link that causes terminal device 2 to access a predetermined resource. The predetermined resource may be, for example, a link to a calendar for specifying the date of an interview, a link that redirects to another website, or a link that allows a file to be downloaded to terminal device 2, but is not limited to these. In Figure 7, operation element 515a is a link that displays a calendar. When the user of terminal device 2 selects operation element 515a via the operation unit 25, the calendar is displayed on the display 24 in response to the operation element 515a. Next, the user of terminal device 2 selects a date on the calendar via the operation unit 25, and the control unit 23 of terminal device 2 receives the selected date as a response to message 515. Note that messages containing operation elements are not limited to the link shown in Figure 7.

[0068] Figure 8 shows a second example of a chat screen including operational elements. The chat screen 500b is displayed, for example, on the display 24 of the terminal device 2. The chat screen 500b displays a chat area 501b and an electronic information area 502. In Figure 8, explanations that overlap with those of the chat screen 500 shown in Figure 6 are omitted as appropriate.

[0069] The chat area 501b displays the chat history, which shows the history of chats conducted between terminal device 2 and chat device 1. In Figure 8, the chat area 501b includes response 523, message 516, and input area 530. Response 523 is the response from the user of terminal device 2 to the message. Message 516 is generated based on the content of response 523.

[0070] Message 516 is information generated by a machine learning model. Message 516 includes an operation element 516a. Operation element 516a is an input form that accepts input from terminal device 2. The input form may be, for example, an email address input form, or an address, telephone number and recipient name input form, but is not limited to these. The information entered in the input form may differ depending on the content of the message. In Figure 8, operation element 516a is an input form that accepts an email address input. When a user of terminal device 2 enters an email address into operation element 516a via the operation unit 25, the control unit 23 of terminal device 2 accepts the entered email address as a response to message 516.

[0071] (Step S117) The control unit 23 of terminal device 2 sends the received operation content as a response to chat device 1. For example, if the received operation content is a date on a calendar, the control unit 23 sends the date to chat device 1. The control unit 13 of chat device 1 retrieves the date from terminal device 2 as a response corresponding to the message. For example, if the received operation content is an email address, the control unit 23 sends the email address to chat device 1. The control unit 13 of chat device 1 retrieves the email address from terminal device 2 as a response corresponding to the message.

[0072] (Step S118) The control unit 13 of the chat device 1 executes processing according to the content of the operation element. The content of the operation may include, for example, a date on a calendar, an email address, or other input information. For example, if the content of the operation is a date on a calendar, the control unit 13 may execute processing to notify the person in charge of the date. The notification means may be email, a chat application, or other communication means. A chat application refers to, for example, an application that has the function of sending and receiving messages between users. The notification may be performed by the control unit 13 or by the machine learning model of the language model device 3. Also, if the content of the operation is an email address, the control unit 13 may send an email to the received email address. The content of the email to be sent may be set based on the message displayed on the terminal device 2. For example, if the message proposes sending sales materials, a proposal, or a quotation, an email with the corresponding files attached may be sent. These attached files may be predetermined files or files generated by a machine learning model. Note that the processing according to the content of the operation is not limited to these.

[0073] (Second Embodiment) Next, the chat system 1000 of the second embodiment will be described. The chat system 1000 of the second embodiment differs from the first embodiment in that, instead of the machine learning model generating a message containing the operation elements, it generates operation element identification information. The differences from the first embodiment will be explained below. Note that the system configuration and the functional configuration of each device in the second embodiment will be omitted from the explanation as appropriate for parts that are common with the first embodiment.

[0074] <How to chat> The chat method performed by the chat system 1000 will be described. Figure 9 is a flowchart showing an example of the chat method in the second embodiment.

[0075] (Step S201) The control unit 13 of the chat device 1 requests the language model device 3 to generate a message. Specifically, based on the response obtained from the terminal device 2, the control unit 13 inputs instruction information for generating the next message to be sent to the terminal device 2 into the machine learning model of the language model device 3. The instruction information may use predetermined prompts, and may include the response obtained from the terminal device 2 and the chat history up to that point. The instruction information may also include information to cause the machine learning model to determine whether or not to include an operation element in the generated message. For example, the instruction information may include a request for the machine learning model to determine whether or not to include an operation element in the message and to generate a message according to the determination result. In this case, if the machine learning model determines that an operation element should be included, the control unit 13 requests the machine learning model to generate operation element identification information to identify the operation element, and if it determines that it should not be included, it requests the machine learning model to generate a message. Note that the determination conditions may differ for each operation element. Determination may also be made for each determination condition. In this case, if even one of the determination conditions is met, operation element identification information is generated to identify the operation element for which that determination condition was met.

[0076] (Step S202) The machine learning model of the language model device 3 determines whether or not to include an operational element in the generated message based on the instruction information.

[0077] (Step S203) If the machine learning model determines that the message does not contain any manipulation elements (step S202: No), the process proceeds to step S103. In step S103, the machine learning model of the language model device 3 generates a message that does not contain any manipulation elements.

[0078] (Step S204) If the machine learning model determines that an operation element should be included (step S202: Yes), the machine learning model generates operation element identification information. There may be one or more operation element identification information entries. For example, operation element identification information may be generated for multiple operation elements that satisfy the determination condition among multiple operation elements. In this way, the machine learning model can determine whether or not to include an operation element in a message based on the user's response and chat history, and generate a message that does not include the operation element or operation element identification information depending on the determination result.

[0079] (Step S205) The language model device 3 transmits information identifying the operation element to the chat device 1.

[0080] (Step S206) The control unit 13 of the chat device 1 generates a message based on the operation element identification information. Specifically, the control unit 13 generates a message corresponding to the type of operation element included in the operation element identification information. For example, if the type of operation element is a link, the control unit 13 generates a message that includes the link, the destination of the link, and text prompting the user to click the link. If multiple operation element identification information is generated, the control unit 13 generates a message that includes the operation element corresponding to each operation element identification information.

[0081] <Summary> As described above, according to the above embodiment, a chat system 1000 configured to communicate with a terminal device 2 is realized, comprising one or more control units 13, the control units 13 acquire responses from the terminal device 2 to messages generated by a machine learning model, input the responses to the machine learning model, determine whether or not to include operation elements operated on the terminal device 2 in the message, generate a message including the operation elements if it determines to include them, and generate a message without the operation elements if it determines not to include them, and display the message generated by the machine learning model on the terminal device 2.

[0082] In the above-described embodiment, the chat system 1000 generates messages based on the user's responses and chat history, and determines whether or not to include operational elements in the messages. This allows operational elements to be presented at an appropriate time according to the user's situation. As a result, the user can be smoothly guided to the desired operation without being required to input excessive information. Therefore, it is possible to facilitate scheduling meetings, sending materials, or transitioning to external services, thereby increasing the likelihood of business negotiations.

[0083] Since messages are generated by a machine learning model, the administrator of the chat system 1000 is burdened with the task of individually creating message content. While conventional systems required pre-preparing multiple message patterns corresponding to user responses, this embodiment dynamically generates messages based on user responses and chat history. This reduces the effort required to create messages and improves operational flexibility.

[0084] The administrator of the chat system 1000 can pre-set conditions for determining operation elements in the chat device 1 or prompt, and the system can then include operation elements in the message at the appropriate time according to the user's response and chat history. This eliminates the need to individually set the timing and conditions for presenting operation elements, thereby reducing the burden on the construction and operation of the chat system 1000.

[0085] By having the administrator of the chat system 1000 record electronic information 121 in the storage unit 12 of the chat device 1, the machine learning model can generate messages based on the electronic information 121. Therefore, the administrator of the chat system 1000 can start operating the chat system 1000 simply by uploading the electronic information 121, without needing to design individual message rules or display conditions for operation elements. This reduces the workload during the introduction and operation of the chat system 1000.

[0086] The control unit 13 may simultaneously display at least a portion of the electronic information 121, the response, and the message generated by the machine learning model as a chat screen on the display 24 of the terminal device 2. With this configuration, the user can proceed with the chat while referring to the content of the displayed electronic information 121. As a result, a response is input that corresponds to the user's interest or level of understanding based on the electronic information 121. Therefore, the chat system 1000 can include appropriate operation elements in the message based on the response, and efficiently guide the user to the desired operation.

[0087] <Variation> The message may be generated by the machine learning model to include at least a portion of the electronic information 121. In this case, the control unit 13 requests the machine learning model to generate a message based on the content of the response obtained from the terminal device 2 and the electronic information recorded in the storage unit 12. Figure 10 shows an example of a chat screen displaying a message that includes at least a portion of the electronic information. The chat screen 500c is displayed, for example, on the display 24 of the terminal device 2. The chat screen 500c displays a chat area 501c and an electronic information area 502. In Figure 10, explanations that overlap with the chat screen 500 shown in Figure 6 are omitted as appropriate.

[0088] The chat area 501c displays the chat history, which shows the history of chats conducted between terminal device 2 and chat device 1. In Figure 10, the chat area 501c includes response 524, message 517, and input area 530. Response 524 is the response from the user of terminal device 2 to the message. Message 517 is generated based on the content of response 524.

[0089] Message 517 is information generated by a machine learning model. Message 517 includes partial information 517a, which represents a part of the electronic information 121, and an operation element 517b. Partial information 517a is information that represents at least a part of the electronic information 121. For example, partial information 517a includes a part of the electronic information 121 displayed in the electronic information area 502 that supplements or explains the content of message 517. In this way, by including a part of the electronic information 121 in the message, the user can resolve their questions without having to search for the electronic information 121 displayed in the electronic information area 502. Therefore, the burden on the user can be further reduced.

[0090] The operation element 517b is an input form that accepts input from the terminal device 2. In Figure 10, the operation element 517b is an input form that accepts the input of an email address. When the user of the terminal device 2 enters an email address into the operation element 517b via the operation unit 25, the control unit 23 of the terminal device 2 accepts the entered email address as a response to message 517.

[0091] The message may be generated by the machine learning model to include multiple operation elements. Figure 11 shows an example of a chat screen displaying a message containing multiple operation elements. The chat screen 500d is displayed, for example, on the display 24 of the terminal device 2. The chat screen 500d displays a chat area 501d and an electronic information area 502. In Figure 11, explanations that overlap with those of the chat screen 500 shown in Figure 6 are omitted as appropriate.

[0092] The chat area 501d displays the chat history, which shows the history of chats conducted between terminal device 2 and chat device 1. In Figure 11, the chat area 501d includes response 525, message 518, and input area 530. Response 525 is the response from the user of terminal device 2 to the message. Message 518 is generated based on the content of response 525.

[0093] Message 518 is information generated by a machine learning model. Message 518 includes operation element 518a and operation element 518b. Thus, message 518 is generated including multiple operation elements. Operation element 518a is an input form that accepts input from terminal device 2. In Figure 11, operation element 518a is an input form that accepts the input of an email address. When the user of terminal device 2 enters an email address into operation element 518a via the operation unit 25, the control unit 23 of terminal device 2 accepts the entered email address as a response to message 518. Operation element 518b is a link that displays a calendar. When the user of terminal device 2 selects operation element 518b via the operation unit 25, the calendar is displayed on the display 24 in response to the operation on operation element 518b. Next, the user of terminal device 2 selects a date on the calendar via the operation unit 25, and the control unit 23 of terminal device 2 accepts the selected date as a response to message 518.

[0094] Thus, since a single message can contain multiple interaction elements, even if multiple interaction elements satisfy the judgment criteria, multiple options can be displayed on the screen according to the user's situation or interests. This allows the user to select the interaction that best suits their needs from among the displayed elements. Therefore, compared to the case where only one interaction element is displayed, the accuracy of guiding the user to the desired action can be improved.

[0095] Furthermore, the operation element may also be a settlement element that initiates a transaction with the user of terminal device 2. For example, if the operation element corresponds to a purchase or application, the user may initiate settlement processing as a response to a message by operating the operation element via the operation unit 25. Settlement processing may be performed in the control unit 13 of chat device 1, or by sending a settlement request to an external settlement system. For example, the system may be configured to send transaction information for settlement processing from terminal device 2 in response to user operation, and then execute settlement processing. In addition, prior to settlement processing, processing may be performed to accept input of user authentication information or payment information. Note that the transaction content and settlement method are not limited to these.

[0096] Furthermore, although the above-described embodiment assumed that the language model device 3 has a machine learning model, it is not limited to this. The machine learning model may, for example, be stored in the memory unit 12 of the chat device 1. Such a machine learning model may be an existing model or a model developed by the developer of the chat device 1.

[0097] <Note> This embodiment includes the following disclosures.

[0098] (Note 1) A chat system configured to communicate with a terminal, It comprises one or more control units, The control unit, The response to the message generated by the machine learning model is obtained from the terminal. The response is input to the machine learning model, and the machine learning model is requested to determine whether or not to include the operation element performed on the terminal in the message and to generate a message according to the determination result. The message generated by the machine learning model is displayed on the terminal. Chat system.

[0099] (Note 2) The previous message includes multiple of the aforementioned operational elements, The chat system described in Appendix 1.

[0100] (Note 3) The system further includes a storage unit that stores electronic information relating to services provided to the user, which is displayed on the terminal that accesses the chat system. The control unit requests the machine learning model to generate a message based on the content of the response and the electronic information. The terminal displays a message containing at least a portion of the electronic information generated by the machine learning model. The chat system described in Appendix 1.

[0101] (Note 4) The aforementioned operating element is a link that allows the terminal to access a predetermined resource. The chat system described in Appendix 1.

[0102] (Note 5) The control unit, in response to an operation on the link, causes the terminal to display a calendar for specifying a date. The chat system described in Appendix 4.

[0103] (Note 6) The aforementioned operating element is an input form that accepts input from the terminal. The chat system described in Appendix 1.

[0104] (Note 7) The control unit accepts the input of an email address from the terminal to the input form. The chat system described in Appendix 6.

[0105] (Note 8) The aforementioned operating element is a settlement element that initiates a transaction with the user of the terminal. The chat system described in Appendix 1.

[0106] (Note 9) The control unit displays the electronic information, the response, and the message generated by the machine learning model on the same screen of the terminal. The chat system described in Appendix 1.

[0107] (Note 10) The aforementioned operating elements are set by the user. The chat system described in Appendix 1.

[0108] (Note 11) If the control unit determines that the operation element should be included in the message, it generates a message that includes the operation element. If it is determined that it does not contain the element, it requests that a message be generated that does not include the operation element. The chat system described in Appendix 1.

[0109] (Note 12) If the control unit determines that the operation element should be included in the message, it requests the generation of operation element identification information that identifies the operation element. A message containing the operation element is generated based on the operation element identification information. The chat system described in Appendix 1.

[0110] (Note 13) A chat device configured to communicate with a terminal, Equipped with a control unit, The control unit, The response to the message generated by the machine learning model is obtained from the terminal. The response is input to the machine learning model, the machine learning model is instructed to determine whether or not to include the operation element performed on the terminal in the message, and to generate a message according to the determination result. The message generated by the machine learning model is displayed on the terminal. Chat device.

[0111] (Note 14) A chat method performed by a chat system configured to communicate with a terminal, Obtaining the response to the message generated by the machine learning model from the terminal, The above response is input to the machine learning model, the machine learning model is instructed to determine whether or not to include the operation element performed on the terminal in the message, and to generate a message according to the determination result, This includes displaying the message generated by the machine learning model on the terminal, How to chat.

[0112] (Note 15) A program to be executed on a chat system configured to communicate with a terminal, Obtaining the response to the message generated by the machine learning model from the terminal, The above response is input to the machine learning model, the machine learning model is instructed to determine whether or not to include the operation element performed on the terminal in the message, and to generate a message according to the determination result, This includes displaying the message generated by the machine learning model on the terminal, program.

[0113] The embodiments disclosed herein should be considered in all respects to be illustrative and not restrictive. The scope of the present invention is indicated by the claims, not in the sense described above, and is intended to include all modifications in the sense and scope equivalent to the claims. Furthermore, the present invention is not limited to the embodiments described above, and various modifications are possible within the scope of the claims, and embodiments obtained by appropriately combining the technical means disclosed in different embodiments are also included in the technical scope of the present invention. [Explanation of Symbols]

[0114] 1: Chat device 2: Terminal device 3: Language Model Device 12: Storage part 13: Control Unit 24: Display 1000: Chat System

Claims

1. A chat system configured to communicate with a terminal, It comprises one or more control units, The control unit, The response to the message generated by the machine learning model is obtained from the terminal. Based on the response, a prompt is input to the machine learning model for generating the next message to be sent to the terminal, causing the machine learning model to determine whether or not to include the operation element performed on the terminal in the message. If it determines that the operation element should be included in the message, it generates a message containing the operation element. If it is determined that it does not contain the element, it requests that a message be generated that does not contain the operation element. The message generated by the machine learning model is displayed on the terminal. Chat system.

2. The aforementioned message includes multiple of the aforementioned operation elements, The chat system according to claim 1.

3. The system further includes a storage unit that stores electronic information relating to services provided to the user, which is displayed on the terminal that accesses the chat system. The control unit requests the machine learning model to generate a message based on the content of the response and the electronic information. The terminal displays a message containing at least a portion of the electronic information generated by the machine learning model. The chat system according to claim 1.

4. The aforementioned operating element is a link that allows the terminal to access a predetermined resource. The chat system according to claim 1.

5. The control unit, in response to an operation on the link, causes the terminal to display a calendar for specifying a date. The chat system according to claim 4.

6. The aforementioned operating element is an input form that accepts input from the terminal. The chat system according to claim 1.

7. The control unit accepts the input of an email address from the terminal to the input form. The chat system according to claim 6.

8. The aforementioned operating element is a settlement element that initiates a transaction with the user of the terminal. The chat system according to claim 1.

9. The control unit displays the electronic information, the response, and the message generated by the machine learning model on the same screen of the terminal. The chat system according to claim 3.

10. The aforementioned operating elements are set by the user. The chat system according to claim 1.

11. If the control unit determines that the operation element should be included in the message, it requests the generation of operation element identification information that identifies the operation element. A message containing the operation element is generated based on the operation element identification information. The chat system according to claim 1.

12. A chat device configured to communicate with a terminal, Equipped with a control unit, The control unit, The response to the message generated by the machine learning model is obtained from the terminal. Based on the response, a prompt is input to the machine learning model for generating the next message to be sent to the terminal, causing the machine learning model to determine whether or not to include the operation element performed on the terminal in the message. If it determines that the operation element should be included in the message, it generates a message containing the operation element. If it is determined that it does not contain the element, it requests that a message be generated that does not contain the operation element. The message generated by the machine learning model is displayed on the terminal. Chat device.

13. A chat method performed by a chat system configured to communicate with a terminal, Obtaining the response to the message generated by the machine learning model from the terminal, Based on the response, a prompt is input to the machine learning model for generating the next message to be sent to the terminal, causing the machine learning model to determine whether or not to include the operation element performed on the terminal in the message. If it determines that the operation element should be included in the message, it generates a message containing the operation element. If it is determined that it does not contain the element, it will be requested to generate a message that does not contain the operation element. This includes displaying the message generated by the machine learning model on the terminal, How to chat.

14. A program to be executed on a chat system configured to communicate with a terminal, Obtaining the response to the message generated by the machine learning model from the terminal, Based on the response, a prompt is input to the machine learning model for generating the next message to be sent to the terminal, causing the machine learning model to determine whether or not to include the operation element performed on the terminal in the message. If it determines that the operation element should be included in the message, it generates a message containing the operation element. If it is determined that it does not contain the element, it will be requested to generate a message that does not contain the operation element. This includes displaying the message generated by the machine learning model on the terminal, program.

Images