Artificial intelligence system, program for artificial intelligence system

Abstract

This system provides an artificial intelligence (AI) system that allows users to specify a target in their browser and easily interact with a pre-trained AI tool with the desired content within that browser. [Solution] The artificial intelligence system 100 comprises a user terminal 110, a first server 120, and a second server 130. When a predetermined operation occurs while at least one of text TX, an image, and a file FL is displayed or specified in the browser 111 of the user terminal 110, the browser 111 displays a predetermined area panel SP that can display the interaction with the pre-trained artificial intelligence means 131 of the second server 130 via the first server 120.

Description

Technical Field

[0001] The present invention relates to an artificial intelligence system including a user terminal and a server, where the user terminal performs display or audio output based on the server, and a program thereof.

Background Art

[0002] Conventionally, a chatbot has a content inquiry step of asking the chat user for the content that the chat user wants to be reminded of on the chat screen of the chat user's terminal, a content registration step of registering the chat data received from the chat user's terminal after the content inquiry step as reminder content in a database, a date and time inquiry step of asking the chat user for the date and time when the chat user wants to notify the content on the chat screen of the chat user's terminal, a date and time registration step of registering the chat data received from the chat user's terminal after the date and time inquiry step as a reminder date and time in the database, a reminder date and time determination step of accessing the date and time control unit of the chat server to determine whether the current date and time has reached the registered reminder date and time, and a transmission notification step of transmitting the reminder content registered in the database to the chat user's terminal when it is determined that the current date and time has reached the registered reminder date and time. A program of a chat system having these steps is known (for example, Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, the conventional chat system programs described above were structured such that users could only send text data by typing on a keyboard, and notifications from the chatbot would be displayed as text data. This presented a problem: it was difficult for users to specify a target in their browser and easily interact with a pre-trained artificial intelligence tool with the desired content within that browser.

[0005] Therefore, the present invention solves the problems of the prior art described above, and that is, the object of the present invention is to provide an artificial intelligence system and program that allows a user to specify an object in a browser and easily interact with a pre-trained artificial intelligence means with desired content in that browser. [Means for solving the problem]

[0006] The invention according to claim 1 is an artificial intelligence system comprising a user terminal and a server, wherein the user terminal displays or outputs audio based on the server, and when a predetermined operation occurs while at least one of text, images, and files is displayed or specified in the browser of the user terminal, the browser displays a predetermined area panel that can display the interaction with the pre-trained artificial intelligence means of the server, thereby solving the aforementioned problems.

[0007] The invention according to claim 2 further solves the aforementioned problems by having, in addition to the configuration of the artificial intelligence system described in claim 1, when an icon button provided on the browser is operated as a predetermined operation, the browser displays a plurality of item menu buttons, and when one of the plurality of item menu buttons is selected, the browser displays a predetermined area panel on which the interaction with the pre-trained artificial intelligence means of the server can be displayed, and the predetermined area panel has a button for the content of one of the selected plurality of item menu buttons.

[0008] The invention according to claim 3 further solves the aforementioned problems by having, in addition to the configuration of the artificial intelligence system described in claim 2, the predetermined area panel having a text / image information display field designated as a resource in the browser, a prompt setting / input field for instructions and commands to the pre-trained artificial intelligence means, and an execution button for the contents of the item.

[0009] The invention according to claim 4, in addition to the configuration of the artificial intelligence system described in claim 2 or claim 3, is that the content of one of the items in the plurality of item menu buttons is "meeting time scheduling", and when the item menu button for "meeting time scheduling" is operated, the browser displays the received message text information specified in the browser as a resource in the text information display field of the predetermined area panel, and displays an execution button for "meeting time scheduling" on the predetermined area panel, and when the execution button is operated, the browser accesses the schedule management system server based on pre-set user information to obtain schedule information for the user of the user terminal, and uses the server's pre-trained artificial intelligence means to obtain scheduled blanks in the schedule information The system further solves the aforementioned problems by having the following configuration: Based on the information, multiple candidate meeting dates and times are selected and displayed in a predetermined area panel, along with a button to create a reply message; when the button to create a reply message is operated, the user terminal sends the received message text information, the multiple candidate meeting dates and times, and a prompt to create a reply message to the server's pre-trained artificial intelligence means; the browser displays the reply message draft information generated by the pre-trained artificial intelligence means in a predetermined area panel, along with a button to reflect it in the message creation screen / field; and when the button to reflect it in the message creation screen / field is operated, the browser displays the reply message draft information in the reply message window / field.

[0010] The invention according to claim 5 further solves the aforementioned problems by having a configuration that, in addition to the configuration of the artificial intelligence system described in claim 4, displays a button indicating that the reply message draft information generated by the pre-trained artificial intelligence means will be reflected in the message creation screen / field, and when the user selects to tentatively book an appointment in the user's schedule, the button indicating that it will be reflected in the message creation screen / field is operated, and the user terminal transmits information indicating that multiple candidate meeting dates and times have been tentatively booked to the schedule management system server.

[0011] The invention according to claim 6, in addition to the configuration of the artificial intelligence system described in claim 2 or claim 3, is that the browser displays the user's schedule information for today based on the user's schedule information on the schedule management system server, and the content of one of the multiple item menu buttons is "Create a daily report", and when the item menu button for "Create a daily report" is operated, the browser displays the schedule information for today, which is specified in the browser as a resource, in the text information display field of a predetermined area panel, and also displays an execution button for "Create a daily report", and when the execution button is operated, the user terminal displays the schedule information for today, Furthermore, the system sends a prompt to the server's pre-trained artificial intelligence means to create a daily report, the browser displays the daily report draft information generated by the pre-trained artificial intelligence means in a predetermined area panel, and also displays a button to reflect it in the message creation screen / field or a daily report file creation button in the predetermined area panel, and when the button to reflect it in the message creation screen / field is operated, the browser displays the daily report draft information in the message window / field, or when the daily report file creation button is operated, the browser outputs a file of the daily report draft information to a predetermined location, thereby further solving the aforementioned problems.

[0012] The invention according to claim 7, in addition to the configuration of the artificial intelligence system described in claim 2 or claim 3, is that the browser displays the user's weekly schedule information based on the user's schedule information on the schedule management system server, and the content of one of the multiple item menu buttons is "Create a weekly report", and when the item menu button for "Create a weekly report" is operated, the browser displays the weekly schedule information specified in the browser as a resource in the text information display field of a predetermined area panel, and also displays an execution button for "Create a weekly report", and when the execution button is operated, the user terminal displays the weekly schedule The system further solves the aforementioned problems by sending information and a prompt to create a weekly report to the server's pre-trained artificial intelligence means, the browser displays the draft weekly report information generated by the pre-trained artificial intelligence means in a predetermined area panel, and also displays a button to reflect it in the message creation screen / field or a weekly report file creation button in the predetermined area panel, and when the button to reflect it in the message creation screen / field is operated, the browser displays the draft weekly report information in the message window / field, and when the weekly report file creation button is operated, the browser outputs a file of the draft weekly report information to a predetermined location.

[0013] The invention according to claim 8, in addition to the configuration of the artificial intelligence system described in claim 2 or claim 3, includes the following: the browser displays the user's schedule information for a predetermined period based on the user's schedule information on the schedule management system server; the content of one of the multiple item menu buttons is "to schedule a meeting"; when the item menu button for "to schedule a meeting" is operated, the browser displays an execution button for "to schedule a meeting" on a predetermined area panel; and when the execution button is operated, the user terminal accesses the schedule management system server based on pre-set user information to obtain schedule information for the user of the user terminal, and uses the server's pre-trained artificial intelligence means to process the schedule information based on the schedule gap information. The system further solves the aforementioned problems by having the user terminal select multiple candidate meeting dates and times, displaying the multiple candidate meeting dates and times in a predetermined area panel, and displaying a message creation button in the predetermined area panel, and when the message creation button is operated, sending the multiple candidate meeting dates and times and a prompt to create a message for the purpose of meeting arrangement to the server's pre-trained artificial intelligence means, and having the browser display the message draft information generated by the pre-trained artificial intelligence means in a predetermined area panel, and displaying a button in the predetermined area panel to reflect it in the message creation screen / field, and when the button to reflect it in the message creation screen / field is operated, the browser displays the message draft information in the message window / field.

[0014] The invention according to claim 9, in addition to the configuration of the artificial intelligence system described in claim 2 or claim 3, when one or more files are specified in the browser and an icon button provided in the browser is operated, the server extracts language character information data as an electronic information source from the specified one or more files and divides it into multiple divided language character file data, converts each of the divided divided language character file data into vector coordinate system divided language data using a vector coordinate transformation means which is a pre-trained artificial intelligence means, and stores the converted vector coordinate system divided language data in vector indexes in association with the original divided language character file data, and the browser displays multiple item menu buttons, and one of the multiple item menu buttons The content of one item is either to summarize or to ask a question, and when the menu button for the item to summarize or to ask a question is operated, the browser displays a text information display field designated as a resource in the browser, a prompt setting / input field which is an instruction command to the pre-trained artificial intelligence means, and an execution button for the content of the item to summarize or to ask a question on a predetermined area panel, and when the execution button is operated, the user terminal sends prompt information as an instruction command to the server, the server generates text data using the pre-trained artificial intelligence means and sends the generated text data to the user terminal, and the browser of the user terminal displays the received generated text data on a predetermined area panel, thereby further solving the aforementioned problems.

[0015] The invention according to claim 10 further solves the aforementioned problems by having a configuration that, in addition to the configuration of the artificial intelligence system described in claim 2 or claim 3, the browser displays a talk screen for message exchanges between a user and other users, and when an icon button provided on the browser is operated while one or more message text groups are specified in the browser, the browser displays a plurality of item menu buttons, and when one of the plurality of item menu buttons is selected, the browser displays a predetermined area panel that can display the exchange with the server's pre-trained artificial intelligence means, and the predetermined area panel has a button for the content of one of the selected plurality of item menu buttons.

[0016] The invention according to claim 11 further solves the aforementioned problems by having the following configuration in addition to the configuration of the artificial intelligence system described in claim 2 or claim 3: the browser displays a file in a cell display format having multiple cells, and when an icon button provided in the browser is operated while a sheet having multiple cells is specified in the browser, the browser displays multiple item menu buttons, the content of one of the multiple item menu buttons is to create a format, and when the item menu button for creating a format is operated, the browser displays a prompt setting / input field which is an instruction command to the pre-trained artificial intelligence means, and an execution button for the content of the item on a predetermined area panel, and when the execution button is operated, the user terminal sends prompt information from the prompt setting / input field to the server, and the browser displays the generated text data generated by executing text generation in a tabular format using the pre-trained artificial intelligence means on a predetermined area panel and displays a button on a predetermined area panel indicating that it should be reflected on the sheet, and when the button indicating that it should be reflected on the sheet is operated, the browser displays the generated text data in tabular format on the sheet.

[0017] The invention according to claim 12, in addition to the configuration of the artificial intelligence system described in claim 2 or claim 3, is as follows: The browser displays a file in cell display format having text data in multiple cells, and when an icon button provided in the browser is operated while a sheet having multiple cells is specified in the browser, the browser displays multiple item menu buttons, and the content of one of the item menu buttons is "to analyze", and when the item menu button for "to analyze" is operated, the browser displays the text data in the file in cell display format specified in the browser as a resource in the text information display field of a predetermined area panel, and displays an execution button for the content of the item in the predetermined area panel, and when the execution button is operated, the user terminal displays the text data of multiple cells and the "to analyze" prompt The system further solves the aforementioned problems by sending input information to a server, which uses pre-trained artificial intelligence means to extract and generate grouping information based on the content of the text data of the multiple cells it receives, and then grouping the data according to the grouping information, sending the grouping information, the number of items in each group after grouping, and the correspondence information with the cell numbers that contained the text data before grouping to the user terminal, and the browser displaying the grouping information, the number of items in each group after grouping, and the correspondence information with the cell numbers that contained the text data before grouping on a predetermined area panel, and displaying a button to reflect it on the sheet, and when the button to reflect it on the sheet is operated, the browser displays the generated text data in tabular format on the sheet.

[0018] The invention according to claim 13 further solves the aforementioned problems by having a configuration that, in addition to the configuration of the artificial intelligence system described in claim 2 or claim 3, the browser displays a customer relationship management screen for customer information and interactions with customers, and when an icon button provided on the browser is operated while customer information and interaction information are specified in the browser, the browser displays a plurality of item menu buttons, and when one of the plurality of item menu buttons is selected, the browser displays a predetermined area panel on which interactions with the server's pre-trained artificial intelligence means can be displayed, and the predetermined area panel has a button for the content of one of the selected plurality of item menu buttons.

[0019] The invention according to claim 14 is a program for an artificial intelligence system that a user terminal displays or outputs audio based on a server, and solves the above-mentioned problems by causing a computer to execute the following steps: an icon operation determination step which determines whether an icon button provided in the browser has been operated when at least one of text, images, and files is displayed and specified in the browser of the user terminal; an item menu selection operation determination step which determines whether one of the item menu buttons has been selected when the icon button has been operated and the browser displays a plurality of item menu buttons; and a panel display step which determines whether one of the item menu buttons has been selected when one of the item menu buttons has been selected and the browser displays a predetermined area panel which is capable of displaying the interaction with the server's pre-trained artificial intelligence means and has a button for the content of one of the selected item menu buttons. [Effects of the Invention]

[0020] The artificial intelligence system of the present invention, comprising a user terminal and a server, not only allows the user terminal to display or output audio based on the server, but also achieves the following unique effects.

[0021] According to the artificial intelligence system of the invention according to claim 1, when an object is displayed, specified, and subjected to a predetermined operation in a browser, a predetermined area panel is displayed and can communicate freely with the pre-trained artificial intelligence means. Therefore, the user can easily communicate with the pre-trained artificial intelligence means with the desired content in the predetermined area panel with the content displayed and specified in the browser as the object. That is, while displaying text or the like in one browser, it is possible to easily communicate with the pre-trained artificial intelligence means with the desired content in the predetermined area panel with the displayed content as the object.

[0022] According to the artificial intelligence system of the invention according to claim 2, in addition to the effects achieved by the invention according to claim 1, when an object is displayed, specified, an icon button is operated, and an item menu button is selected and operated in a browser, a predetermined area panel having buttons for the content of that item is displayed and can communicate freely with the pre-trained artificial intelligence means. Therefore, the user can select a desired item menu in the browser and easily communicate with the pre-trained artificial intelligence means with the content of the desired item in the predetermined area panel.

[0023] According to the artificial intelligence system of the invention according to claim 3, in addition to the effects achieved by the invention according to claim 2, when, for example, summary creation or reply document creation is selected as an item menu, resources are displayed, and it is possible to freely set the length specification of the summary text, the specification of including a greeting as the content of the reply document, etc. Therefore, the user can easily perform summary creation or reply document creation by communicating with the pre-trained artificial intelligence means.

[0024] According to the artificial intelligence system of the invention according to claim 4, in addition to the effects achieved by the invention according to claim 2 or claim 3, when the user operates the execution button, a plurality of meeting candidate date and time information is picked up by the pre-trained artificial intelligence means and displayed in the predetermined area panel. Therefore, the user can save the trouble of checking the schedule information, which is so-called calendar information, in the schedule management system. Furthermore, when the user operates the reply message creation button, reply message scenario information reflecting a plurality of meeting candidate date and time information by the pre-trained artificial intelligence means is displayed in the predetermined area panel, so the user can easily prepare reply message scenario information for schedule adjustment using the pre-trained artificial intelligence means. Also, when the user operates the button for reflecting in the message creation screen / field, the reply message scenario information is set in the reply message window / field, so the user can easily reach the state just before sending by simply operating the send button.

[0025] According to the artificial intelligence system of the invention according to claim 5, in addition to the effects achieved by the invention according to claim 4, since the intention of reservation is registered at a plurality of meeting candidate date and times of the user's schedule information, it is possible to avoid the situation where other schedules are mistakenly entered at a plurality of meeting candidate date and times during the schedule adjustment stage. Furthermore, even when the execution button for adjusting the meeting time for another matter is operated, the provisional reservation date and time that has already been tentatively entered will not be listed as the meeting candidate date and time for the other matter, so it is possible to avoid the situation where provisional reservations overlap.

[0026] According to the artificial intelligence system of the invention according to claim 6, in addition to the effects achieved by the invention according to claim 2 or claim 3, when the user operates the item menu button for creating a daily report and then the execution button while the browser is displaying the schedule information for today, the daily report scenario information is generated and displayed by the pre-trained artificial intelligence means, so the user can easily prepare the daily report draft. Furthermore, when the button for reflecting in the message creation screen / field or the daily report file creation button is operated, the state just before sending the daily report message is reached, or the daily report file is output to a predetermined location, so the user can easily submit the daily report.

[0027] According to the artificial intelligence system of the invention of claim 7, in addition to the effects of the invention of claim 2 or claim 3, when the user is viewing weekly schedule information in a browser and operates the item menu button for creating a weekly report and then operates the execute button, the pre-trained artificial intelligence means generates and displays draft weekly report information, so the user can easily prepare a draft weekly report. Furthermore, by clicking the button to reflect the message in the message creation screen / field, or the button to create the weekly report file, the system returns to the state just before sending the weekly report message, or the weekly report file is output to a designated location, allowing users to easily submit their weekly reports.

[0028] According to the artificial intelligence system of the invention of claim 8, in addition to the effects achieved by the invention of claim 2 or claim 3, when the user operates the execution button, multiple candidate meeting date and time information is picked up by the pre-trained artificial intelligence means and displayed on a predetermined area panel, so the user can avoid the trouble of checking the details of schedule information, which is so-called calendar information, in the schedule management system. Furthermore, when a user presses the message creation button, a pre-trained artificial intelligence system displays a draft message reflecting multiple possible meeting dates and times on a designated area panel. This allows the user to easily prepare a draft message for scheduling using the pre-trained artificial intelligence system. Furthermore, when a user presses a button to reflect the message in the message creation screen / field, the message draft information is set in the message window / field, making it easy for the user to reach a state just before sending, where all they have to do is press the send button.

[0029] According to the artificial intelligence system of claim 9, in addition to the effects of the invention of claim 2 or claim 3, for example, by selecting one or more files on a drive such as a cloud computer or a local computer and operating an icon button, language text data is associated as vector data and saved in a vector index as so-called RAG (Retrieval Augmented Generation) training data, which is an external information source for the artificial intelligence system. Therefore, users can easily create RAG (Retrieval Augmented Generation) training data. Furthermore, when a user operates an execution button to summarize or ask a question, the pre-trained artificial intelligence generates text summarizing or answering the question based on the scope of the RAG training data. This generated text is then displayed in a designated area panel as the summary or answer to the question, allowing the user to easily obtain a summary or answer to a question within the scope of the RAG training data.

[0030] According to the artificial intelligence system of the invention of claim 10, in addition to the effects of the invention of claim 2 or claim 3, when a target is specified in the browser displaying the talk screen, the icon button is operated and the item menu button is selected, a predetermined area panel having buttons for the content of that item is displayed, and interaction with the pre-trained artificial intelligence means becomes possible. Therefore, the user can select the desired item menu in the browser and easily interact with the pre-trained artificial intelligence means with the content of the desired item in the predetermined area panel.

[0031] According to the artificial intelligence system of the invention of claim 11, in addition to the effects of the invention of claim 2 or claim 3, when a cell display file is displayed in a browser, if the user operates the icon button and the item menu button for creating a format, enters the desired prompt information in the prompt setting / input field, and operates the execute button, the generated text data in a format that matches the prompt information is generated by the pre-trained artificial intelligence means and then displayed in a predetermined area panel, so that the user can easily prepare the generated text data in a format as a format proposal. Furthermore, when a user clicks a button to apply the changes to the sheet, the generated text data in tabular format is reflected in the sheet of the cell-formatted file, making it easy for users to create the format for cell-formatted files. For example, you can easily create formats for documents such as quotations and invoices.

[0032] According to the artificial intelligence system of the invention of claim 12, in addition to the effects of the invention of claim 2 or claim 3, when the user operates an icon button, an item menu button indicating analysis, and an execution button, the text data of multiple cells is grouped based on its content, and information on the content of the grouping, the number of items in each group after grouping, and the correspondence information with the cell numbers that contained the text data before grouping are displayed on a predetermined area panel. As a result, the user can easily prepare data for the grouped analysis results of the text data of multiple cells. Furthermore, when the user clicks a button to apply the results to the sheet, the grouped analysis results are reflected in a sheet of a file in cell format, allowing for the analysis of text data by reflecting the analysis results in a file in cell format. For example, if employee survey comments are entered into multiple cells as an example of text data, and the user operates the icon button, the menu button for analysis, and the execute button, the employee survey comments will be grouped into categories based on their content. The system will then display information about the grouping, the number of comments in each group, and the correspondence between the cell numbers containing the original text data and the corresponding employee survey comments. This allows users to easily group and analyze the content of employee survey comments.

[0033] According to the artificial intelligence system of the invention of claim 13, in addition to the effects of the invention of claim 2 or claim 3, when a target is specified in the browser displaying the customer relationship management screen, the icon button is operated and the item menu button is selected, a predetermined area panel having buttons for the content of that item is displayed, and interaction with the pre-trained artificial intelligence means becomes possible. Therefore, the user can select the desired item menu in the browser and easily interact with the pre-trained artificial intelligence means with the content of the desired item in the predetermined area panel.

[0034] According to the program of the artificial intelligence system of the invention of claim 14, similar to the effect achieved by the invention of claim 2, when an object is specified in the browser, an icon button is operated, and an item menu button is selected, a predetermined area panel having buttons for the content of that item is displayed, and interaction with the pre-trained artificial intelligence means becomes possible. Therefore, the user can select a desired item menu in the browser and easily interact with the pre-trained artificial intelligence means with the content of the desired item in the predetermined area panel. In other words, while displaying text or other content in a single browser, it is possible to easily interact with a pre-trained artificial intelligence system using a designated area panel to specify desired content based on the displayed content. [Brief explanation of the drawing]

[0035] [Figure 1] A diagram illustrating the concept of an artificial intelligence system, which is the first embodiment of the present invention. [Figure 2] A chart illustrating an example of the operation of an artificial intelligence system, which is the first embodiment of the present invention. [Figure 3] (A) and (B) are diagrams showing how a reply email is created in the browser of a user terminal of an artificial intelligence system, which is the first embodiment of the present invention. [Figure 4] (A) and (B) are diagrams showing how a summary of an incoming email is created in the browser of a user terminal of an artificial intelligence system, which is the first embodiment of the present invention. [Figure 5] (A) and (B) are diagrams showing how the browser of a user terminal of an artificial intelligence system, which is the first embodiment of the present invention, detects problems with received emails. [Figure 6] (A) and (B) are diagrams showing how the browser of a user terminal of an artificial intelligence system, which is the first embodiment of the present invention, adjusts the sender and meeting time for an incoming email. [Figure 7] (A) and (B) are diagrams showing the registration of a provisional reservation in a calendar / schedule management system when coordinating meeting times with the sender of an incoming email in the browser of a user terminal of an artificial intelligence system, which is the first embodiment of the present invention. [Figure 8] (A) and (B) are diagrams showing how a summary of the entire thread related to an received email is created in the browser of a user terminal of an artificial intelligence system, which is the first embodiment of the present invention. [Figure 9] (A) and (B) are diagrams showing how a daily report is created from calendar and schedule information in the browser of a user terminal of an artificial intelligence system, which is a second embodiment of the present invention. [Figure 10] (A) and (B) are diagrams showing how a second embodiment of the present invention, an artificial intelligence system, allows users to specify a meeting time by selecting a partner from calendar and schedule information in the browser of their user terminal. [Figure 11](A) and (B) are diagrams showing how RAG learning is performed by specifying a file from drive information in the browser of a user terminal of an artificial intelligence system, which is a third embodiment of the present invention. [Figure 12] (A) and (B) are diagrams showing how a reply message is created in the browser of a user terminal of an artificial intelligence system, which is a fourth embodiment of the present invention, in response to a message on the chat screen. [Figure 13] (A) and (B) are diagrams showing how a format is created in a cell display file in the browser of a user terminal of an artificial intelligence system, which is a fifth embodiment of the present invention. [Figure 14] (A) and (B) are diagrams showing how text data in a cell-display format file is analyzed in the browser of a user terminal of an artificial intelligence system, which is a fifth embodiment of the present invention. [Figure 15] (A) and (B) are diagrams showing how an artificial intelligence system, which is a sixth embodiment of the present invention, generates an email body from customer relationship management information in a browser on a user terminal, in cooperation with a customer relationship management system, which is an example of the company's own system. [Figure 16] (A) and (B) are figures showing the selection of an image range in the browser of a user terminal of an artificial intelligence system, which is a seventh embodiment of the present invention, and the situation when a question is asked about the content of the image. [Figure 17] (A) and (B) are figures showing the process of reading the text of a web page displayed in the browser of a user terminal of an artificial intelligence system, which is a seventh embodiment of the present invention, and the process of asking a question about the content of the web page. [Figure 18] (A) and (B) are figures showing the process of reading the text of a selected area on a web page displayed in the browser of a user terminal of an artificial intelligence system, which is a seventh embodiment of the present invention, and the process of asking a question about the content of the selected area's text. [Modes for carrying out the invention]

[0036] The artificial intelligence system of the present invention comprises a user terminal and a server, and when a predetermined operation is performed while at least one of text, images, and files is displayed or specified in the browser of the user terminal, the browser is configured to display a predetermined area panel on which interaction with the server's pre-trained artificial intelligence means can be displayed. In this configuration, the user can easily interact with the pre-trained artificial intelligence means with desired content in the predetermined area panel, targeting the content displayed or specified in the browser. Any specific embodiment of the system is acceptable as long as it is configured in this way. Furthermore, the program for the artificial intelligence system of the present invention may have any specific embodiment as long as it allows the user to easily select a desired item menu in the browser and interact with the pre-trained artificial intelligence means with the content of the desired item in the predetermined area panel, by causing the computer to execute the following steps: an icon operation determination step which determines whether an icon button provided in the browser has been operated when at least one of text, images, and files is displayed and specified in the browser of the user terminal's browser; an item menu selection operation determination step which determines whether one of the item menu buttons has been selected when the icon button has been operated and the browser displays a plurality of item menu buttons; and a panel display step which determines whether one of the item menu buttons has been selected when one of the item menu buttons has been selected and the browser displays a predetermined area panel which displays the interaction with the pre-trained artificial intelligence means of the server and has a button for the content of one of the selected item menu buttons.

[0037] For example, a user terminal can be any device that displays data on the user terminal's display browser, such as a desktop personal computer terminal, a notebook personal computer terminal, a smartphone terminal, a tablet terminal, a smartwatch terminal, or a glasses terminal, and that connects to a server via a communication network including a wide area network (such as the Internet), a local network, or a telephone line to send and receive data. Furthermore, the server may be a cloud server created in a cloud environment, and the number of physical servers that make up the server may be one or more. Furthermore, the pre-trained artificial intelligence means (text generation means) is a text generation means that generates text data and is composed of an interactive model called a large-scale language model, such as ChatGPT (Generative Pre-trained Transformer) (hereinafter referred to as ChatGPT), and may be configured on a single server or on multiple servers in the cloud. Furthermore, in this context, "cell display format file" also refers to a spreadsheet file, which is a so-called row and column cell (mass) format in which the input cell is specified by the row and column number and symbol. Furthermore, a key feature of "cell display format files" is that the width and height of cells can be freely changed on a row-by-row basis. Furthermore, the layout of text and numbers within cells can be freely configured on a cell-by-cell basis. A "cell display format file" can refer not only to a sheet, which is a collection of cells, but also to a file containing multiple sheets. [Examples]

[0038] The first embodiment of the present invention, an artificial intelligence system 100, will be described below with reference to Figures 1 to 8(B). Here, Figure 1 is a diagram showing the concept of the artificial intelligence system 100, which is the first embodiment of the present invention; Figure 2 is a chart showing an example of operation of the artificial intelligence system 100, which is the first embodiment of the present invention; Figure 3(A) shows the process of performing a predetermined operation to display the side panel SP while the received message text information ML is displayed in the browser 111 of the user terminal 110 of the artificial intelligence system 100, which is the first embodiment of the present invention; Figure 3(B) shows the process of creating a reply email using the side panel SP of the browser 111; Figure 4(A) shows the process of performing a predetermined operation to display the side panel SP while the received message text information ML is displayed in the browser 111 of the user terminal 110 of the artificial intelligence system 100, which is the first embodiment of the present invention; Figure 4(B) shows the process of creating a summary using the side panel SP of the browser 111; Figure 5(A) shows the process of performing a predetermined operation to display the side panel SP while the received message text information ML is displayed in the browser 111 of the user terminal 110 of the artificial intelligence system 100, which is the first embodiment of the present invention; Figure 5 Figure 6(A) shows how problems are discovered in the received message text information ML using the side panel SP of browser 111, which is the first embodiment of the artificial intelligence system 100 of the present invention, with the received message text information ML displayed in the browser 111 of the user terminal 110, and a predetermined operation is performed to display the side panel SP. Figure 6(B) shows how the sender of the received message text information ML and the meeting time are coordinated using the side panel SP of browser 111, which is the first embodiment of the artificial intelligence system 100 of the present invention, with the meeting time coordinated using the provisional reservation of the calendar / schedule management system. Figure 7(B) shows how the browser 111 has made a provisional reservation of the calendar / schedule information SD. Figure 8(A) shows how the side panel SP is displayed in the browser 111 of the user terminal 110 of the artificial intelligence system 100 of the present invention, with the entire thread related to the received message text information ML displayed, and a predetermined operation is performed to display the side panel SP.Figure 8(B) shows how to create a summary of the entire thread using the side panel SP of browser 111.

[0039] As shown in Figure 1, the first embodiment of the present invention, artificial intelligence system 100, comprises a user terminal 110 and a first server 120 and a second server 130 as servers. Of these, the first server 120 has a bot that also functions as a display program. On the other hand, the second server 130 has ChatGPT, which is an example of a pre-trained artificial intelligence means 131. Furthermore, from a technical standpoint, the first server 120 and the second server 130 may have a configuration that is physically common to each other, or they may have a configuration that is physically different. Furthermore, the user terminal 110 is configured to communicate freely with ChatGPT, which is an example of a pre-trained artificial intelligence means 131 of the second server 130, via the first server 120.

[0040] Furthermore, as a top-level concept of the technical idea, when a predetermined operation occurs while a text TX or file FL is displayed and specified in the browser 111 of the user terminal 110, the browser 111 is configured to display a side panel SP, which is an example of a predetermined area panel that can display interactions with Chat GPT, an example of a pre-trained artificial intelligence means 131 of the second server 130, via the first server 120.

[0041] More specifically, when at least one of text TX, an image, and a file FL is displayed or specified in the browser 111 of the user terminal 110, and the first icon button 111a, which is an example of an icon button provided in the browser 111, is operated as an example of a predetermined operation to display a side panel SP, which is an example of a predetermined area panel, the browser 111 displays a plurality of item menu buttons. Furthermore, as an example of a predetermined operation to display the side panel SP, when one of several item menu buttons is selected, the browser 111 displays the side panel SP as an example of a predetermined area panel that can display interactions with Chat GPT, which is an example of a pre-trained artificial intelligence means 131 of the second server 130, via the first server 120. For the purposes of this explanation, we will use the side panel SP as an example of a predetermined area panel. However, in terms of technical concept, the location of the predetermined area panel in browser 111 does not necessarily have to be on the side relative to the center; it may be above or below. In this configuration, the side panel SP has a button that contains the content of one of the selected items from a menu of multiple items.

[0042] As a result, when a target is specified in the browser 111 and the first icon button 111a is operated to select an item menu button, a side panel SP with buttons for the content of that item is displayed, allowing for seamless interaction with Chat GPT, which is an example of a pre-trained artificial intelligence means 131. As a result, users can easily interact with Chat GPT, an example of a pre-trained artificial intelligence means 131, by selecting a desired item menu in the browser 111 and using the content of the desired item in the side panel SP. In other words, while displaying text and other content in a single browser 111, it is possible to easily interact with Chat GPT, an example of a pre-trained artificial intelligence means 131, using the displayed content as the target of a side panel SP, which is an example of a predetermined area panel, with the desired content.

[0043] Next, we will explain in detail an example of how the artificial intelligence system 100 operates. As shown in Figure 2, in step S1, as an icon operation determination step (predetermined operation determination step), the browser 111 of the user terminal 110 determines whether the first icon button 111a, which is an example of an icon button provided in the browser 111, has been operated as an example of a predetermined operation, while at least one of text TX, an image, and a file FL is displayed and specified in the browser 111. More specifically, as shown in Figure 3(A), the browser 111 of the user terminal 110 displays the received message text information ML as an example of at least one of the following: text TX, image, and file FL. Here, it is sufficient that at least one of the text TX, image, and file FL is specified by the browser 111 displaying at least one of the text TX, image, and file FL. In other words, the method of specification is not limited to selecting at least one of the text TX, image, and file FL with the cursor; it is sufficient to define a range of at least one of the text TX, image, and file FL. Furthermore, the browser 111 has a first icon button 111a and a side panel icon button 111b. If it is determined that the first icon button 111a has been operated, the process proceeds to step S2; otherwise, step S1 is repeated.

[0044] In step S2, as shown in Figure 3(A), the browser 111 displays multiple item menu buttons as an item menu display step. More specifically, the browser 111 displays menu buttons such as "Create Reply Email Document" menu button 112a, "Summarize Email Content" menu button 112b, "Find Problems in Email" menu button 112c, and "Schedule Meeting" menu button 112d as examples of multiple item menu buttons.

[0045] In step S3, as an item menu selection operation determination step (predetermined operation determination step), the browser 111 determines whether or not one of the multiple item menu buttons has been selected as an example of a predetermined operation. More specifically, let's assume that the "Create Reply Email Document" menu button 112a is selected as an example of an item menu button. Thus, if it is determined that a selection operation has been performed, the process proceeds to step S4; on the other hand, if it is determined that there is no selection yet, step S3 is repeated.

[0046] In step S4, as shown in Figure 3(B), the browser 111 displays a side panel SP, which is an example of a predetermined area panel, as a panel display step. Here, the side panel SP is capable of displaying interactions with Chat GPT, which is an example of a pre-trained artificial intelligence means 131 of the second server 130, via the first server 120, and has a button for the content of one of the selected items in a plurality of item menu buttons. Then, in step S3, when the "Create Reply Email Document" menu button 112a is selected, the side panel SP is equipped with a "Create Email Body" button SP1a as a button corresponding to the content of the selected "Create Reply Email Document" menu button 112a.

[0047] To explain the side panel SP in more detail, the side panel SP includes a text information display field SP1b as an example of a text / image information display field, a prompt input field SP1c as a prompt setting / input field, an advanced settings button, a "Create email body" button SP1a, a generated text display field SP1f, a redo button SP1g, a copy button SP1h, a "Add to email creation screen" button SP1k which indicates that the message will be reflected in the message creation screen / field, and a navigation bar SP2. Of these, the text information display field SP1b receives the received message text information ML, which is an example of the text TX specified in the browser 111, as resource information, and is entered by the browser 111. Additionally, the prompt input field SP1c is equipped with a voice input button SP1d and a prompt wizard button SP1e. For example, when the voice input button SP1d is pressed by the user, the microphone on the user terminal 110 is turned on, enabling voice input. When the user voice-inputs "Create a thank-you email," the phrase "Create a thank-you email" is entered into the prompt input field SP1c. When the Prompt Wizard button SP1e is pressed by the user, the Prompt Wizard window opens, allowing for detailed configuration of the prompt settings. Additionally, when the user clicks the advanced settings button, the following options will be displayed: formatting settings, tone settings, length settings, your role settings, and language settings. In the format settings section, you can select one of the following purposes: email to customers, email to colleagues, email to business partners, report email, blog-style email, proposal email, or apology email. This allows for fine-tuning of wording and formatting. Additionally, the tone settings allow you to choose from four styles of expression: formal, gentle, businesslike, passionate, quiet, or empathetic. This allows the phrasing to be fine-tuned. Furthermore, the length setting allows you to select one of three lengths for the generated text: medium, short, or long. This allows the length to be tuned. Additionally, in the role setting section, you can select one of the following for your position: Unspecified, Leader, Businessman, Sales Member, Technical Member, Administrative Department Member, Planning Department Member, Company Executive, Teacher / Staff, or Field Worker. This allows for fine-tuning of the perspective from which comments are made. Furthermore, in the language settings, you can select one of the following languages ​​for the generated text: Japanese, English, Chinese, Vietnamese, or Korean. This allows for tuning to determine which language comments will be written in. Furthermore, it is desirable that the generated text display area SP1f and other elements below the "Create Email Body" button SP1a be displayed after the "Create Email Body" button SP1a has been pressed. The reason is that, in terms of the order of operations, the "Create Email Body" button SP1a should be pressed first, and only then should the subsequent information be displayed, making it easier for the user to understand.

[0048] In step S5, as a step to determine whether an execution button has been operated, the browser 111 determines whether the "Create email body" button SP1a, which is an example of a button representing the content of one of the selected item menu buttons, has been operated. If it is determined that the operation has been performed, proceed to step S6; otherwise, if it is determined that the operation has not yet been performed, repeat step S5.

[0049] In step S6, as a text generation step, the user terminal 110 sends prompt information to the first server 120, including the received message text information ML displayed in the text information display field SP1b as an example of resource information, and instructions to create an email. Then, the first server 120 generates text data using Chat GPT, which is an example of a pre-trained artificial intelligence means 131 of the second server 130, based on the received message text information ML and prompt information. In step S7, as a text display step, the first server 120 sends the generated text data GT obtained to the user terminal 110. Then, the browser 111 of the user terminal 110 displays the received generated text data GT in the generated text display field SP1f of the side panel SP.

[0050] As a result, as mentioned above, when a target is specified in the browser 111 and the first icon button 111a is operated to select an item menu button, a side panel SP with buttons for the content of that item is displayed, allowing for seamless interaction with Chat GPT, which is an example of a pre-trained artificial intelligence means 131. As a result, users can easily interact with Chat GPT, an example of a pre-trained artificial intelligence means 131, by selecting a desired item menu in the browser 111 and using the content of the desired item in the side panel SP. In other words, while displaying text and other content in a single browser 111, it is possible to easily interact with Chat GPT, an example of a pre-trained artificial intelligence means 131, using the displayed content as the target of a side panel SP, which is an example of a predetermined area panel, with the desired content.

[0051] Furthermore, in this embodiment, the side panel SP is provided with a re-execution button SP1g, a copy button SP1h, and an "Add to email creation screen" button SP1k, which indicates that the message will be reflected in the message creation screen / field. Of these, if the re-execution button SP1g is pressed, the action that occurred when the "Create email body" button SP1a was pressed will be performed again.

[0052] More specifically, steps S5 to S7 are executed to determine whether the execution button has been pressed. For example, if the user views the generated text data GT displayed for the first time in the generated text display field SP1f of the side panel SP and is not entirely satisfied, the user terminal 110 will then re-execute by pressing the re-execution button SP1g, causing the user terminal 110 to send the received message text information ML and prompt information, which are examples of resource information, to the first server 120 again. Then, the first server 120, based on the received message text information ML and prompt information, uses Chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, to generate text data again. In this case, a generated text data GT is created with slightly different expressions and content compared to the previous version. Then, the first server 120 sends the generated text data GT obtained to the user terminal 110. Then, the browser 111 of the user terminal 110 displays the received generated text data GT in the generated text display field SP1f of the side panel SP. In other words, the generated text data GT will be displayed with slightly different expressions and content compared to the previous version.

[0053] Furthermore, when the copy button SP1h is operated by the user, the generated text data GT displayed in the generated text display field SP1f on the side panel SP is stored in the memory unit of the user terminal 110, which is a temporary storage area also known as the clipboard. Users can paste the generated text data GT into emails or document files using the paste operation.

[0054] Furthermore, let's assume that the user has clicked the "Add to email composition screen" button SP1k, which is a button that indicates that the message will be reflected in the message composition screen / field. Then, as shown in Figure 3(B), the browser 111 displays a reply email window 113, which is an example of a reply message window, and reflects and displays the generated text data GT displayed in the generated text display field SP1f of the side panel SP in the reply email window 113. This sets the reply email data (generated text data GT) into the reply email window 113, which is an example of a reply message window, and puts the email in a state just before sending, where only the send button needs to be pressed. As a result, users can easily prepare reply emails.

[0055] Additionally, the side panel SP is equipped with a navigation bar SP2. The navigation bar SP2 includes, as an example, buttons for creating email body text (SP2a), creating email body summary text (SP2b), creating thread summary text (SP2c), identifying problems (SP2d), editing outgoing email document (SP2e), and scheduling meeting times (SP2f). When these buttons are pressed, the content of the side panel SP is configured to switch to reflect the content of the button that was pressed. In other words, when the first icon button 111a is operated and one item menu button is selected from among multiple item menu buttons, a side panel SP containing the corresponding button is displayed. After it is displayed, the contents of the side panel SP can be freely switched using the buttons on the navigation bar SP2. Furthermore, when the side panel icon button 111b provided in the browser 111 is operated, a side panel SP is displayed that has buttons with the same content as when the "Create Reply Email Document" menu button 112a is operated, as an example of the default settings. In other words, as an example of a predetermined operation to display the side panel SP, the side panel icon button 111b may be operated.

[0056] In this embodiment, the side panel SP, which is a predetermined area panel, has a text information display field SP1b specified as a resource in the browser 111, a prompt input field SP1c which is a prompt setting / input field for instructions to the chat GPT, which is an example of a pre-trained artificial intelligence means 131, and a "Create email body" button SP1a which is an example of an execution button for the contents of an item. This means that when you select items such as "Create Summary" or "Create Reply Document" from the menu, resources will be displayed, allowing you to freely configure settings such as specifying the length of the summary or whether to include a thank you message in the reply document. As a result, users can easily create summaries and reply documents by interacting with ChatGPT, which is an example of a pre-trained artificial intelligence tool 131.

[0057] As shown in Figure 4(A), when the first icon button 111a provided on the browser 111 is operated as an example of a predetermined operation to display the side panel SP, with the received message text information ML specified as an example of a text TX or file FL in the browser 111 of the user terminal 110, the browser 111 displays multiple item menu buttons. Furthermore, as an example of a predetermined operation to display the side panel SP, let's assume that the "Email Content Summary" menu button 112b is selected as one of several item menu buttons.

[0058] Then, as shown in Figure 4(B), the browser 111 displays the side panel SP. The side panel SP includes, for example, a text information display field SP1b, a length setting field SP1n which is a prompt setting / input field, a "Summarize Document" button SP1m which is an execution button for the content of the item, a generated text display field SP1f, a re-execution button SP1g, and a copy button SP1h. Furthermore, it is desirable that the generated text display area SP1f and other elements below the "Summarize Document" button SP1m be displayed after the "Summarize Document" button SP1m has been operated. The reason is that, in terms of the order of operations, the "Summarize Document" button SP1m should be operated first, and only then should the subsequent information be displayed, making it easier for the user to understand.

[0059] In the text information display area SP1b, when the side panel SP is displayed or when the display is switched to the content of the summary item by operating the buttons on the navigation bar SP2, the received message text information ML is loaded as an example of a text TX or file FL specified by the browser 111, and the received message text information ML is displayed as a resource.

[0060] In addition, above the text information display area SP1b, there are tabs labeled "Email Content" and "Entire Thread". If one email is selected and the first icon button 111a and the "Summary of Email Content" menu button 112b are selected, the "Email Content" tab will be displayed, and the text information display field SP1b will be populated with the text information ML of the selected received message. When the "Entire Thread" tab is accessed, the browser 111 loads the email text information for the entire thread, including the specified received message text information ML and other emails related to that thread, and displays it in the text information display field SP1b.

[0061] Additionally, the SP1n length setting field allows users to select one of the following options for the length of the generated text: automatic, short, medium, or long. Then, when the "Summarize Document" button SP1m is pressed, the user terminal 110 sends the received message text information ML, which is displayed in the text information display field SP1b as an example of resource information, prompt information about creating a summary and its length, to the first server 120. Then, the first server 120 generates text data as a summary using ChatGPT, which is an example of a pre-trained artificial intelligence means 131 of the second server 130, based on the received message text information ML and prompt information. Then, the first server 120 sends the generated text data GT, which is the summary obtained, to the user terminal 110.

[0062] Then, the browser 111 of the user terminal 110 displays the generated text data GT, which is the received summary, in the generated text display field SP1f of the side panel SP. In this way, for the text TX specified in the browser 111, the generated text data GT as a summary can be easily obtained using Chat GPT, which is an example of a pre-trained artificial intelligence means 131, by operating the "Summarize Document" button SP1m on the side panel SP.

[0063] Next, as shown in Figure 5(A), when the first icon button 111a provided on the browser 111 is operated as an example of a predetermined operation to display the side panel SP, with the received message text information ML specified as an example of a text TX or file FL in the browser 111 of the user terminal 110, the browser 111 displays a number of item menu buttons. Furthermore, as an example of a predetermined operation to display the side panel SP, let's assume that the "Find problems in emails" menu button 112c is selected as one of several item menu buttons.

[0064] Then, as shown in Figure 5(B), the browser 111 displays the side panel SP. The side panel SP includes, for example, a text information display field SP1b, an "Identify Problems" button SP1p which is an execution button for the content of the item, a generated text display field SP1f, a re-execution button SP1g, and a copy button SP1h. Furthermore, it is desirable that the generated text display area SP1f and other elements below the "Identify Problems" button SP1p be displayed after the "Identify Problems" button SP1p has been operated. The reason is that, in terms of the order of operations, the "Find the Problem" button SP1p should be pressed first, and only then should the subsequent information be displayed, making it easier for the user to understand.

[0065] In the text information display area SP1b, when the side panel SP is displayed or when the display is switched to the content of the summary item by operating the buttons on the navigation bar SP2, the received message text information ML is loaded as an example of a text TX or file FL specified by the browser 111, and the received message text information ML is displayed as a resource.

[0066] Then, when the "Find the problem" button SP1p is pressed, the user terminal 110 sends the received message text information ML, which is displayed in the text information display field SP1b as an example of resource information, and prompt information about finding the problem to the first server 120. Then, based on the received message text information ML and prompt information, the first server 120 uses ChatGPT, an example of a pre-trained artificial intelligence means 131 of the second server 130, to determine whether there are any problems and to generate text data about the problems. Then, the first server 120 sends the presence or absence of problems and generated text data GT about the problems to the user terminal 110.

[0067] Then, the browser 111 of the user terminal 110 displays the presence or absence of received problems and the generated text data GT about the problems in the generated text display field SP1f of the side panel SP. In this way, by operating the "Problem Detection" button SP1p on the side panel SP for the text TX specified in the browser 111, it is possible to easily obtain information about the presence or absence of problems and generated text data GT about the problems using Chat GPT, which is an example of a pre-trained artificial intelligence means 131.

[0068] As shown in Figure 6(A), when the first icon button 111a provided on the browser 111 is operated as an example of a predetermined operation to display the side panel SP, with the received message text information ML specified as an example of a text TX or file FL in the browser 111 of the user terminal 110, the browser 111 displays multiple item menu buttons. Furthermore, as an example of a predetermined operation to display the side panel SP, let's assume that the "Meeting Time Scheduling" menu button 112d is selected as one of several item menu buttons.

[0069] Then, as shown in Figure 6(B), the browser 111 displays the side panel SP. The side panel SP includes, for example, a text information display field SP1b, an execution button for the item's content called "Schedule a Meeting" button SP1q, a candidate date display field SP1r which is an example of a generated text display field SP1f, an advanced settings button, a "Create Email Body" button SP1a which is a button for creating a reply message, a generated text display field SP1f, a selection field SP1s for tentatively booking an event on the calendar, a re-execution button SP1g, a copy button SP1h, and a "Add to Email Creation Screen" button SP1k which indicates that the message will be reflected in the message creation screen / field.

[0070] Furthermore, it is desirable that the display of candidate dates, such as SP1r (an example of the generated text display field SP1f below the "Schedule Meeting Time" button SP1q), occurs after the "Schedule Meeting Time" button SP1q has been operated. The reason is that, in terms of the order of operations, the "Schedule Meeting" button SP1q should be operated first, and only then should the subsequent options be displayed, making it easier for the user to understand. Furthermore, it is desirable that the generated text display area SP1f and other elements below the "Create Email Body" button SP1a be displayed after the "Create Email Body" button SP1a has been pressed. The reason is that, in terms of the order of operations, the "Create Email Body" button SP1a should be pressed first, and only then should the subsequent information be displayed, making it easier for the user to understand.

[0071] In the text information display area SP1b, when the side panel SP is displayed or when the display is switched to the content of the summary item by operating the buttons on the navigation bar SP2, the received message text information ML is loaded as an example of a text TX or file FL specified by the browser 111, and the received message text information ML is displayed as a resource. In other words, one of the items in the menu buttons is for scheduling a meeting. When the menu button for scheduling a meeting is operated, the browser 111 displays the received message text information ML, which is specified as a resource in the browser 111, in the text information display field SP1b of the side panel SP, which is a predetermined area panel, and also displays the "Schedule Meeting" button SP1q, which is an execution button for scheduling a meeting, on the side panel SP.

[0072] First, when the "Schedule Meeting" button SP1q is pressed, the browser 111 accesses the schedule management system server based on pre-configured user information and obtains schedule information SD (including schedule availability information) for the user on the user terminal 110. Along with this, the user terminal 110 sends a prompt to the first server 120 indicating that it will search for and list multiple schedule information SDs (including schedule availability information) for the acquired user. In this case, the operation of the "Meeting Time Adjustment" button SP1q may trigger the first server 130 to access the schedule management system server based on a schedule information acquisition command from the user terminal 110, acquire schedule information SD (including schedule availability information) for the user of the user terminal 110, and also receive prompt information from the user terminal 110 indicating that it will search for and list multiple schedule information SD (including schedule availability information) for the acquired user.

[0073] Then, based on the received schedule information SD and prompt information, the first server 120 uses ChatGPT, an example of the pre-trained artificial intelligence means 131 of the second server 130, to generate text data as candidate date information for scheduling a meeting. Then, the first server 120 sends the generated text data GT, which is obtained as candidate date information for scheduling the meeting, to the user terminal 110.

[0074] Then, the browser 111 of the user terminal 110 displays the generated text data GT, which is the received meeting time adjustment candidate date information, in the candidate date display field SP1r, which is also the generated text display field SP1f of the side panel SP. Along with this, browser 111 displays the "Create Email Body" button SP1a, which is a button for creating a reply message, on the side panel SP. In this way, by operating the "Schedule a Meeting" button SP1q on the side panel SP for the text TX specified in the browser 111, it is possible to easily obtain generated text data GT as candidate date information for scheduling a meeting using chat GPT, which is an example of a pre-trained artificial intelligence means 131. Additionally, the candidate date display field SP1r includes a copy button SP1h. Furthermore, in the SP1r suggested date display area, suggested dates for scheduling meetings are displayed in a highly editable format. In other words, the suggested meeting dates displayed in the SP1r section can be freely changed by the user.

[0075] Next, when the "Create email body" button SP1a is pressed, the user terminal 110 sends prompt information to the first server 120, including the received message text information ML displayed in the text information display field SP1b as an example of resource information, the candidate date information for scheduling multiple meetings displayed in the candidate date display field SP1r, and the intention to create an email for the purpose of scheduling a meeting. Then, the first server 120 uses ChatGPT, an example of the pre-trained artificial intelligence means 131 of the second server 130, to generate text data as a proposed reply email for scheduling a meeting, which is a proposed reply message, based on the received message text information ML, candidate date information for scheduling multiple meetings, and prompt information. Then, the first server 120 sends the generated text data GT, which is a draft reply email for scheduling the meeting, to the user terminal 110. Then, the browser 111 of the user terminal 110 displays the generated text data GT, which is the received draft reply email for scheduling the meeting, in the generated text display field SP1f of the side panel SP. The generated text data GT, intended as a draft reply email for scheduling a meeting, includes information on multiple possible meeting dates displayed in the candidate date display field SP1r.

[0076] Furthermore, if the information transmitted from the user terminal 110 to the chat GPT, which is an example of a pre-trained artificial intelligence means 131 on the second server 130 via the first server 120, includes information that identifies a company or individual, such as an address, company name, personal name, or email address, the browser 111 on the user terminal 110 or the first server 120 may be configured to mask (convert to dummy information with the same attributes) the information that identifies the company or individual before sending it to the chat GPT, which is an example of a pre-trained artificial intelligence means 131 on the second server 130. In the generated text data GT obtained, the masked portions may be converted back to the original company or individual identifying information (by inversely converting the converted dummy information back to the original information) before being displayed in the generated text display field SP1f. As a result, ChatGPT, an example of a pre-trained artificial intelligence method 131, does not learn information that identifies companies or individuals. As a result, it is possible to avoid the leakage of company or individual-identifiable information to other terminals via Chat GPT, which is an example of a pre-trained artificial intelligence means 131.

[0077] Furthermore, let's assume that the user has clicked the "Add to email composition screen" button SP1k, which is a button that indicates that the message will be reflected in the message composition screen / field. Then, as shown in Figure 6(B), the browser 111 displays a reply email window 113, which is an example of a reply message window, and reflects and displays the generated text data GT displayed in the generated text display field SP1f of the side panel SP in the reply email window 113.

[0078] This sets the reply email data (generated text data GT) into the reply email window 113, which is an example of a reply message window, and puts the email in a state just before sending, where only the send button needs to be pressed. As a result, users can easily prepare reply emails to schedule meetings. In other words, when the user operates the "Schedule Meeting" button SP1q, which is the execution button, Chat GPT, which is an example of the pre-trained artificial intelligence means 131, picks up multiple candidate meeting date and time information MT and displays them in the candidate date display field SP1r of the side panel SP, which is a predetermined area panel. As a result, users can avoid the hassle of checking schedule information (SD), which is essentially calendar information, within the schedule management system.

[0079] Furthermore, when the user operates the "Create Email Body" button SP1a, which is a button for creating a reply message, the Chat GPT, which is an example of a pre-trained artificial intelligence means 131, displays a draft reply message information that reflects multiple meeting candidate date and time information MT on the side panel SP, which is a predetermined area panel. As a result, users can easily prepare draft reply messages for scheduling using Chat GPT, which is an example of a pre-trained artificial intelligence means 131. Furthermore, when a user operates the "Add to email composition screen" button SP1k, which indicates that the message will be reflected in the message composition screen / field, the reply message draft information is set in the reply email window 113, which is an example of a reply message window / field. As a result, users can easily reach the point just before sending, where all they have to do is press the send button.

[0080] Furthermore, in this embodiment, as shown in Figure 6(B), when the browser 111 displays the generated text data GT, which is a draft reply email for scheduling a meeting and is generated by Chat GPT, an example of a pre-trained artificial intelligence means 131, on a predetermined area panel called the side panel SP, and also displays the "Add to email creation screen" button SP1k, which is a button indicating that the message will be reflected in the message creation screen / field, it also displays a selection field SP1s for provisionally booking an appointment in a calendar that the user can select, indicating that the appointment will be provisionally booked in the user's schedule.

[0081] Then, in the selection field SP1s for tentatively booking an event on the calendar, the user selects to tentatively book an event on their schedule, and the "Add to email composition screen" button SP1k, which is the button that reflects the event in the message composition screen, is operated. At this time, the user terminal 110 is configured to send information to the schedule management system server indicating that it has made a provisional reservation for multiple meeting date and time information MT. More specifically, as shown in Figure 7(A), when the user selects to tentatively book an appointment in the selection field SP1s for tentatively booking an appointment in the calendar, and the "Add to email creation screen" button SP1k is operated, the browser 111 displays a tentative booking confirmation window 114 for user confirmation. Then, when the "Make a reservation" operation is performed, the user terminal 110 sends information to the schedule management system server indicating that a provisional reservation has been made for multiple meeting date and time information MT. In conjunction with this, the user terminal 110 obtains link information for the schedule management system server, and the browser 111 displays a calendar display link button SP1t containing the link information on the side panel SP. When the "Add to email composition screen" button SP1k was pressed, it included information on three possible meeting dates, and accordingly, three calendar display link buttons SP1t were displayed on the side panel SP.

[0082] Let's assume that the calendar display link button SP1t is operated by the user. Then, the user terminal 110 accesses the schedule management system server based on the link destination information of the calendar display link button SP1t. Then, as shown in Figure 7(B), the browser 111 displays a calendar / schedule screen 115 for the user's schedule information SD. On the calendar / schedule screen 115, the provisional reservation schedule information KSJ is displayed for the proposed meeting date and time, based on the information indicating that the three proposed meeting dates registered on the schedule management system server have been tentatively booked.

[0083] As a result, provisional reservation schedule information KSJ is registered for multiple candidate meeting dates and times in the user's schedule information SD, indicating that the reservation is tentative. As a result, it becomes possible to avoid accidentally scheduling other appointments at multiple potential meeting times during the scheduling process. Furthermore, even if you use the "Schedule Meeting" button SP1q, which is an action button for scheduling a meeting for a separate matter, the date and time of a tentatively scheduled meeting that has already been entered will not be listed as a candidate date and time for the meeting for that separate matter. As a result, it becomes possible to avoid situations where multiple provisional bookings overlap.

[0084] As shown in Figure 8(A), when a group of threads related to one email from the received message text information ML is specified as an example of a text TX or file FL in the browser 111 of the user terminal 110, the second icon button 111c is displayed below the thread in the browser 111. Then, as an example of a predetermined operation to display the side panel SP, when the second icon button 111c provided in the browser 111 is operated, the browser 111 displays multiple item menu buttons. As an example of multiple menu buttons, the system is configured to display the "Summary of the entire thread" menu button 112e, the "Identify problems in the entire thread" menu button 112f, and the "Schedule a meeting" menu button 112d. As an example of a predetermined operation to display the side panel SP, let's assume that the "Summary of the entire thread" menu button 112e is selected as one of several item menu buttons.

[0085] Then, as shown in Figure 8(B), the browser 111 displays the side panel SP. The side panel SP includes, for example, a text information display field SP1b, a length setting field SP1n which is a prompt setting / input field, a "Summarize Document" button SP1m which is an execution button for the content of the item, a generated text display field SP1f, a re-execution button SP1g, and a copy button SP1h. Furthermore, it is desirable that the generated text display area SP1f and other elements below the "Summarize Document" button SP1m be displayed after the "Summarize Document" button SP1m has been operated. The reason is that, in terms of the order of operations, the "Summarize Document" button SP1m should be operated first, and only then should the subsequent information be displayed, making it easier for the user to understand.

[0086] The text information display area SP1b reads the received message text information ML for the entire thread specified by the browser 111, and displays the received message text information ML for the entire thread as a resource. Then, when the "Summarize Document" button SP1m is pressed, the user terminal 110 sends the received message text information ML of the entire thread as an example of resource information, prompt information about creating a summary and its length to the first server 120. Then, the first server 120 generates text data as a summary using ChatGPT, which is an example of a pre-trained artificial intelligence means 131 of the second server 130, based on the received message text information ML and prompt information of the entire thread.

[0087] Then, the first server 120 sends the generated text data GT, which is the summary obtained, to the user terminal 110. Then, the browser 111 of the user terminal 110 displays the generated text data GT, which is the received summary, in the generated text display field SP1f of the side panel SP. In this way, for the text TX of the entire thread specified in the browser 111, the generated text data GT as a summary can be easily obtained using Chat GPT, which is an example of a pre-trained artificial intelligence means 131, by operating the "Summarize Document" button SP1m on the side panel SP.

[0088] Furthermore, when the email document editing button SP2e on the navigation bar SP2 of the side panel SP is operated, the browser 111 displays the text information display field SP1b and an example of an execution button, the "Edit Email Body" button, on the side panel SP. Then, when the text TX of the email draft is entered into the text information display field SP1b and the "Edit Email Body" button is operated, the generated text data GT as the revised email draft is generated by chatGPT, which is an example of a pre-trained artificial intelligence means 131, and displayed in the generated text display field SP1f. I will omit the detailed explanation.

[0089] The artificial intelligence system 100, which is the first embodiment of the present invention obtained in this way, comprises a user terminal 110 and a first server 120 and a second server 130 as servers. When at least one of text TX, images, and files FL is displayed and specified in the browser 111 of the user terminal 110, and the user operates the side panel icon button 111b as an example of a predetermined operation (or the first icon button 111a or the second icon button 111c is used and an item menu button is operated), the browser 111 displays a side panel SP, which is an example of a predetermined area panel that can display the interaction with Chat GPT, an example of a pre-trained artificial intelligence means 131 of the second server 130, via the first server 120. As a result, the user can easily interact with Chat GPT, an example of a pre-trained artificial intelligence means 131, with desired content on the side panel SP, targeting the content displayed and specified in the browser 111.

[0090] Furthermore, as a predetermined operation, when the first icon button 111a or the second icon button 111c is operated as an example of an icon button provided in the browser 111, the browser 111 displays multiple item menu buttons, and when the "Create Reply Email Document" menu button 112a is selected as one of the multiple item menu buttons, the browser 111 displays a side panel SP that can display the interaction with Chat GPT, which is an example of a pre-trained artificial intelligence means 131 of the second server 130 via the first server 120, and the side panel SP has a "Create Email Body" button SP1a as an example of a button representing the content of the item of the "Create Reply Email Document" menu button 112a as one of the selected multiple item menu buttons, the user can easily interact with Chat GPT, which is an example of a pre-trained artificial intelligence means 131, by selecting the desired item menu in the browser 111 and using the content of the desired item in the side panel SP.

[0091] Furthermore, the side panel SP is configured to include a text information display field SP1b as a text / image information display field specified as a resource in the browser 111, a prompt input field SP1c as an example of a prompt setting / input field for instructions to Chat GPT, which is an example of a pre-trained artificial intelligence means 131, and a "Create email body" button SP1a as an example of an execution button for the contents of an item. As a result, the user can easily create summaries and reply documents by interacting with Chat GPT, which is an example of a pre-trained artificial intelligence means 131.

[0092] Furthermore, when the content of one of the multiple item menu buttons is "Meeting Time Adjustment," and the "Meeting Time Adjustment" menu button 112d, which is an item menu button indicating "Meeting Time Adjustment," is operated, the browser 111 displays the received message text information ML, which is specified as a resource in the browser 111, in the text information display field SP1b of the side panel SP, and also displays the "Meeting Time Adjustment" button SP1q, which is an execution button indicating "Meeting Time Adjustment," on the side panel SP. When the "Meeting Time Adjustment" button SP1q, which is an execution button, is operated, the browser 111 accesses the schedule management system server based on pre-configured user information to obtain schedule information SD for the user of the user terminal 110, and via the first server 120, uses chat GPT, which is an example of a pre-trained artificial intelligence means 131 of the second server 130, to pick up multiple meeting candidate date and time information MT based on the schedule blank information in the schedule information SD, and displays them on the side panel SP. Both display the "Create Email Body" button SP1a, which is a button for creating a reply message. When the "Create Email Body" button SP1a is operated, the user terminal 110 sends the received message text information ML, the information on multiple meeting candidate dates and times MT, and a prompt indicating that a reply message will be created to Chat GPT, an example of a pre-trained artificial intelligence means 131 on the second server 130, via the first server 120. The browser 111 displays the generated text data GT, which is the reply message draft information generated by Chat GPT, an example of a pre-trained artificial intelligence means 131, on the side panel SP, and also displays the "Add to Email Creation Screen" button SP1k, which indicates that the information will be reflected in the message creation screen / field. When the "Add to Email Creation Screen" button SP1k is operated, the browser 111 displays the generated text data GT, which is the reply message draft information, in the reply email window 113, an example of a reply message window / field. This configuration allows the user to, so to speak,This eliminates the need to manually check schedule information (SD) in a schedule management system. Users can easily prepare draft reply messages for schedule adjustments using Chat GPT, an example of a pre-trained artificial intelligence tool 131, and then simply press the send button to reach the pre-send state.

[0093] Furthermore, when the browser 111 displays the generated text data GT, which is a draft reply message information generated by chat GPT, an example of a pre-trained artificial intelligence means 131, on the side panel SP, and also displays the "Add to email creation screen" button SP1k, which is a button indicating that the message will be reflected in the message creation screen / field, it also displays the option to tentatively reserve an appointment in the user's schedule in the selection field SP1s for tentatively reserving an appointment in the calendar. When the option to tentatively reserve an appointment in the user's schedule is selected and the "Add to email creation screen" button SP1k, which is a button indicating that the message will be reflected in the message creation screen / field, is operated, the user terminal 110 sends information to the schedule management system server indicating that multiple candidate meeting date and time information MT have been tentatively reserved. This configuration prevents other appointments from being mistakenly scheduled for multiple candidate meeting dates and times during the scheduling adjustment stage, and also prevents tentative reservations from overlapping even if the "Adjust meeting time" button SP1q, which is an execution button for adjusting the meeting time for a separate matter, is operated.

[0094] Furthermore, the program of the artificial intelligence system 100, which is the first embodiment of the present invention, includes an icon operation determination step S1 which determines whether a first icon button 111a or a second icon button 111c has been operated as an example of an icon button provided in the browser 111 of the user terminal 110, while at least one of text TX, an image, and a file FL is displayed and specified in the browser 111 of the user terminal 110; an item menu selection operation determination step S2, S3 which determines whether, when the first icon button 111a or the second icon button 111c is operated, the browser 111 displays a plurality of item menu buttons and whether, for example, the "Create Reply Email Document" menu button 112a has been selected as one of the plurality of item menu buttons; and for example, the "Create Reply Email Document" menu button 112a has been selected as one of the plurality of item menu buttons. When the "Create reply email document" menu button 112a is selected, the browser 111 displays the interaction with Chat GPT, an example of a pre-trained artificial intelligence means 131 on the second server 130 via the first server 120. By causing the computer to execute a panel display step S4, which displays a side panel SP, an example of a predetermined area panel, having a "Create email body" button SP1a as an example of a button representing the content of the item of the "Create reply email document" menu button 112a, as one of the selected item menu buttons, the user can easily interact with Chat GPT, an example of a pre-trained artificial intelligence means 131, by selecting the desired item menu in the browser 111 and using the content of the desired item in the side panel SP, the effect is enormous. [Examples]

[0095] Next, an artificial intelligence system 100, which is a second embodiment of the present invention, will be described based on Figures 9(A) to 10(B). Here, Figure 9(A) shows a predetermined operation to display the side panel SP while the user's schedule information SD is displayed in the browser 111 of the user terminal 110 of the artificial intelligence system 100, which is a second embodiment of the present invention. Figure 9(B) shows a predetermined operation to display the side panel SP while the user's schedule information SD is displayed in the side panel SP of the browser 111. Figure 10(A) shows a predetermined operation to display the side panel SP while the user's schedule information SD is displayed in the browser 111 of the user terminal 110 of the artificial intelligence system 100, which is a second embodiment of the present invention. Figure 10(B) shows a predetermined operation to specify a meeting time by selecting a partner from the calendar schedule information SD in the side panel SP of the browser 111. The artificial intelligence system 100 of the second embodiment is the same as the artificial intelligence system 100 of the first embodiment, but with the target specified in the browser 111 of the user terminal 110 changed from email to schedule information SD. Since many elements are common with the artificial intelligence system 100 of the first embodiment, a detailed explanation of the common items will be omitted.

[0096] In the artificial intelligence system 100, which is a second embodiment of the present invention, as shown in Figure 9(A), the browser 111 of the user terminal 110 displays the user's schedule information SD for today based on the user's schedule information SD of the schedule management system server. Browser 111 displays the user's weekly schedule information (SD), but today, which is one day of the week, is selected. Then, when the first icon button 111a provided on the browser 111 is operated as an example of a predetermined operation to display the side panel SP, with the schedule information SD for today specified as an example of a text TX or file FL in the browser 111 of the user terminal 110, the browser 111 displays multiple item menu buttons. As an example of multiple menu buttons, there is a "Create Daily Report" menu button 112g for creating a daily report, a "Create Weekly Report" menu button 112h for creating a weekly report, a "Create Monthly Report" menu button 112k for creating a monthly report, and a "Schedule Meeting" menu button 112d for scheduling a meeting. Furthermore, as an example of a predetermined operation to display the side panel SP, let's assume that the "Create Daily Report" menu button 112g is selected as one of several item menu buttons.

[0097] Then, as shown in Figure 9(B), the browser 111 displays the side panel SP. The side panel SP includes, for example, a text information display field SP1b, a length setting field SP1n which is a prompt setting / input field, a "Create Report" button SP1u which is an execution button for the contents of the item, a generated text display field SP1f, a re-execution button SP1g, a copy button SP1h, a "Add to Email Creation Screen" button SP1k which is a button to reflect in the message creation screen / field, and a "Submit as File" button SP1v which is a daily report file creation button. Furthermore, it is desirable that the generated text display area SP1f and other elements below the "Create Report" button SP1u be displayed after the "Create Report" button SP1u has been pressed. The reason is that, in terms of the order of operations, the "Create Report" button SP1u should be clicked first, and only then should the subsequent information be displayed, making it easier for the user to understand.

[0098] When the side panel SP is displayed or when the display is switched to the contents of the daily report creation items by operating the buttons on the navigation bar SP2, the schedule information SD for today is loaded as an example of a text TX or file FL specified by the browser 111, and the schedule information SD for today is displayed as a resource.

[0099] In addition, above the text information display area SP1b, there are tabs for "Daily Content," "Weekly Content," and "Monthly Content." If a daily schedule information SD is specified and the first icon button 111a or the "Create Daily Report" menu button 112g is selected, the "Daily Contents" tab will be displayed, and the specified daily schedule information SD will be entered into the text information display field SP1b. When the "Weekly Content" tab is accessed, the browser 111 loads the schedule information SD for the week, including the specified day, and displays it in the text information display field SP1b. Similarly, when the "Monthly Contents" tab is accessed, the browser 111 loads the monthly schedule information SD, including the specified day, and displays it in the text information display field SP1b.

[0100] Then, when the "Create Report" button SP1u is pressed, the user terminal 110 sends the daily schedule information SD displayed in the text information display field SP1b as an example of resource information, prompt information regarding the creation of a daily report and its length to the first server 120. Then, the first server 120, based on the received schedule information SD and prompt information, uses ChatGPT, which is an example of a pre-trained artificial intelligence means 131 of the second server 130, to generate text data as daily report draft information. Then, the first server 120 sends the generated text data GT, which is obtained as daily report draft information, to the user terminal 110.

[0101] Then, the browser 111 of the user terminal 110 displays the generated text data GT, which is the received daily report draft information, in the generated text display field SP1f of the side panel SP. Along with this, Browser 111 displays either the "Add to email composition screen" button SP1k, which indicates that the message will be reflected in the message composition screen / field, or the "Submit as file" button SP1v, which is a button for creating a daily report file, on the side panel SP. In this way, by operating the "Create Daily Report" button on the side panel SP for the text TX specified in the browser 111, it is possible to easily obtain generated text data GT as a report using Chat GPT, which is an example of a pre-trained artificial intelligence means 131. Then, when the "Add to email composition screen" button SP1k, which is a button that indicates that the information will be reflected in the message composition screen / field, is operated, the browser 111 opens an email sending window as an example of a message window / field, and displays the daily report draft information in the email sending window. Alternatively, when the "Submit as File" button SP1v, which serves as the daily report file creation button, is operated, the browser 111 is configured to output a file FL containing the daily report draft information to a pre-configured location.

[0102] As a result, when the user is viewing today's schedule information SD in browser 111 and operates the "Create Daily Report" menu button 112g, which is an item menu button for creating a daily report, and then operates the "Create Report" button SP1u, which is an execution button, the daily report draft information is generated and displayed by ChatGPT, which is an example of a pre-trained artificial intelligence means 131. As a result, users can easily prepare draft daily reports. Furthermore, pressing the "Add to Email Creation Screen" button SP1k, which indicates that the message will be reflected in the message creation screen / field, or the "Submit as File" button SP1v, which is a button for creating a daily report file, will return the system to the state just before sending the daily report message, or the daily report file FL will be output to a designated location. As a result, users can easily submit their daily reports.

[0103] Furthermore, in Figure 9(A), assume that the first icon button 111a was operated, and the "Create Weekly Report" menu button 112h was operated. Browser 111 then displays the side panel SP. At this time, the side panel SP will display the "Weekly Content" tab. In other words, the schedule information SD for one week, including the specified day, is read by the browser 111 and displayed in the text information display field SP1b. Then, when the "Create Report" button SP1u is pressed, the user terminal 110 sends the weekly schedule information SD displayed in the text information display field SP1b as an example of resource information, prompt information regarding the creation of a weekly report and its length to the first server 120. Then, the first server 120 generates text data as weekly report draft information using ChatGPT, which is an example of a pre-trained artificial intelligence means 131 of the second server 130, based on the received schedule information SD and prompt information. Then, the first server 120 sends the generated text data GT, which is obtained as weekly report draft information, to the user terminal 110.

[0104] Then, the browser 111 of the user terminal 110 displays the generated text data GT, which is the received weekly report draft information, in the generated text display field SP1f of the side panel SP. Along with this, browser 111 displays either the "Add to email composition screen" button SP1k, which indicates that the message will be reflected in the message composition screen / field, or the "Submit as file" button SP1v, which is a button for creating a weekly report file, on the side panel SP. In this way, by operating the "Create Weekly Report" button on the side panel SP for the text TX specified in the browser 111, it is possible to easily obtain generated text data GT as a report using Chat GPT, which is an example of a pre-trained artificial intelligence means 131. Then, when the "Add to email composition screen" button SP1k, which is a button that indicates that the information will be reflected in the message composition screen / field, is operated, the browser 111 opens an email sending window as an example of a message window / field, and displays the weekly report draft information in the email sending window. Alternatively, when the "Submit as File" button SP1v, which serves as the weekly report file creation button, is operated, the browser 111 is configured to output a file FL containing the weekly report draft information to a pre-configured location.

[0105] As a result, when the browser 111 is displaying weekly schedule information SD, if the user operates the "Create Weekly Report" menu button 112h, which is an item menu button for creating a weekly report, and then operates the "Create Report" button SP1u, which is an execution button, weekly report draft information will be generated and displayed by ChatGPT, which is an example of a pre-trained artificial intelligence means 131. As a result, users can easily prepare draft weekly reports. Furthermore, pressing the "Add to Email Creation Screen" button SP1k, which indicates that the message will be reflected in the message creation screen / field, or the "Submit as File" button SP1v, which is a button for creating a weekly report file, will return the system to the state just before sending the weekly report message, or the weekly report file FL will be output to a designated location. As a result, users can easily submit their weekly reports. In Figure 9(A), if the first icon button 111a is operated and the "Create Monthly Report" menu button 112k is operated, the same processing as when the "Create Weekly Report" menu button 112h is operated is performed, and the browser 111 of the user terminal 110 displays the generated text data GT as received monthly report draft information in the generated text display field SP1f of the side panel SP. I will omit the detailed explanation.

[0106] As shown in Figure 10(A), the browser 111 of the user terminal 110 displays the user's schedule information SD for a predetermined period based on the user's schedule information SD on the schedule management system server. When a predetermined schedule information SD for a predetermined period is specified in the browser 111 of the user terminal 110 as an example of a text TX or file FL, and the first icon button 111a provided in the browser 111 is operated as an example of a predetermined operation to display the side panel SP, the browser 111 displays multiple item menu buttons. Furthermore, as an example of a predetermined operation to display the side panel SP, let's assume that the "Meeting Time Scheduling" menu button 112d is selected as one of several item menu buttons.

[0107] Then, as shown in Figure 10(B), the browser 111 displays the side panel SP. The side panel SP includes, for example, a "Specify recipient in address book" button SP1w, a recipient information display field SP1x, a "Schedule meeting" button SP1q, a candidate date display field SP1r which is also a generated text display field SP1f, an advanced settings button, a "Create email body" button SP1a, a generated text display field SP1f, a selection field SP1s for tentatively booking an event in the calendar, a redo button SP1g, a copy button SP1h, and an "Add to email creation screen" button SP1k.

[0108] Regarding the timing of displaying the "Specify recipient from address book" button SP1w and the "Schedule meeting time" button SP1q below the recipient information display field SP1x, it is desirable that these buttons be displayed after the "Specify recipient from address book" button SP1w has been operated and the recipient information has been entered into the recipient information display field SP1x. The reason is that, in terms of the order of operations, the "Specify recipient in address book" button SP1w should be operated first, and the recipient information should be entered into the recipient information display field SP1x, after which the following information should be displayed, making it easier for the user to understand.

[0109] In other words, when the "Schedule Meeting" menu button 112d, which is a menu button for scheduling a meeting, is operated, the browser 111 displays the "Specify Contact in Address Book" button SP1w and the contact information display field SP1x on the side panel SP. Then, when the "Specify recipient from address book" button SP1w is pressed, the browser 111 displays the name and email address information of other users that have been previously registered in the address book. Next, when another user's name or email address is selected from the address book, the selected user's name or email address is entered into the recipient information display field SP1x. Then, browser 111 displays the "Schedule Meeting" button SP1q on the side panel SP, which is an execution button for scheduling a meeting. When the "Schedule Meeting" button SP1q is pressed by the user, the browser 111 accesses the schedule management system server based on pre-configured user information and obtains schedule information SD for the user on the user terminal 110, similar to the embodiment described above. Along with this, the user terminal 110 sends the acquired schedule information SD for the user, and prompt information indicating that it will search for and list multiple available schedule slots, to the first server 120.

[0110] Then, based on the received schedule information SD and prompt information, the first server 120 uses ChatGPT, an example of the pre-trained artificial intelligence means 131 of the second server 130, to generate text data as candidate date information for scheduling a meeting. Then, the first server 120 sends the generated text data GT, which is obtained as candidate date information for scheduling the meeting, to the user terminal 110. Then, the browser 111 of the user terminal 110 displays the generated text data GT, which is the received meeting time adjustment candidate date information, in the candidate date display field SP1r, which is also the generated text display field SP1f of the side panel SP. Along with this, browser 111 displays the "Create Email Body" button SP1a, which is a button for creating a reply message, on the side panel SP. In this way, by operating the "Schedule a Meeting" button SP1q on the side panel SP for the text TX specified in the browser 111, it is possible to easily obtain generated text data GT as candidate date information for scheduling a meeting using chat GPT, which is an example of a pre-trained artificial intelligence means 131.

[0111] Next, when the "Create email body" button SP1a is pressed, the user terminal 110 sends prompt information to the first server 120, including the recipient information displayed in the recipient information display field SP1x as an example of resource information, the candidate date information for scheduling multiple meetings displayed in the candidate date display field SP1r, and the intention to create an email for the purpose of scheduling a meeting. Then, based on the received recipient information, candidate date information for scheduling multiple meetings, and prompt information, the first server 120 uses Chat GPT, an example of the pre-trained artificial intelligence means 131 of the second server 130, to generate text data as a draft email for scheduling a meeting, which is message draft information. Then, the first server 120 sends the generated text data GT, which is obtained as a draft email for scheduling the meeting, to the user terminal 110. Then, the browser 111 of the user terminal 110 displays the generated text data GT, which is a draft email for scheduling a meeting, in the generated text display field SP1f of the side panel SP. The generated text data GT, intended as a draft email for scheduling a meeting, includes information on multiple possible meeting dates displayed in the candidate date display field SP1r.

[0112] Furthermore, let's assume that the user has clicked the "Add to email composition screen" button SP1k, which is a button that indicates that the message will be reflected in the message composition screen / field. Then, browser 111 displays a mail window, which is an example of a message window, and reflects the generated text data GT displayed in the generated text display field SP1f of the side panel SP into the mail window for display. This sets the email data (generated text data GT) of the draft meeting time scheduling email, which is the message draft information, into the sending email window, which is an example of a message window, and puts it in a state just before sending, where all that is needed is to press the send button. As a result, users can easily prepare emails to schedule meetings. The other buttons and other components are the same as in the previously described embodiment, so a detailed explanation will be omitted.

[0113] In other words, when the user operates the "Schedule Meeting" button SP1q, which is the execution button, Chat GPT, which is an example of the pre-trained artificial intelligence means 131, picks up multiple candidate meeting date and time information MT and displays them in the candidate date display field SP1r of the side panel SP, which is a predetermined area panel. As a result, users can avoid the hassle of checking the details of schedule information (SD), which is essentially calendar information, within the schedule management system.

[0114] Furthermore, when the user operates the "Create Email Body" button SP1a, which is a message creation button, a message draft information reflecting multiple meeting candidate date and time information MT is displayed on the side panel SP, which is a predetermined area panel, by Chat GPT, an example of a pre-trained artificial intelligence means 131. As a result, users can easily prepare message drafts for scheduling using ChatGPT, which is an example of a pre-trained artificial intelligence means 131. Furthermore, when a user operates the "Add to Email Creation Screen" button SP1k, which is a button that indicates that the message will be reflected in the message creation screen / field, the message draft information is set in the sending email window, which is an example of a message window / field. As a result, users can easily reach the point just before sending, where all they have to do is press the send button.

[0115] In the artificial intelligence system 100, which is a second embodiment of the present invention obtained in this way, the browser 111 displays the user's schedule information SD for today based on the user's schedule information SD of the schedule management system server, and when the content of one of the multiple item menu buttons is "Create a daily report", and the "Create Daily Report" menu button 112g, which is the item menu button for "Create a daily report", is operated, the browser 111 displays the schedule information SD for today, which is specified in the browser 111 as a resource, in the text information display field SP1b of the side panel SP, and also displays the "Create Report" button SP1u, which is an example of an execution button for "Create a daily report", and when the "Create Report" button SP1u, which is an example of an execution button, is operated, the user terminal 110 sends the schedule information SD for today and the prompt to "Create a daily report" to the chat GPT, which is an example of a pre-trained artificial intelligence means 131 of the second server 130, via the first server 120, and the browser 11 1 displays generated text data GT, which is daily report draft information generated by chat GPT, an example of a pre-trained artificial intelligence means 131, on the side panel SP, and also displays a "Add to email creation screen" button SP1k, an example of a button that reflects the information in the message creation screen / field, or a "Submit as file" button SP1v, an example of a button that creates a daily report file, on the side panel SP. When the "Add to email creation screen" button SP1k, an example of a button that reflects the information in the message creation screen / field, is operated, the browser 111 displays the generated text data GT, which is daily report draft information, in the sending email window as an example of a message window / field, or when the "Submit as file" button SP1v, an example of a button that creates a daily report file, is operated, the browser 111 outputs a file FL containing the generated text data GT, which is daily report draft information, to a pre-set location. As a result, the user can easily prepare a daily report draft and easily submit a daily report.

[0116] Furthermore, when the browser 111 displays the user's weekly schedule information SD based on the user's schedule information SD on the schedule management system server, and the content of one of the multiple item menu buttons is to create a weekly report, and the "Create Weekly Report" menu button 112h, which is the item menu button for creating a weekly report, is operated, the browser 111 displays the weekly schedule information SD specified in the browser 111 as a resource in the text information display field SP1b of the side panel SP, and also displays the "Create Report" button SP1u, which is an example of an execution button for creating a weekly report, and when the "Create Report" button SP1u, which is an example of an execution button, is operated, the user terminal 110 sends the weekly schedule information SD and the prompt to create a weekly report to the chat GPT, which is an example of a pre-trained artificial intelligence means 131 on the second server 130, via the first server 120, and the browser 111 The generated text data GT, which is weekly report draft information generated by Chat GPT, an example of a trained artificial intelligence means 131, is displayed on the side panel SP. The side panel SP also displays an example of a button to reflect it in the message creation screen / field, such as the "Add to email creation screen" button SP1k, or an example of a button to create a weekly report file, such as the "Submit as file" button SP1v. When the "Add to email creation screen" button SP1k is operated, the browser 111 displays the generated text data GT, which is weekly report draft information, in the sending email window as an example of a message window / field. When the "Submit as file" button SP1v is operated, the browser 111 outputs a file FL containing the weekly report draft information to a pre-set location. This configuration allows users to easily prepare weekly report drafts and easily submit weekly reports.

[0117] Furthermore, when the browser 111 displays the user's schedule information SD for a predetermined period based on the user's schedule information SD on the schedule management system server, and the content of one of the multiple item menu buttons is to adjust the meeting time, and the "Adjust Meeting Time" menu button 112d, which is the item menu button for adjusting the meeting time, is operated, the browser 111 displays the "Adjust Meeting Time" button SP1q, which is the execution button for adjusting the meeting time, on the side panel SP, and when the "Adjust Meeting Time" button SP1q, which is the execution button, is operated, the user terminal 110 accesses the schedule management system server based on pre-configured user information to obtain the schedule information SD for the user of the user terminal 110, and via the first server 120, the second server 130 uses Chat GPT, which is an example of a pre-trained artificial intelligence means 131, to pick up multiple meeting candidate date and time information MT based on the schedule blank information in the schedule information SD, and the browser 111 then selects multiple meeting The system displays candidate date and time information MT on the side panel SP, along with a message creation button, "Create Email Body" button SP1a, on the side panel SP. When the "Create Email Body" button SP1a is operated, the user terminal 110 sends multiple candidate meeting date and time information MT and a prompt indicating the creation of a message for meeting coordination purposes to Chat GPT, an example of a pre-trained artificial intelligence means 131 on the second server 130, via the first server 120. The browser 111 displays the generated text data GT, which is message draft information generated by Chat GPT, an example of a pre-trained artificial intelligence means 131, on the side panel SP, along with a "Add to Email Creation Screen" button SP1k, which indicates that the message draft information will be reflected in the message creation screen / field. When the "Add to Email Creation Screen" button SP1k is operated, the browser 111 displays the generated text data GT, which is message draft information, as an example of a message window / field in the sending email window.The user can avoid the trouble of checking the details of schedule information SD, which is essentially calendar information, in the schedule management system. Furthermore, when the user operates the "Create Email Body" button SP1a, which is a message creation button, the user can easily prepare generated text data GT, which is message draft information for schedule adjustment, using Chat GPT, an example of a pre-trained artificial intelligence means 131. When the user operates the "Add to Email Creation Screen" button SP1k, which indicates that the message will be reflected in the message creation screen / field, the user can easily reach a state just before sending, where all that remains is to press the send button. The effects are enormous. [Examples]

[0118] Next, an artificial intelligence system 100, which is a third embodiment of the present invention, will be described based on Figures 11(A) to 11(B). Here, Figure 11(A) shows a predetermined operation being performed in the browser 111 of the user terminal 110 of the artificial intelligence system 100, which is a third embodiment of the present invention, to display the side panel SP with the file FL specified from the drive information, and Figure 11(B) shows the RAG learning being performed from the text data extracted from file FL in the side panel SP of the browser 111. The artificial intelligence system 100 of the third embodiment is the same as the artificial intelligence system 100 of the first embodiment, but the target specified in the browser 111 of the user terminal 110 has been changed from email to a file FL on a drive, etc. Since many elements are common with the artificial intelligence system 100 of the first embodiment, a detailed explanation of the common items will be omitted.

[0119] In the artificial intelligence system 100, which is a third embodiment of the present invention, as shown in Figure 11(A), the browser 111 of the user terminal 110 displays a drive display screen 116 based on drive information of a recording unit (storage area) such as a cloud server. The drive display screen 116 shows information about the FL files saved on the drive. Then, when one or more files FL are specified in the browser 111 as an example of a text TX or file FL, and the first icon button 111a provided in the browser 111 is operated as an example of a predetermined operation to display the side panel SP, the browser 111 displays multiple item menu buttons. As an example of multiple menu buttons, there is a "RAG Learning" menu button 112m for loading and training data for RAG (Retrieval Augmented Generation), which is an external information source for the artificial intelligence system 100; a "Create Summary" menu button 112n for creating a summary; and a "Ask a Question" menu button 112p for asking a question. Furthermore, as an example of a predetermined operation to display the side panel SP, let's assume that the "RAG Learning" menu button 112m is selected as one of several item menu buttons.

[0120] Then, as shown in Figure 11(B), the browser 111 displays the side panel SP. The side panel SP includes, for example, a text information display field SP1b, an "RAG Learning" button SP1y which is an execution button for the content of the item, a question input field SP1z which is a prompt setting / input field, a re-execution button SP1g, and a copy button SP1h. Additionally, the question input field SP1z is equipped with a voice input button SP1d and a prompt wizard button SP1e. Furthermore, it is desirable that the question input field SP1z and other elements below the "RAG Learning" button SP1y be displayed after the "RAG Learning" button SP1y has been operated. The reason is that, in terms of the order of operations, the "RAG Learning" button SP1y should be pressed first, and only then should the subsequent information be displayed, making it easier for the user to understand.

[0121] When the side panel SP is displayed or when the display is switched to the content of the RAG learning item by operating a button on the navigation bar SP2, the first server 120 reads information from one or more files FL as an example of a text TX or file FL specified by the browser 111, and then sends it to the user terminal 110, and the language character information data is displayed as a resource in the text information display field SP1b.

[0122] Then, when the "RAG Learning" button SP1y is pressed, the user terminal 110 sends the language character information data displayed in the text information display field SP1b to the first server 120. Then, the first server 120 divides the language character information data into multiple split language character file data. Then, the first server 120 converts the divided language character file data into vector coordinate system divided language data using the Embedding model, which is an example of a vector coordinate transformation means that is a pre-trained artificial intelligence means. At the same time, it associates the converted vector coordinate system divided language data with the original divided language character file data and stores them in vector indices. Here, a vector index refers to a list, database, or conceptual spatial map in a memory unit used to quickly find a specific object from a large amount of vector coordinate system data. Furthermore, the vector coordinate transformation means, also known as the embedding model, is designed to convert input data into data in a vector coordinate system. For example, the vector components are determined based on each element of the data to be transformed and its intensity.

[0123] Although the "RAG Learning" button SP1y was used as the trigger in this example, the operation of the first icon button 111a could also be used as the trigger to execute the process up to saving to the vector index. Once saving to the vector index is complete, browser 111 displays the prompt settings / input field SP1z (question input field), the "Ask a question" button, and the "Summarize" button on the side panel SP. This allows you to ask questions about the contents of a specified file (FL), obtain answers to those questions, and create summaries. Since its detailed configuration and operation are the same as those of the embodiment described above, a detailed explanation will be omitted.

[0124] Furthermore, the browser 111 displays multiple menu buttons, and the content of one of these menu buttons is either a "Create Summary" menu button 112n indicating the creation of a summary, or a "Ask a Question" menu button 112p indicating the creation of a question. Then, when the "Create Summary" menu button 112n, which indicates the intention to summarize, is operated, the browser 111 displays the text information display field SP1b, which is specified as a resource in the browser 111, the length setting field SP1n, which is a prompt setting / input field for instructions to the chat GPT, which is an example of a pre-trained artificial intelligence means 131, and the "Summarize" button, which is an execution button for the content of summarizing or asking a question, on the side panel SP. When the "Summarize" button, which is the execution button, is pressed, the user terminal 110 sends prompt information to the first server 120 as an instruction command. Then, the first server 120 generates text data using Chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, and sends the generated text data GT to the user terminal 110. In this process, ChatGPT, an example of a pre-trained artificial intelligence means 131 of the second server 130, generates text data as a summary by referring to a vector index. The browser 111 of the user terminal 110 is configured to display the received generated text data GT on the side panel SP.

[0125] Furthermore, when the "Ask a Question" menu button 112p is pressed, the browser 111 displays the text information display field SP1b, which is specified as a resource in the browser 111, the question input field SP1z, which is a prompt setting / input field for instructions to the chat GPT, which is an example of a pre-trained artificial intelligence means 131, and the "Ask a Question" button, which is an execution button for the content of the question, on the side panel SP. When a question is entered into the question input field SP1z and the "Ask a Question" button, which is the execution button, is pressed, the user terminal 110 sends prompt information to the first server 120 as an instruction command. Then, the first server 120 generates text data using Chat GPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, and sends the generated text data GT to the user terminal 110. In this process, the first server 120 converts prompt information, including the question content, into question content vector data using a vector coordinate transformation means. Then, the first server 120 refers to the vector index and picks up one or more vector coordinate system segmented language data that have a high correlation with the question content vector data.

[0126] Next, the first server 120 uses the pre-transformed segmented language character file data associated with the picked-up vector coordinate system segmented language data, along with prompt information, to generate text data as a response using Chat GPT, an example of a pre-trained artificial intelligence means 131 on the second server 130. For example, the system can select the language data after vector coordinate system segmentation that is closest in relation to the previous one, or the language data after segmentation that corresponds to the top-ranking closest related language data after vector coordinate system segmentation, and generate text data as a response based on that. The browser 111 of the user terminal 110 is configured to display the received generated text data GT on the side panel SP.

[0127] As a result, for example, when one or more files FL on a drive such as a cloud computer or a local computer are selected and the first icon button 111a is operated, the language text data is associated as vector data and saved to a vector index as so-called RAG (Retrieval Augmented Generation) training data, which is an external information source for the artificial intelligence system 100. As a result, users can easily create training data for RAG (Retrieval Augmented Generation). Furthermore, when the user operates the execution button to summarize or ask a question, the ChatGPT, which is an example of a pre-trained artificial intelligence means 131, generates text as a summary or answer to the question based on the scope of the RAG training data, and the generated text is displayed on the side panel SP as a summary or answer to the question. As a result, users can easily obtain summaries or answers to questions within the scope of the RAG training data.

[0128] In the artificial intelligence system 100, which is the third embodiment of the present invention obtained in this way, when the first icon button 111a provided on the browser 111 is operated with one or more files FL specified in the browser 111, the first server 120 extracts language character information data as an electronic information source from the one or more specified files FL and divides it into multiple divided language character file data, converts the divided divided language character file data into vector coordinate system divided language data using a vector coordinate transformation means which is a pre-trained artificial intelligence means, and stores the converted vector coordinate system divided language data in vector indexes in association with the divided language character file data before conversion, and the browser 111 displays multiple item menu buttons (112m, 112n, 112p), and the content of one of the items on the multiple item menu buttons is either to summarize or to ask a question. When the menu buttons (112n, 112p) for items to summarize or ask a question are operated, the browser 111 displays the text information display field SP1b specified as a resource in the browser 111, the prompt setting / input field which is an instruction command for Chat GPT, an example of a pre-trained artificial intelligence means 131, and the execution button for the content of summarizing or asking a question on the side panel SP. When the execution button is operated, the user terminal 110 sends prompt information as an instruction command to the first server 120, the first server 120 generates text data using Chat GPT, an example of a pre-trained artificial intelligence means 131 of the second server 130, and sends the generated text data GT to the user terminal 110, and the browser 111 of the user terminal 110 displays the received generated text data GT on the side panel SP. Augmented Generation (RAG) offers significant advantages, such as the ease with which training data can be created, and the ability for users to easily obtain summaries or answers to questions within the scope of the RAG training data. [Examples]

[0129] Next, an artificial intelligence system 100, which is a fourth embodiment of the present invention, will be described based on Figures 12(A) to 12(B). Here, Figure 12(A) shows a predetermined operation being performed in the browser 111 of the user terminal 110 of the artificial intelligence system 100, which is a fourth embodiment of the present invention, to display the side panel SP with a message specified on the talk screen 117, and Figure 12(B) shows a process of creating a reply message for the specified message on the side panel SP of the browser 111. The artificial intelligence system 100 of the fourth embodiment is the same as the artificial intelligence system 100 of the first embodiment, but with the target specified in the browser 111 of the user terminal 110 changed from email to chat messages. Since many elements are common with the artificial intelligence system 100 of the first embodiment, a detailed explanation of the common items will be omitted.

[0130] In the artificial intelligence system 100, which is a fourth embodiment of the present invention, as shown in Figure 12(A), the browser 111 of the user terminal 110 displays a talk screen 117 (also called a chat screen) showing the message exchange between the user and other users. Here, "chat screen 117" refers to a screen used to exchange messages with one or more people, and is configured to allow users to add messages and continue the conversation on the same screen.

[0131] Then, when the first icon button 111a provided on the browser 111 is operated as an example of a predetermined operation to display the side panel SP, with one or more message text groups, which are received message text information ML, specified as an example of a text TX or file FL in the browser 111 of the user terminal 110, the browser 111 displays multiple item menu buttons. As an example of multiple menu buttons, there is a "Create Reply Chat Document" menu button 112q for creating a reply message, a "Summarize Chat Content" menu button 112r for creating a summary, a "Find Problems in Chat" menu button 112s for identifying problems, and a "Schedule Meeting" menu button 112d for scheduling a meeting. Furthermore, regarding how to specify the received message text information (ML), which is a group of message texts, one example is to select them by clicking and checking the boxes. Furthermore, as an example of a predetermined operation to display the side panel SP, let's assume that the "Create Reply Chat Document" menu button 112q is selected as one of several item menu buttons.

[0132] Then, as shown in Figure 12(B), the browser 111 displays the side panel SP. The side panel SP includes, for example, a text information display field SP1b, a prompt input field SP1c for prompt settings and input, an advanced settings button, a "Create Message Body" button SP3a which is an execution button for the contents of the item, a generated text display field SP1f, a redo button SP1g, a copy button SP1h, a "Add to Message Creation Screen" button SP3b which indicates that the message will be reflected in the message creation screen / field, and a navigation bar SP2. Furthermore, it is desirable that the generated text display area SP1f and other elements below the "Create Message Body" button SP3a be displayed after the "Create Message Body" button SP3a has been pressed. The reason is that, in terms of the order of operations, the "Create Message Body" button SP3a should be pressed first, and only then should the subsequent information be displayed, making it easier for the user to understand.

[0133] Of these, the text information display field SP1b receives the received message text information ML, which is an example of the text TX specified in the browser 111, as resource information, and is entered by the browser 111. Additionally, the prompt input field SP1c is equipped with a voice input button SP1d and a prompt wizard button SP1e. For example, when the voice input button SP1d is pressed by the user, the microphone on the user terminal 110 is turned on, enabling voice input. When the user voice-inputs "Create a thank-you message," the message "Create a thank-you message" is entered into the prompt input field SP1c.

[0134] Then, when the "Create Message Body" button SP3a is pressed, the user terminal 110 sends prompt information to the first server 120, including the received message text information ML displayed in the text information display field SP1b as an example of resource information, and the intention to create a reply message. Then, the first server 120 generates text data as a reply message using Chat GPT, which is an example of a pre-trained artificial intelligence means 131 of the second server 130, based on the received message text information ML and prompt information. Then, the first server 120 sends the generated text data GT obtained to the user terminal 110. Then, the browser 111 of the user terminal 110 displays the received generated text data GT in the generated text display field SP1f of the side panel SP.

[0135] Furthermore, let's assume that the user has clicked the "Add to Message Creation Screen" button SP3b, which is a button that indicates that the message will be reflected in the message creation screen / field. Then, as shown in Figure 12(B), the browser 111 reflects the generated text data GT displayed in the generated text display field SP1f of the side panel SP into the message input field 117a and displays it. This sets the reply message data (generated text data GT) into the message input field 117a, which is an example of a reply message window, and puts the message in a state just before sending, where only the send button needs to be pressed. As a result, users can easily prepare reply messages.

[0136] The operation and configuration when the "Summarize Chat Content" menu button 112r (to create a summary), the "Find Problems in Chat" menu button 112s (to identify problems), and the "Schedule Meeting" menu button 112d (to adjust meeting time) are operated are the same as in the previously described embodiment. I will omit the detailed explanation. In other words, when a target is specified in the browser 111 that displays the talk screen 117, and the first icon button 111a is operated to select an item menu button, a side panel SP having buttons for the content of that item is displayed, and interaction with Chat GPT, which is an example of a pre-trained artificial intelligence means 131, becomes possible. As a result, users can easily interact with Chat GPT, an example of a pre-trained artificial intelligence means 131, by selecting a desired item menu in the browser 111 and using the content of the desired item in the side panel SP.

[0137] In the artificial intelligence system 100, which is the fourth embodiment of the present invention obtained in this way, the browser 111 displays a talk screen 117 regarding message exchanges between the user and other users, and when the first icon button 111a provided on the browser 111 is operated while one or more message texts, which are received message text information ML, are specified in the browser 111, the browser 111 displays a plurality of item menu buttons (112q, 112r, 112s, 112d), and when one of the plurality of item menu buttons is selected, the browser 111 displays the first server 12 The system displays a side panel SP that can freely show interactions with Chat GPT, an example of a pre-trained artificial intelligence means 131 of the second server 130, via 0. The side panel SP is configured to have buttons (SP3a, SP1m, SP1p, SP1q) for the content of one of several selected item menu buttons. This allows the user to easily interact with Chat GPT, an example of a pre-trained artificial intelligence means 131, by selecting a desired item menu in the browser 111 that displays the talk screen 117 and displaying the content of the desired item in the side panel SP. The effect is significant. [Examples]

[0138] Next, a fifth embodiment of the present invention, the artificial intelligence system 100, will be described based on Figures 13(A) to 14(B). Here, Figure 13(A) shows a predetermined operation to display the side panel SP while a file EX in cell display format is displayed in the browser 111 of the user terminal 110 of the artificial intelligence system 100, which is the fifth embodiment of the present invention; Figure 13(B) shows a predetermined operation to display the side panel SP while prompt information is entered in the side panel SP of the browser 111; Figure 14(A) shows a predetermined operation to display the side panel SP while a file EX in cell display format containing text data STX is displayed in the browser 111 of the user terminal 110 of the artificial intelligence system 100, which is the fifth embodiment of the present invention; and Figure 14(B) shows a predetermined operation to display the side panel SP of the browser 111 while text data STX is analyzed. The artificial intelligence system 100 of the fifth embodiment is the same as the artificial intelligence system 100 of the first embodiment, but the target specified in the browser 111 of the user terminal 110 has been changed from email to a cell-display format file EX. Since many elements are common with the artificial intelligence system 100 of the first embodiment, a detailed explanation of the common items will be omitted.

[0139] In the fifth embodiment of the present invention, the artificial intelligence system 100, as shown in Figure 13(A), the browser 111 of the user terminal 110 displays a cell-format display file display screen 118. This cell format file display screen 118 shows a file EX in cell format, which has multiple cells. Then, when a file EX in a cell display format having multiple cells is specified as an example of a text TX or file FL in the browser 111 of the user terminal 110, and the first icon button 111a provided in the browser 111 is operated as an example of a predetermined operation to display the side panel SP, the browser 111 displays multiple item menu buttons. As an example of multiple menu buttons, there is a "Create Format" menu button 112t for creating a format, a "Run AI Analysis (Category)" menu button 112u for performing an analysis, and a "Run AI Analysis (Includes Negative and Positive Elements)" menu button 112v for analyzing including negative and positive elements. Furthermore, as an example of a predetermined operation to display the side panel SP, let's assume that the "Create Format" menu button 112t is selected as one of several item menu buttons.

[0140] Then, as shown in Figure 13(B), the browser 111 displays the side panel SP. The side panel SP includes, for example, a prompt input field SP1c for setting and inputting prompts, a detailed settings button, an execution button for the item content called "Create Format" button SP3c, a generated text display field SP1f, a re-execution button SP1g, a copy button SP1h, and a button called "Reflect to Cell Sheet" SP3d to reflect the changes in the sheet. Furthermore, it is desirable that the generated text display area SP1f and other elements below the "Create Format" button SP3c be displayed after the "Create Format" button SP3c has been pressed. The reason is that, in terms of the order of operations, the "Create Format" button SP3c should be pressed first, and only then should the subsequent information be displayed, making it easier for the user to understand.

[0141] Then, when a prompt setting / input field, such as the prompt input field SP1c, is entered with the message "Create a travel expense application format," and the "Create format" button SP3c is pressed, the user terminal 110 sends the prompt information to the first server 120. Then, based on the received prompt information, the first server 120 uses ChatGPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, to generate tabular text data as a format for a cell display file. More specifically, since it is a travel expense application format, it generates tabular text data where information such as the travel expense application title, document number, application date, and applicant is entered into each cell.

[0142] Here, the tabular text data (tabular generated text data GT) includes tab characters and other symbols so that data is entered into each cell. In other words, it is tabular text data that is reflected in each cell on the sheet of FileEX when data is pasted into FileEX in cell format. Then, the first server 120 sends the generated tabular text data GT, which is the format for a cell display file obtained, to the user terminal 110. Then, the browser 111 of the user terminal 110 displays the generated text data GT, which is in tabular format as a format for the received cell display file, in the generated text display field SP1f of the side panel SP. Along with this, browser 111 displays a "Reflect to Cell Sheet" button SP3d on the side panel SP, which indicates that the changes will be reflected in the sheet.

[0143] In this way, for a file EX in cell display format specified in browser 111, by operating the "Create Format" button SP3c on the side panel SP, it is possible to easily obtain generated tabular text data GT as a format for a file in cell display format using ChatGPT, which is an example of a pre-trained artificial intelligence means 131. Furthermore, as shown in Figure 13(B), when the "Reflect to Cell Sheet" button SP3d, which indicates that the data should be reflected in the sheet, is operated, the browser 111 is configured to display the generated text data GT in tabular format on a sheet of the file EX in cell display format.

[0144] As a result, when a user is viewing a cell-formatted file EX in a browser 111, and they operate the first icon button 111a, the "Create Format" menu button 112t which is an item menu button indicating the creation of a format, enter the desired prompt information into the prompt input field SP1c, which is an example of a prompt setting / input field, and then operate the "Create Format" button SP3c which is an execution button, generated text data GT in a format corresponding to the prompt information is generated by chat GPT, which is an example of a pre-trained artificial intelligence means 131, and then displayed on the side panel SP. As a result, users can easily prepare the generated text data GT as a format proposal. Furthermore, when the user operates the "Reflect to Cell Sheet" button SP3d, which indicates that the data should be reflected in the sheet, the generated text data GT in tabular format is reflected in the sheet of the file EX, which is in cell display format. As a result, users can easily create files in a cell-based format. For example, you can easily create formats for application forms, quotations, invoices, and other documents.

[0145] Furthermore, as shown in Figure 14(A), the browser 111 of the user terminal 110 displays a file EX in cell display format, which contains text data STX in multiple cells. For example, consider an employee survey, where multiple cells contain text data (STX) representing employee comments. Then, when a single sheet containing multiple cells into which text data STX is entered as an example of a text TX or file FL is specified in the browser 111 of the user terminal 110, and the first icon button 111a provided in the browser 111 is operated as an example of a predetermined operation to display the side panel SP, the browser 111 displays multiple item menu buttons. Furthermore, as an example of a predetermined operation to display the side panel SP, let's assume that the "Execute AI Analysis (Category)" menu button 112u is selected as one of several item menu buttons.

[0146] Then, as shown in Figure 14(B), the browser 111 displays the side panel SP. The side panel SP includes, for example, a text information display field SP1b, an execution button for the item content called "Execute AI Analysis (Category)" SP3e, a generated text display field SP1f, a re-execution button SP1g, a copy button SP1h, and a button to reflect the changes in the sheet called "Reflect in Cell Sheet" SP3d.

[0147] Furthermore, it is desirable that the generated text display area SP1f and other elements below the "Execute AI Analysis (Category Classification)" button SP3e be displayed after the "Execute AI Analysis (Category Classification)" button SP3e has been operated. The reason is that, in terms of the order of operations, the "Execute AI Analysis (Category Classification)" button SP3e should be operated first, and only then should the subsequent items be displayed, making it easier for the user to understand.

[0148] When the side panel SP is displayed or when the display is switched to an item that indicates it will be analyzed by operating the buttons on the navigation bar SP2, the text data STX from multiple cells in a cell-display format file EX is read as an example of a text TX or file FL specified by the browser 111, and the text data STX is displayed as a resource.

[0149] Then, when the "Execute AI Analysis (Category Classification)" button SP3e is pressed, the user terminal 110 sends the text data STX displayed in the text information display field SP1b as an example of resource information, and prompt information indicating that analysis will be performed, to the first server 120. Then, the first server 120 uses ChatGPT, an example of a pre-trained artificial intelligence means 131 of the second server 130, to generate tabular text data as analysis results, based on the text data STX and prompt information of the multiple cells received by the second server 130.

[0150] More specifically, the first server 120 extracts and generates grouped content information GR based on the content of the text data STX of multiple cells received, using chatGPT, which is an example of a pre-trained artificial intelligence means 131. In conjunction with this, the first server 120 uses ChatGPT, which is an example of a pre-trained artificial intelligence means 131, to group the data according to the content information GR for grouping, and obtains tabular generated text data GT that includes the content information GR for grouping, the number of items in each group after grouping NU, and correspondence information RE with the cell number that has the text data STX before grouping.

[0151] Then, the first server 120 sends the generated tabular text data GT, which is the analysis result obtained, to the user terminal 110. Then, the browser 111 of the user terminal 110 displays the generated text data GT, which is the result of the received analysis, in the generated text display field SP1f of the side panel SP. Along with this, browser 111 displays a "Reflect to Cell Sheet" button SP3d on the side panel SP, which indicates that the changes will be reflected in the sheet.

[0152] In this way, for a file EX in cell display format specified in browser 111, by operating the "Execute AI Analysis (Category)" button SP3e on the side panel SP, it is possible to easily obtain tabular generated text data GT as the result of analyzing the text data STX of multiple cells in the file EX in cell display format using chat GPT, which is an example of a pre-trained artificial intelligence means 131. Furthermore, as shown in Figure 14(B), when the "Reflect to Cell Sheet" button SP3d, which indicates that the data should be reflected in the sheet, is operated, the browser 111 is configured to display the generated text data GT in tabular format on a sheet of the file EX in cell display format.

[0153] As a result, when a user operates the first icon button 111a, the "Execute AI Analysis (Category)" menu button 112u which is an item menu button indicating that analysis should be performed, and the "Execute AI Analysis (Category)" button SP3e which is an execution button for the content of the items, the text data STX of multiple cells are grouped according to their content, and the content information GR of the grouping, the number information NU of the items in each group after grouping, and the correspondence information RE with the cell numbers that contained the text data STX before grouping are displayed on the side panel SP. As a result, users can easily prepare grouped analysis results for text data (STX) from multiple cells.

[0154] Furthermore, when the user operates the "Reflect to Cell Sheet" button SP3d, which indicates that the data should be reflected in the sheet, the grouped analysis results data will be reflected in the sheet of the EX file in cell format. As a result, the analysis results can be reflected in a cell-formatted file (EX) and used to analyze text data (STX). For example, if employee survey comments are entered into multiple cells as an example of text data STX, and you operate the first icon button 111a, the menu button 112u which is the item menu button for performing the analysis, and the button SP3e which is the execution button for the item content, the employee survey comments will be grouped into categories based on their content, and the content information GR of the grouping will show what categories there are, the number of comments in each group after grouping will show how many comments there are in that category, and the correspondence information RE which employee survey comments correspond to the cell numbers that contained the text data STX before grouping. As a result, users can easily group and analyze the content of employee survey comments.

[0155] In Figure 14(A), if the first icon button 111a is operated and the "Execute AI Analysis (with Negative / Positive)" menu button 112v is operated, the "Execute AI Analysis (with Negative / Positive)" button will be displayed on the side panel SP. Then, when the "Execute AI Analysis (with Negative / Positive)" button is pressed, ChatGPT, an example of a pre-trained artificial intelligence means 131, determines whether the text data STX of multiple cells is positive or negative based on its content, and then groups them accordingly. Generated text data GT is then created, which includes content information GR for the grouping by positive and negative, information NU for the number of items in each group after grouping, and correspondence information RE with the cell number that contained the text data STX before grouping. The analyzed results, categorized into positive and negative, are then displayed in the generated text display area SP1f on the side panel SP. Further details will be omitted.

[0156] In the fifth embodiment of the present invention, artificial intelligence system 100 obtained in this way, when the browser 111 displays a file EX in a cell display format having multiple cells, and when the first icon button 111a provided on the browser 111 is operated while a single sheet having multiple cells is specified in the browser 111, the browser 111 displays multiple item menu buttons, and the content of one of the item menu buttons is to create a format, and when the "Create Format" menu button 112t, which is the item menu button for creating a format, is operated, the browser 111 displays a prompt input field SP1c as an example of a prompt setting / input field which is an instruction command for chat GPT, which is an example of a pre-trained artificial intelligence means 131, and an execution button for the content of the item, "Create Format" button SP3c on the side panel SP, and the execution button When the "Create Format" button SP3c is operated, the user terminal 110 sends prompt information from the prompt input field SP1c, which is a prompt setting / input field, to the first server 120. The browser 111 then uses Chat GPT, an example of a pre-trained artificial intelligence means 131, to generate text in a tabular format and displays the generated text data GT on the side panel SP. It also displays the "Reflect to Cell Sheet" button SP3d on the side panel SP, which indicates that the data should be reflected on the sheet. When the "Reflect to Cell Sheet" button SP3d is operated, the browser 111 displays the tabular generated text data GT on the sheet of the cell display format file EX. This configuration allows the user to easily prepare the generated text data GT as a format proposal, and furthermore, allows the user to easily create the format of the cell display format file EX.

[0157] Furthermore, when the browser 111 displays a file EX in cell display format containing text data STX in multiple cells, and a single sheet containing multiple cells is specified in the browser 111, and the first icon button 111a provided in the browser 111 is operated, the browser 111 displays multiple item menu buttons, and the content of one of the item menu buttons is "to analyze", and the menu button 112u, which is the item menu button for "to analyze", is operated, the browser 111 displays the text data STX in the file EX in cell display format specified in the browser 111 as a resource in the text information display field SP1b of the side panel SP, and also displays the execution button for the item content, "Execute AI Analysis (Category)" button SP3e, on the side panel SP, and when the execution button "Execute AI Analysis (Category)" button SP3e is operated, the user terminal 110 sends the text data STX of multiple cells and the prompt information indicating "to analyze" to the first server 120, and the first Server 120 extracts and generates grouping content information GR based on the content of the text data STX of multiple cells received by server 120 using Chat GPT, an example of a pre-trained artificial intelligence means 131, and groups the data according to the grouping content information GR. It then sends the generated text data GT in tabular format to the user terminal 110, which includes the grouping content information GR, the number of items in each group after grouping NU, and correspondence information RE with the cell numbers containing the text data STX before grouping. Browser 111 displays the generated text data GT in tabular format, the grouping content information GR, the number of items in each group after grouping NU, and the correspondence information RE with the cell numbers containing the text data STX before grouping on the side panel SP, and also displays a "Reflect to Cell Sheet" button SP3d indicating that the data should be reflected on the sheet. When the "Reflect to Cell Sheet" button SP3d is pressed, Browser 111 displays the generated text data GT in tabular format on the sheet.The ability for users to easily prepare grouped analysis results for text data (STX) from multiple cells, and to reflect the analysis results in a cell-display file (EX) for analyzing the text data (STX), demonstrates its significant effectiveness. [Examples]

[0158] Next, the sixth embodiment of the present invention, the artificial intelligence system 100, will be described based on Figures 15(A) to 15(B). Here, Figure 15(A) shows a predetermined operation to display the side panel SP while the browser 111 of the user terminal 110 of the artificial intelligence system 100, which is the sixth embodiment of the present invention, is displaying the customer relationship management screen 119 of a customer relationship management system, which is an example of the company's own system. Figure 15(B) shows the process of inputting prompt information on the side panel SP of the browser 111 to create an email body from customer relationship management information. The artificial intelligence system 100 of the sixth embodiment is modified in which the target specified in the browser 111 of the user terminal 110 of the artificial intelligence system 100 of the first embodiment is changed from email to customer relationship management information on the customer relationship management screen 119 of the customer relationship management system. Since many elements are common with the artificial intelligence system 100 of the first embodiment, a detailed explanation of the common items will be omitted.

[0159] In the sixth embodiment of the present invention, the artificial intelligence system 100, as shown in Figure 15(A), the browser 111 of the user terminal 110 displays the customer relationship management screen 119 of a customer relationship management system, which is an example of the company's own system. This customer relationship management screen 119 displays customer relationship management information, including customer information and interaction information such as the customer's company information and status information regarding sales-related matters with the customer. Then, when customer relationship management information is specified as an example of a text TX or file FL in the browser 111 of the user terminal 110, and the first icon button 111a provided in the browser 111 is operated as an example of a predetermined operation to display the side panel SP, the browser 111 displays multiple item menu buttons. As an example of multiple menu buttons, there is a "Status Check" menu button 112w for checking the status, a "Create Email Body" menu button 112x for composing an email, a "Request Next Action" menu button 112y for requesting the next action, and a "Schedule Meeting" menu button 112d for scheduling a meeting. Furthermore, as an example of a predetermined operation to display the side panel SP, let's assume that the "Create email body" menu button 112x is selected as one of several item menu buttons.

[0160] Then, as shown in Figure 15(B), the browser 111 displays the side panel SP. The side panel SP includes, for example, a text information display field SP1b, a prompt input field SP1c which is a prompt setting / input field, an advanced settings button, an execution button for the item's content called "Create Email Body" button SP1a, a generated text display field SP1f, a redo button SP1g, a copy button SP1h, and an "Add to Email Creation Screen" button SP1k which indicates that the message will be reflected in the message creation screen / field. Furthermore, it is desirable that the generated text display area SP1f and other elements below the "Create Email Body" button SP1a be displayed after the "Create Email Body" button SP1a has been pressed. The reason is that, in terms of the order of operations, the "Create Email Body" button SP1a should be pressed first, and only then should the subsequent information be displayed, making it easier for the user to understand.

[0161] When the side panel SP is displayed or when the display is switched to the content of the item for creating an email by operating the buttons on the navigation bar SP2, the customer relationship management information is loaded as an example of a text TX or file FL specified by the browser 111 and displayed as a resource in the text information display field SP1b. Customer relationship management information includes email recipient information and contact person information.

[0162] Additionally, the prompt input field SP1c is equipped with a voice input button SP1d and a prompt wizard button SP1e. For example, when the voice input button SP1d is pressed by the user, the microphone on the user terminal 110 is turned on, enabling voice input. When the user voice-inputs "Create a year-end greeting message," the phrase "Create a year-end greeting message" is entered into the prompt input field SP1c.

[0163] Then, when the "Create Message Body" button SP3a is pressed, the user terminal 110 sends prompt information to the first server 120, including customer relationship management information displayed in the text information display field SP1b as an example of resource information, and information about creating a year-end greeting message. Then, the first server 120 generates text data as a message using Chat GPT, which is an example of a pre-trained artificial intelligence means 131 of the second server 130, based on the received customer relationship management information and prompt information. Then, the first server 120 sends the generated text data GT obtained to the user terminal 110. Then, the browser 111 of the user terminal 110 displays the received generated text data GT in the generated text display field SP1f of the side panel SP.

[0164] Furthermore, let's assume that the user has clicked the "Add to Message Creation Screen" button SP3b, which is a button that indicates that the message will be reflected in the message creation screen / field. Then, browser 111 opens the mail sending window and reflects the generated text data GT displayed in the generated text display area SP1f of the side panel SP into the mail sending window for display. This sets the message data (generated text data GT) into an example of a message window / field, such as a mail window for sending, and puts it in a state just before sending, where all that's left is to press the send button. As a result, users can easily prepare messages.

[0165] Furthermore, if the "Check Status" menu button 112w, which indicates the user is checking the status, is pressed, the browser 111 will display the text information display field SP1b and the "Check Status" button, which is an execution button for the item's content, on the side panel SP. Then, when the "Check Status" button is pressed, the user terminal 110 sends prompt information to the first server 120, such as customer relationship management information and status information displayed in the text information display field SP1b as an example of resource information. Then, the first server 120 generates text data as a response using ChatGPT, which is an example of a pre-trained artificial intelligence means 131 of the second server 130, based on the received customer relationship management information and prompt information. Then, the first server 120 sends the generated text data GT obtained to the user terminal 110. Then, the browser 111 of the user terminal 110 displays the received generated text data GT in the generated text display field SP1f of the side panel SP.

[0166] Furthermore, if the "Request Next Action" menu button 112y, which indicates the user is requesting a suggestion for the next action, is clicked, the browser 111 displays the text information display field SP1b and the "Request Next Action" button, which is an execution button for the item's content, on the side panel SP. When the "Propose Next Action" button is operated, the user terminal 110 transmits to the first server 120 the customer relationship management information displayed in the text information display column SP1b as an example of resource information, and prompt information regarding proposing the next action and the like. Then, based on the received customer relationship management information and prompt information, the first server 120 uses ChatGPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, to execute the generation of text data as a proposal. Then, the first server 120 transmits the obtained generated text data GT to the user terminal 110.

[0167] Then, the browser 111 of the user terminal 110 displays the received generated text data GT in the generated text display column SP1f of the side panel SP. The operations and configurations when the "Meeting Time Adjustment" menu button 112d for adjusting the meeting time is operated are the same as those in the foregoing embodiments. A detailed description thereof is omitted.

[0168] As a result, when a target is specified in the browser 111 that displays the customer relationship management screen 119 and the first icon button 111a is operated and an item menu button is selected, a side panel SP having buttons for the content of that item is displayed, enabling free interaction with ChatGPT, which is an example of the pre-trained artificial intelligence means 131. As a result, the user can easily interact with ChatGPT, which is an example of the pre-trained artificial intelligence means 131, with the content of the desired item in the side panel SP by selecting the desired item menu in the browser 111.

[0169] In the artificial intelligence system 100, which is the sixth embodiment of the present invention obtained in this way, the browser 111 displays a customer relationship management screen 119 having customer information and customer relationship management information regarding interactions with customers, and when the first icon button 111a provided on the browser 111 is operated while customer information and interaction information are specified in the browser 111, the browser 111 displays a plurality of item menu buttons (112w, 112x, 112y, 112d), and when one of the plurality of item menu buttons (112w, 112x, 112y, 112d) is selected, the browser 111 displays the first server The browser 111 displays a side panel SP that can freely show interactions with Chat GPT, an example of a pre-trained artificial intelligence means 131 of the second server 130, via the browser 120. The side panel SP is configured to have buttons (SP3a, SP1q, etc.) for the content of one of several selected item menu buttons (112w, 112x, 112y, 112d). As a result, the user can select a desired item menu in the browser 111 and easily interact with Chat GPT, an example of a pre-trained artificial intelligence means 131, with the content of the desired item in the side panel SP, thus achieving significant benefits. [Examples]

[0170] Next, the seventh embodiment of the present invention, the artificial intelligence system 100, will be described based on Figures 16(A) to 18(B). Here, Figure 16(A) shows the selection of an image range in the browser 111 of the user terminal 110 of the artificial intelligence system 100, which is the seventh embodiment of the present invention; Figure 16(B) shows the state when a question is asked about the content of the image; Figure 17(A) shows the reading of the text of the web page displayed in the browser 111 of the user terminal 110 of the artificial intelligence system 100, which is the seventh embodiment of the present invention; Figure 17(B) shows the state when a question is asked about the content of the web page; Figure 18(A) shows the reading of the text of the selected area of ​​the web page displayed in the browser 111 of the user terminal 110 of the artificial intelligence system 100, which is the seventh embodiment of the present invention; and Figure 18(B) shows the state when a question is asked about the content of the text of the selected area. The artificial intelligence system 100 of the seventh embodiment is a modified version of the artificial intelligence system 100 of the first embodiment, with changes to the target specified in the browser 111 of the user terminal 110 and the method of specification. Since many elements are common with the artificial intelligence system 100 of the first embodiment, a detailed explanation of the common items will be omitted.

[0171] In the seventh embodiment of the present invention, artificial intelligence system 100, as shown in Figure 16(A), the navigation bar SP2 is provided with a question button SP2g for images, a question button SP2h for the open content, and a question button SP2k for the selected content. The browser 111 on the user terminal 110 is displaying a website of choice. Next, let's assume that the user has selected the question button SP2g in the image of the navigation bar SP2. Then, the content of the side panel SP changes to questions presented in images.

[0172] Here, the contents of the side panel SP were switched by operating the buttons on the navigation bar SP2. As mentioned earlier, when the first icon button 111a in the browser 111 is operated, the browser 111 displays multiple item menu buttons. As an example of multiple menu buttons, there may be a "Ask a question with an image" menu button for asking a question with an image, a "Ask a question about the currently open content" menu button for asking a question about the currently open content, and a "Ask a question about the selected content" menu button for asking a question about the selected content. And of course, when those menu buttons are selected, the browser 111 may be configured to display the side panel SP containing their contents.

[0173] When the "Ask a Question with an Image" button SP2g is selected, the side panel SP includes, for example, an image information display field SP3f as an example of a text / image information display field, a prompt input field SP1c, and an execution button, the "Ask a Question with an Image" button SP3g. Then, let's assume that the user has selected an image of a graph from any website displayed in browser 111. Next, let's assume that the image information display field SP3f is selected and the paste operation is performed. Then, as shown in Figure 16(B), the image of the selected graph is entered into the image information display field SP3f. Alternatively, when an image is selected, the browser 111 may be configured to input the selected image into the image information display field SP3f without requiring a paste operation.

[0174] Then, suppose the user enters a prompt message in the prompt input field SP1c asking "Please tell me the trend of the graph," and the "Ask a question with an image" button SP3g is pressed. The user terminal 110 then sends the image displayed in the image information display field SP3f as an example of resource information, along with a prompt message stating "Please tell me the trend of the graph," to the first server 120. Based on the received image and prompt information, the first server 120 uses ChatGPT, an example of a pre-trained artificial intelligence means 131 of the second server 130, to generate text data about the content of the image. In this case, ChatGPT, an example of a pre-trained artificial intelligence means 131, recognizes the content of an image using an image recognition means. Then, the first server 120 sends generated text data GT about the content of the obtained image to the user terminal 110.

[0175] Then, the browser 111 of the user terminal 110 displays the generated text data GT about the content of the received image in the generated text display field SP1f of the side panel SP. In this way, by operating the "Ask a question with an image" button SP3g on the side panel SP for the image specified in the browser 111, it is possible to easily obtain generated text data GT about the content of the image using Chat GPT, which is an example of a pre-trained artificial intelligence means 131.

[0176] Furthermore, as shown in Figure 17(A), assume that the user has selected the question button SP2h in relation to the content currently displayed in the navigation bar SP2. Then, the content of the side panel SP changes to a question about what is currently open. Here, the content of the side panel SP was switched by operating the buttons on the navigation bar SP2. However, as mentioned above, the side panel SP with the relevant content may also be displayed by operating the first icon button 111a in the browser 111 and then operating the displayed "Ask a question about the open content" menu button.

[0177] When the question button SP2h is selected for the open content, the side panel SP includes, for example, a "Load URL page" button SP3h, a text information display field SP1b, a prompt input field SP1c, and an execution button, the "Ask a question" button SP3k. Let's assume that the user has activated the "Load URL Page" button SP3h. Browser 111 then inputs the text data of the displayed website into the text information display field SP1b.

[0178] Next, as shown in Figure 17(B), suppose the user enters prompt information to the effect of "Please tell me the details" into the prompt input field SP1c, and the "Ask a question" button SP3k is pressed. The user terminal 110 then sends the text data displayed in the text information display field SP1b, which is an example of resource information, and the prompt information "Please tell me the contents" to the first server 120. Next, the first server 120, based on the received text data and prompt information, uses ChatGPT, an example of the pre-trained artificial intelligence means 131 of the second server 130, to generate text data about the content of the resource text (website) displayed in the text information display field SP1b. Then, the first server 120 sends the generated text data GT about the content of the website it obtained to the user terminal 110.

[0179] Then, the browser 111 of the user terminal 110 displays the generated text data GT about the content of the received resource text in the generated text display field SP1f of the side panel SP. In this way, by operating the "Ask a Question" button SP3k on the side panel SP, it is possible to easily obtain generated text data GT about the content of a website specified in the browser 111 using Chat GPT, which is an example of a pre-trained artificial intelligence means 131.

[0180] Furthermore, as shown in Figure 18(A), assume that the user has selected the question button SP2k based on the selection in the navigation bar SP2. Then, the content of the side panel SP changes to a question asking about your selections. Here, the content of the side panel SP was switched by operating the buttons of the navigation bar SP2. However, as described above, when the first icon button 111a in the browser 111 is operated and the displayed "Ask about the selection" menu button is operated, the side panel SP with the corresponding content may be displayed.

[0181] When the "Ask about the selection" button SP2k is selected and operated, the side panel SP is provided with, as an example, a text information display column SP1b, a prompt input column SP1c, and a "Ask" button SP3k which is an execution button. And suppose that the text of the website displayed by the browser 111 is selected by a drag operation or the like according to the user's operation. Then, the browser 111 inputs the selected text data of the displayed website into the text information display column SP1b.

[0182] Subsequently, as shown in FIG. 18(B), suppose that the user inputs prompt information to the effect of "Please tell me the content" into the prompt input column SP1c and operates the "Ask" button SP3k. Then, the user terminal 110 transmits the text data displayed in the text information display column SP1b, which is an example of resource information, and the prompt information to the effect of "Please tell me the content" to the first server 120. Subsequently, based on the received text data and prompt information, the first server 120 uses ChatGPT, which is an example of the pre-trained artificial intelligence means 131 of the second server 130, to generate text data about the content of the text (selected part of the website) of the resource displayed in the text information display column SP1b. Then, the first server 120 transmits the generated text data GT about the content of the selected part of the obtained website to the user terminal 110.

[0183] Then, the browser 111 of the user terminal 110 displays the generated text data GT about the content of the selected section of the received website in the generated text display field SP1f of the side panel SP. In this way, by operating the "Ask a Question" button SP3k on the side panel SP, it is possible to easily obtain generated text data GT about the content of the selected area of ​​the website specified in the browser 111 using Chat GPT, which is an example of a pre-trained artificial intelligence means 131. [Explanation of Symbols]

[0184] 100 ··· Artificial Intelligence System 110 ··· User terminal 111 ··· Browser 111a... First icon button 111b... Side panel icon button 111c... Second icon button 112a... "Create Reply Email Document" menu button 112b... "Email Summary" menu button 112c... "Find problems in email" menu button 112d... "Schedule a meeting" menu button 112e... "Summary of the entire thread" menu button 112f... "Discover problems throughout the thread" menu button 112g... "Create Daily Report" menu button 112h... "Create Weekly Report" menu button 112k... "Create Monthly Report" menu button 112m... "RAG Learning" menu button 112n... "Create Summary" menu button 112p... "Ask a question" menu button 112q... "Create reply chat document" menu button 112r... "Summary of Chat Content" menu button 112s... "Discover problems in chat" menu button 112t... "Create Format" menu button 112u... "Execute AI Analysis (Category)" menu button 112v... "Execute AI Analysis (Negative / Positive Results Included)" menu button 112w... "Status Check" menu button 112x... "Create email body" menu button 112y... Menu button for "Request next action" 113... Reply email window (reply message window / field) 114 ··· Provisional Reservation Confirmation Window 115 ··· Calendar / Schedule Screen 116... Drive display screen 117... Talk screen (chat screen) 117a... Input field for sending message (Reply message window field) 118 ··· Cell format display file display screen 119 ··· Customer Relationship Management Screen 120 ··· Server 1 130 ··· Server 2 131 ··· Pre-trained artificial intelligence methods (large-scale language models, chat GPT) SP... Side panel (designated area panel) SP1a... "Create email body" button SP1b... Text information display area (text and image information display area) SP1c... Prompt input field (Prompt settings / input field) SP1d... Voice input button SP1e... Prompt Wizard Button SP1f... Generated text display field SP1g... Retry button SP1h... Copy button SP1k... "Add to email composition screen" button (a button indicating that the message will be reflected in the message composition screen / field) SP1m... "Summarize Document" button SP1n... Length setting field (prompt setting / input field) SP1p... "Identify Problems" button SP1q... "Schedule a meeting" button SP1r... Candidate date display field (generated text display field) SP1s... Selection field for tentatively booking an event on the calendar SP1t... Calendar display link button SP1u... "Create Report" button SP1v... "Submit as file" button (Daily report file creation button) SP1w... "Select recipient from address book" button SP1x... Opponent information display field SP1y... "RAG Learning" button SP1z... Question input field (prompt settings input field) SP2 ··· Navigation bar SP2a... Email body creation button SP2b... Email body summary creation button SP2c... Create entire thread summary button SP2d... Problem Detection Button SP2e... Edit sent email document button SP2f... Meeting time scheduling button SP2g... Question button with image SP2h... Question button about the open content SP2k... Question button regarding selection SP3a... "Create Message Body" button SP3b... "Add to message creation screen" button SP3c... "Create Format" button SP3d... "Reflect to Cell Sheet" button (a button indicating that the changes will be reflected in the sheet) SP3e... "Execute AI Analysis (Category Classification)" button SP3f... Image information display area (text and image information display area) SP3g... "Ask a question with an image" button SP3h... "Load URL page" button SP3k... "Ask a question" button TX ··· Text FL ··· File ML... Received message text information MT... Meeting candidate date and time information GT ··· Generated text data (draft reply message information) KSJ ··· Provisional Reservation Schedule Information SD ··· Schedule Information EX... File in cell display format STX ··· Text data GR ··· Group division information NU ··· (Count information within each group after grouping) RE ··· Correspondence information (with the cell number containing the text data before grouping)

Claims

1. An artificial intelligence system comprising a user terminal and a server, wherein the user terminal displays or outputs audio based on the server, An artificial intelligence system characterized in that, when a predetermined operation occurs while at least one of text, images, and files is displayed or selected in the browser of the user terminal, the browser displays a predetermined area panel that can display interactions with the pre-trained artificial intelligence means of the server.

2. As a predetermined operation, when an icon button provided in the browser is operated, the browser displays a number of item menu buttons. When one of the aforementioned multiple item menu buttons is selected, the browser displays a predetermined area panel that can display the interaction with the server's pre-trained artificial intelligence means. The artificial intelligence system according to claim 1, characterized in that the predetermined area panel has a button for the content of one of the selected item menu buttons.

3. The artificial intelligence system according to claim 2, characterized in that the predetermined area panel has a configuration that includes a text and image information display field designated as a resource in the browser, a prompt setting / input field which is an instruction command for a pre-trained artificial intelligence means, and an execution button for the contents of the item.

4. The content of one of the aforementioned menu buttons is "Meeting time scheduling". When the menu button for scheduling the meeting is operated, the browser displays the received message text information specified as a resource in the browser in the text information display field of the designated area panel, and also displays an execution button for scheduling the meeting on the designated area panel. When the execution button is operated, the browser accesses the schedule management system server based on pre-configured user information to obtain schedule information for the user on the user terminal, and uses the server's pre-trained artificial intelligence means to pick up multiple meeting candidate date and time information based on the schedule blank information in the schedule information, displays it in a predetermined area panel, and also displays a reply message creation button. When the reply message creation button is pressed, the user terminal sends the received message text information, multiple meeting date and time information, and a prompt to create a reply message to the server's pre-trained artificial intelligence means. The browser displays the reply message draft information generated by the pre-trained artificial intelligence means in a predetermined area panel and also displays a button to reflect it in the message creation screen / field. The artificial intelligence system according to claim 2 or 3, characterized in that when a button indicating that the message should be reflected in the message creation screen / field is operated, the browser displays the draft reply message information in the reply message window / field.

5. When the browser displays the reply message draft information generated by the pre-trained artificial intelligence means in a predetermined area panel and displays a button to reflect it in the message creation screen / field, it also displays an option to tentatively book an appointment in the user's schedule. The artificial intelligence system according to claim 4, characterized in that when the user selects to tentatively book an appointment in the user's schedule and the button to reflect it in the message creation screen / field is operated, the user terminal transmits information to the schedule management system server indicating that multiple candidate meeting dates and times have been tentatively booked.

6. The aforementioned browser displays the user's schedule information for today based on the user's schedule information on the schedule management system server, and the content of one of the multiple menu buttons is "Create a daily report". When the menu button for creating the daily report is operated, the browser displays the schedule information for today, which was specified in the browser as a resource, in the text information display field of the designated area panel, and also displays an execution button for creating the daily report. When the execution button is pressed, the user terminal sends today's schedule information and a prompt to create a daily report to the server's pre-trained artificial intelligence means. The browser displays the daily report draft information generated by the pre-trained artificial intelligence means in a designated area panel, and also displays a button in the designated area panel indicating that it will be reflected in the message creation screen / field, or a button to create a daily report file. When the button indicating that the message should be reflected in the message creation screen / field is pressed, the browser displays the daily report draft information in the message window / field. Alternatively, the artificial intelligence system according to claim 2 or 3, characterized in that when the daily report file creation button is operated, the browser outputs a file containing the daily report draft information to a pre-set location.

7. The browser displays the user's weekly schedule information based on the user's schedule information on the schedule management system server, and the content of one of the multiple menu buttons is "Create weekly report". When the menu button for creating the weekly report is operated, the browser displays the weekly schedule information specified as a resource in the browser in the text information display field of the designated area panel, and also displays an execution button for creating the weekly report. When the execution button is operated, the user terminal sends a prompt to the server's pre-trained artificial intelligence means indicating that it will create a weekly schedule and a weekly report, and the browser displays the weekly report draft information generated by the pre-trained artificial intelligence means in a predetermined area panel, and also displays a button in the predetermined area panel indicating that it will be reflected in the message creation screen / field or a weekly report file creation button. When the button indicating that the message should be reflected in the message creation screen / field is pressed, the browser displays the weekly report draft information in the message window / field. The artificial intelligence system according to claim 2 or 3, characterized in that when the weekly report file creation button is operated, the browser outputs a file containing the weekly report draft information to a pre-set location.

8. The browser displays the user's schedule information for a specified period based on the user's schedule information on the schedule management system server, and the content of one of the multiple menu buttons is related to scheduling a meeting. When the menu button for scheduling the meeting is operated, the browser displays an execution button for scheduling the meeting on a designated area panel. When the execution button is operated, the user terminal accesses the schedule management system server based on pre-configured user information to obtain schedule information for the user of the user terminal, and uses the server's pre-trained artificial intelligence means to pick up multiple candidate meeting date and time information based on the schedule blank information in the schedule information, and the browser displays the multiple candidate meeting date and time information in a predetermined area panel and also displays a message creation button in a predetermined area panel. When the message creation button is pressed, the user terminal sends multiple meeting date and time options, and a prompt to the server's pre-trained artificial intelligence means indicating that it will create a message for the purpose of scheduling a meeting. The browser displays the message draft information generated by the pre-trained artificial intelligence means in a predetermined area panel and also displays a button in the predetermined area panel indicating that it will be reflected in the message creation screen / field. The artificial intelligence system according to claim 2 or 3, characterized in that when a button indicating that the message should be reflected in the message creation screen / field is operated, the browser displays the message draft information in the message window / field.

9. When one or more files are specified in the browser and an icon button provided in the browser is operated, the server extracts language character data as an electronic information source from the specified one or more files, divides it into multiple divided language character file data, converts each of the divided divided language character file data into vector coordinate system divided language data using a vector coordinate transformation means which is a pre-trained artificial intelligence means, and stores the converted vector coordinate system divided language data in vector indexes in association with the original divided language character file data. The aforementioned browser displays multiple item menu buttons, and the content of one of the items in the menu buttons is either a summary or a question. When the menu button for summarizing or asking a question is operated, the browser displays a text information display field designated as a resource in the browser, a prompt setting / input field for instructions and commands to the pre-trained artificial intelligence means, and an execution button for the content of summarizing or asking a question on a predetermined area panel. When the execution button is pressed, the user terminal sends prompt information to the server as an instruction command. The server generates text data using pre-trained artificial intelligence means and sends the generated text data to the user terminal. The artificial intelligence system according to claim 2 or 3, characterized in that the browser of the user terminal is configured to display the received generated text data in a predetermined area panel.

10. The aforementioned browser displays a chat screen showing the message exchange between the user and other users. When one or more message text groups are specified in the browser, and an icon button provided in the browser is operated, the browser displays multiple item menu buttons. When one of the aforementioned multiple item menu buttons is selected, the browser displays a predetermined area panel that can display the interaction with the server's pre-trained artificial intelligence means. The artificial intelligence system according to claim 2 or 3, characterized in that the predetermined area panel has a button for the content of one of the selected item menu buttons.

11. The browser displays a file in a cell display format that has multiple cells, When a single sheet containing multiple cells is selected in the aforementioned browser, and an icon button provided in the browser is operated, the browser displays multiple item menu buttons, and the content of one of the items in the multiple item menu buttons is to create a format. When the menu button for creating the aforementioned format is operated, the browser displays a prompt setting / input field, which is an instruction command for the pre-trained artificial intelligence means, and an execution button for the content of the item, on a predetermined area panel. When the execution button is pressed, the user terminal sends the prompt information from the prompt setting / input field to the server. The browser uses pre-trained artificial intelligence to generate text in a tabular format, displays the generated text data in a predetermined area panel, and also displays a button in the predetermined area panel indicating that it will be reflected in the sheet. The artificial intelligence system according to claim 2 or 3, characterized in that when a button indicating that the data should be reflected in the sheet is pressed, the browser displays the generated text data in tabular format on the sheet.

12. The browser displays a file in cell format containing text data in multiple cells, When a single sheet containing multiple cells is selected in the aforementioned browser, and an icon button provided in the browser is operated, the browser displays multiple item menu buttons. The content of one of the aforementioned menu buttons is "to perform analysis". When the menu button for the item to perform the analysis is operated, the browser displays the text data of the cell display format file specified in the browser as a resource in the text information display field of the predetermined area panel, and also displays an execution button for the item's content in the predetermined area panel. When the execution button is pressed, the user terminal sends text data from multiple cells and prompt information indicating that it will perform analysis to the server. The server extracts and generates grouping information based on the content of the text data of multiple cells received using a pre-trained artificial intelligence means, groups the data according to the grouping information, and transmits the grouping information, the number of items in each group after grouping, and the correspondence information with the cell numbers containing the text data before grouping to the user terminal. The browser displays the grouping content information, the number of items in each group after grouping, and the correspondence information with the cell numbers containing the text data before grouping in a designated area panel, and also displays a button to reflect this on the sheet. The artificial intelligence system according to claim 2 or 3, characterized in that when a button indicating that the data should be reflected in the sheet is pressed, the browser displays the generated text data in tabular format on the sheet.

13. The aforementioned browser displays a customer relationship management screen containing customer information and interactions with the customer. When customer information and interaction information are specified in the browser, and an icon button provided in the browser is operated, the browser displays multiple item menu buttons. When one of the aforementioned multiple item menu buttons is selected, the browser displays a predetermined area panel that can display the interaction with the server's pre-trained artificial intelligence means. The artificial intelligence system according to claim 2 or 3, characterized in that the predetermined area panel has a button for the content of one of the selected item menu buttons.

14. A program for an artificial intelligence system that displays or outputs audio based on a server on a user terminal, An icon operation determination step that determines whether an icon button provided in the browser has been operated while at least one of text, images, and files is displayed or selected in the browser of the user terminal, When the aforementioned icon button is operated, the browser displays a plurality of item menu buttons and determines whether or not one of the plurality of item menu buttons has been selected; A program for an artificial intelligence system characterized in that, when one of the multiple item menu buttons is selected, the browser causes the computer to execute a panel display step, which displays a predetermined area panel having a button for the content of one of the selected item menu buttons, and which allows interaction with the server's pre-trained artificial intelligence means to be displayed.

Images