AI answering system

Abstract

To provide a system capable of properly giving instructions to subordinates, communicating with subordinates, or improving relationships among subordinates, the following structure is employed: a response system that generates responses to inquiries, has a learning unit that performs machine learning based on inquiries, general information, responses, and satisfaction indices regarding the responses, and has a response generation unit that generates responses to inquiries from a plurality of mobile terminals and transmits the responses to the respective mobile terminals based on the learning results of the learning unit.

Description

Technical Field

[0001] This invention relates to AI response systems, and more specifically, to techniques for effectively utilizing AI constructed from information from multiple information groups. Background Technology

[0002] In the past, as companies grew, they would establish numerous departments to specialize in different functions. This generally led to problems such as time-consuming information transmission and delayed decision-making.

[0003] As a specific example of the drawbacks, sometimes section chiefs explain things to department heads, who then make adjustments between departments, leaving subordinates to wait for the decision. Additionally, the number of consultations sometimes increases, making it impossible to have enough meeting rooms, and the increased absence of superiors due to numerous meetings deprives subordinates of opportunities to report, communicate, and discuss. Furthermore, if subordinates work under a superior who is not adept at coordinating with other departments, they are often preoccupied with inter-departmental adjustments and sometimes unable to perform their original duties.

[0004] Furthermore, sometimes when superiors lack the qualifications to be managers, subordinates are tormented by their unpredictable behavior or fail to receive appropriate business instructions. In such cases, subordinates may hesitate when reporting to, communicating with, or negotiating with their superiors, and may even develop resentment towards them. Conversely, the same situation can occur when subordinates have lower communication skills or other abilities.

[0005] Furthermore, even when management identifies important matters that should be discussed promptly, they are sometimes delayed or concealed for the reasons mentioned above. For sound management, problems and failures within the company can be valuable opportunities for improvement, but employees (supervisors and subordinates) concerned about their position sometimes fail to report them for these reasons. This represents a significant loss for the company.

[0006] Moreover, as a way to connect the intentions and ideas of founders and retired veterans to the future, one can cite corporate philosophy, education, etc., but sometimes these alone are not enough.

[0007] In addition, one can imagine scenarios where communication partners are foreigners or located overseas, and response delays caused by language differences, time differences, or remote locations hinder smooth business operations.

[0008] As a result, work efficiency decreases, and the appeal of the work and the organization diminishes. Moreover, in situations like the superior-subordinate relationship within such companies, where communication with others is necessary, because the other party is a person, it is always necessary to understand their free time, their mood, and to consider their position, personality, language, time zone, location, etc.

[0009] Therefore, regardless of the superior's qualifications or the size of the organization, there is a need for an appropriate system for giving instructions to subordinates, communicating with subordinates, or improving relationships among subordinates.

[0010] In addition, there are situations where there is no suitable person to discuss matters with in private, thus preventing the ability to ask questions or engage in discussions. Therefore, there is a need for a system that allows for appropriate discussions and questioning regardless of the questioner's circumstances.

[0011] Various technologies have been proposed to address this problem. For example, a system has been proposed that utilizes all information, proposals, and improvement plans obtained by all employees for unified management and operation (see Patent Document 1). However, there is no record of communication between employees and managers, thus failing to solve the problem of the present invention.

[0012] Patent document 1: Japanese Patent Application Publication No. 2006-209602. Summary of the Invention

[0013] This invention addresses the problems of inadequate instruction, communication, and relationship improvement among subordinates by effectively utilizing AI and big data to perform tasks previously handled by management (humans). In other words, it solves the problem through a system that responds to inquiries from subordinates based on the learning results of a machine learning-based learning system.

[0014] The AI ​​answering system involved in this invention is an answering system that generates answers to inquiries. Its means is that it has a learning unit and an answer generation unit. The learning unit performs machine learning based on the inquiry, general information, the answer, and the satisfaction index of the answer. Based on the learning results of the learning unit, the answer generation unit generates answers to be sent to each mobile terminal for inquiries from multiple mobile terminals.

[0015] In addition, the present invention utilizes machine learning based on internal company information, in addition to general information.

[0016] Furthermore, the means of the present invention is that the inquiry includes at least any one of the following: report, contact, discussion, opinion, hope, question, worry, dream, likes and dislikes, and self-introduction.

[0017] In addition, the present invention provides that the internal information of the company includes at least one of the following: corporate philosophy, corporate culture, internal regulations, policies, history, contracts, specifications, procedures, stories of past successful employees, customer information, business information, travel reports of business travelers, minutes of internal and external meetings, and employee competency information (qualifications, experience, performance records).

[0018] In addition, the present invention means that the general information includes at least one of the following: biographies, the words of successful people, advisors, self-help books, professional books, comedy studies, humor studies, current affairs materials, Mandarin, arithmetic, science, society, English, laws, departmental orders, ISO, JIS and other standard specifications, and intellectual property information (patents, trademarks, copyrights, etc.).

[0019] Furthermore, the means of the present invention is that the response includes at least one of the following: instruction, suggestion, teaching, guidance, praise, criticism, encouragement, improvement of human comprehensive ability, improvement of the comprehensive ability of social members, business communication, and data production support.

[0020] Furthermore, the present invention allows administrators or users to select a portion of general information.

[0021] Furthermore, the present invention allows for the selection of the respondent's appearance and the expression of their response from multiple roles.

[0022] Furthermore, the present invention involves an answer generation unit that generates an answer based on a weighted average of each piece of information.

[0023] Furthermore, the present invention provides that the response generation unit generates responses in the form of a multi-person dialogue.

[0024] Furthermore, the present invention has the means of having an interpersonal adjustment unit that adjusts interpersonal relationships and business content when the content of the inquiry is about interpersonal relationships, and notifies the inquirer and relevant parties of the answer to adjust interpersonal relationships.

[0025] Furthermore, the present invention includes an importance learning unit that performs machine learning on inquiries, internal company information, general information, and an importance index for the inquiries. Based on the learning results of this importance learning unit, it selects inquiries with higher importance from multiple mobile terminals and notifies the manager.

[0026] The AI ​​response system according to the present invention can always provide appropriate answers to inquiries, thereby improving user peace of mind and satisfaction.

[0027] Furthermore, the AI ​​response system according to the present invention, when used as a business system, facilitates the rationalization and efficiency of business operations, and enables the cultivation of user talent based on employee education simply by using it. Attached Figure Description

[0028] Figure 1 This is a system diagram of an embodiment of the AI ​​response system involved in this invention.

[0029] Figure 2 This is a schematic diagram illustrating the learning process of an embodiment of the AI ​​response system involved in this invention.

[0030] Figure 3 This is a flowchart of an embodiment of the AI ​​response system involved in the present invention.

[0031] Figure 4 This is a diagram showing the query and answer of an embodiment of the AI ​​answering system involved in this invention.

[0032] Figure 5 This is a flowchart of an embodiment of the AI ​​response system involved in this invention, specifically the interpersonal adjustment system.

[0033] Figure 6 This is a display diagram of the interpersonal adjustment system's inquiry and response, representing an embodiment of the AI ​​response system involved in this invention.

[0034] Figure 7 This is a flowchart of an important notification system, which is an embodiment of the AI ​​response system involved in this invention.

[0035] Figure 8 This is a display diagram of the interpersonal adjustment system's inquiry and response, representing an embodiment of the AI ​​response system involved in this invention.

[0036] Figure 9 This is a flowchart illustrating the selection process from multiple AIs in an embodiment of the AI ​​response system of the present invention.

[0037] Figure 10 This is a schematic diagram illustrating an embodiment of the AI ​​response system according to the present invention, in the case of selecting from multiple AIs.

[0038] Figure 11 This is a flowchart and a display diagram of the answers when there are multiple respondents in an embodiment of the AI ​​answering system involved in this invention. Detailed Implementation

[0039] The most significant feature of the AI ​​response system involved in this invention is that it can provide appropriate answers to inquiries regardless of the environment or superior.

[0040] The embodiments of the AI ​​response system according to the present invention will be described below with reference to the accompanying drawings.

[0041] Furthermore, the overall structure and components of the AI ​​response system shown in this embodiment are not limited to the following embodiments, and can be appropriately modified within the scope of the technical concept of the present invention, that is, within the scope of structures, usage forms, etc. that can achieve the same effect.

[0042] Example 1

[0043] according to Figures 1 to 8The present invention will be described below.

[0044] Figure 1 This is a system diagram of an embodiment of the AI ​​response system involved in this invention. Figure 2 This is a schematic diagram illustrating the learning process of an embodiment of the AI ​​response system involved in this invention. Figure 3 This is a flowchart of an embodiment of the AI ​​response system involved in the present invention. Figure 4 This is a diagram showing the query and answer of an embodiment of the AI ​​answering system involved in this invention. Figure 5 This is a flowchart of an embodiment of the AI ​​response system involved in this invention, specifically the interpersonal adjustment system. Figure 6 This is a display diagram of the interpersonal adjustment system's inquiry and response, representing an embodiment of the AI ​​response system involved in this invention. Figure 7 This is a flowchart of an important notification system, which is an embodiment of the AI ​​response system involved in this invention. Figure 8 This is a display diagram of the interpersonal adjustment system's inquiry and response, representing an embodiment of the AI ​​response system involved in this invention.

[0045] AI Response System 1 is a system that responds to inquiries from end users using AI. When used in enterprises, it can be used to answer inquiries from employees. When used by the public, it can be used to provide friendly, conversational responses.

[0046] Both systems consist of an AI-powered response system that answers inquiries and an interpersonal adjustment system that adjusts relationships based on the inquiries. When used in enterprises, an additional system can be added to notify managers of important matters based on the importance of the inquiry.

[0047] The three systems will take appropriate actions based on the content of the inquiry.

[0048] Inquiries include reports, communications, discussions, opinions, hopes, questions, worries, dreams, likes and dislikes, and self-introductions. If it's a company system, reports and communications are the main content.

[0049] Inquirers advance business by receiving responses that correspond to their inquiries, such as instructions, suggestions, teachings, guidance, praise, criticism, business communication, and support in data preparation. Furthermore, in negotiations, they increase their motivation by receiving responses that encourage them and enhance their overall capabilities as members of society.

[0050] Inquiries from employees and the general public often include hopes, questions, worries, dreams, likes and dislikes, and self-introductions. There are situations where the inquirer hopes to ask someone or receive advice from someone. The inquirer is satisfied by receiving advice, instruction, guidance, praise, criticism, encouragement, and content that improves their overall abilities and the overall capabilities of society.

[0051] Alternatively, it's conceivable that during the inquiry, the information could be anonymously sent from the AI ​​to the administrator 82, based on the content, thus gathering accurate information for the administrator 82. However, even if the inquirer's information is anonymous to the administrator 82, it is preferable to retain it as a record in the AI. Even with anonymity, the inquirer's findings, reported achievements, etc., can still be recorded in the AI. Furthermore, it's also possible that the inquirer can choose whether to remain anonymous during the inquiry.

[0052] The system responds to inquiries such as those related to communication, negotiation, and general questions, using AI to provide answers based on general information. The AI ​​learns in advance through deep learning and other methods based on relevant information. The inquirer receives an appropriate answer instantly, within the scope of the AI's learning.

[0053] The interpersonal adjustment system operates during consultations with inquirers regarding their interpersonal relationships. It provides advice on interpersonal troubles and conflicts, and, where possible, contacts and advises the individuals involved, thereby attempting to improve the relationships. In the case of a business-related system, adjustments include reassessing business responsibilities.

[0054] The important matters notification system uses AI to determine whether information and communications from inquirers contain content that the company considers important. If the information is deemed important, the system notifies the management of the content.

[0055] The AI ​​response system 1 consists of an AI response unit 10 and multiple mobile terminals 70. Furthermore, the AI ​​response unit 10 is connected to the external internet 80, internal company information 81, and management 82. The AI ​​response unit 10 and the mobile terminals 70 are connected wirelessly via a dedicated communication network or the internet. In addition, the AI ​​response system 1 can also connect to a desktop PC.

[0056] The AI ​​Response Unit 10 is the main part of the system, consisting of the Terminal Communication Unit 11, the Information Storage Unit 12, the External Communication Unit 13, the Inquiry Acceptance Unit 20, the Learning Unit 30, the Response Generation Unit 40, and the Importance Learning Unit 60.

[0057] AI Response Unit 10 receives inquiries from multiple mobile terminals 70, generates answers using AI, and sends them to the mobile terminals 70. Additionally, it obtains information from the Internet 80 and internal company information 81 as needed. Furthermore, it appropriately notifies managers 82.

[0058] The terminal communication unit 11 is responsible for sending and receiving information with the mobile terminal 70. Communication with the mobile terminal 70 is conducted using wireless communication. The wireless communication can be a dedicated system line or a standard mobile communication device. Communication can be in the form of email or chat. Furthermore, depending on the content of the inquiry, it may include audio such as a telephone call, or video such as a video call.

[0059] Information storage unit 12 is the part that stores information used for AI. It stores general information, internal company information, and other information used in AI learning, or newly acquired information from external sources.

[0060] General information refers to generally known information that may be used in the answer, such as biographies, quotes from successful people, advice from experts, self-help books, professional books, comedy studies, humor studies, current affairs materials, Mandarin, arithmetic, science, society, English, laws, regulations, ISO, JIS and other standards and specifications, as well as intellectual property information (patents, trademarks, copyrights, etc.).

[0061] Using biographies and quotes from successful individuals is effective because they contain content that serves as advice and encouragement to the inquirer. Using books of wisdom, self-help, and professional knowledge is beneficial because they contain advice that can be offered when the inquiry concerns technical aspects or directions for skill improvement. Using comedic and humorous elements is effective in making the answer more relaxing for the inquirer. Using current events is effective in providing answers that consider contemporary changes. Using Mandarin, arithmetic, science, social studies, and English is necessary for generating answers based on fundamental knowledge. Using laws, regulations, standards such as ISO and JIS, and intellectual property information is necessary because answers require a deeper level of knowledge.

[0062] Furthermore, if all information can be randomly obtained as general information, it is conceivable that information derived from errors, ideas deviating from common sense, or emotions might be obtained. Consequently, common-sense judgments are not made. Additionally, it is conceivable that the required information will vary depending on the system's usage pattern and scenario. Therefore, it is preferable to set the information to a format in which users and administrators can appropriately select and determine the required domain, type, scope, and amount of general information in advance or subsequently. By adopting this format, harmful information can be guaranteed to be excluded during AI's judgment and response generation, and the AI ​​can be endowed with the characteristics (professionalism, personality) desired by users and administrators.

[0063] Internal company information refers to crucial information within the company, such as corporate philosophy, corporate culture, internal regulations, policies, history, contracts, specifications, procedures, stories of past successful employees, customer information, operational information, travel reports, minutes of internal and external meetings, and employee competency information (qualifications, experience, performance records). Corporate philosophy often forms the foundation for company culture. Stories of past successful employees sometimes serve as a reference when determining business direction. Customer and operational information directly relate to employees' work. Travel reports and minutes of internal and external meetings accumulate over time as internal company information. Employee competency information serves as a reference when providing personalized advice, building well-balanced teams, and utilizing personnel effectively.

[0064] Furthermore, in addition to its effective use within a single enterprise, this invention can also be envisioned for effective use among multiple enterprises, such as joint ventures and conglomerates. Moreover, its effective use extends beyond enterprises to include all situations requiring human interaction, such as schools, government agencies, group activities, and political environments. Therefore, terms like "within a company" and "enterprise" should not be interpreted literally but rather appropriately based on the context in which they are effectively used.

[0065] Additionally, queries, answers, and answer satisfaction are stored as example problems for use in machine learning. As example problems, queries, answers, and answer satisfaction are grouped together. When a particular query is performed, the answer and answer satisfaction are prepared as a model solution.

[0066] When AI performs deep learning and other machine learning processes, it compares the output answers, provisional answer satisfaction levels, prepared answers as model answers, and answer satisfaction levels during the machine learning process, providing feedback to improve learning accuracy.

[0067] External Communications Department 13 is responsible for acquiring information from external sources and sending reports to Manager 82. When new information exists on the Internet 80 or in the company's internal information system 81, this information is collected and stored in Information Storage Department 12. Additionally, when an inquiry contains important information, the report is sent to Manager 82.

[0068] The inquiry handling unit 20 is the part that preprocesses the inquiries received at the terminal communication unit 11 before handing them over to the learning unit 30.

[0069] The inquiry handling department 20 possesses capabilities for word and sentence analysis 21, contextual understanding 22, and frequent dialogue identification 23. Word and sentence analysis 21 involves analyzing the sentence units used in the inquiry. It primarily analyzes phrases and parts of speech. Through word and sentence analysis 21, the content of the sentence unit is determined.

[0070] Contextual understanding section 22 focuses on identifying the relationships between sentences within the text. This analysis helps to pinpoint the subtle differences between each sentence.

[0071] Frequent dialogue identification 23 is the part that determines whether to input the content of the query into the AI. For queries that can be answered simply by referring to data, such as "What's the weather like tomorrow?" or "What's my schedule for tomorrow?", the AI ​​will not process the answer. For other queries, the content of the query will be sent to the AI.

[0072] The learning unit 30 is the main part of this invention. It is the part that outputs answers to queries. The learning unit 30 has AIs for each domain: general information AI 31 and internal company information AI 32. General information AI 31 is the AI ​​that outputs answers learned from general information, and internal company information AI 32 is the AI ​​that outputs answers learned from internal company information. Depending on the answering method, there may be cases where answers from all AIs are used, or only answers from a portion of the AIs are used.

[0073] AI takes many forms, including methods using neural networks. These methods break down the input query into a mesh-like structure with numerous branches and weights, outputting content that closely approximates the expected answer, learned through example problems. An example of machine learning will be described separately.

[0074] The response generation unit 40 is responsible for arranging the responses output by the learning unit 30 for each inquirer. This includes situations where the inquirer expects a gentle response or a straightforward one. It is also responsible for adjusting the overall style of the responses.

[0075] For example, if the output from Learning Department 30 is "Don't worry about your surroundings, be confident." and the questioner expects a gentle response, such as "Although a sudden change may be difficult, don't worry about your surroundings, be confident, and things will get better.", then it becomes an expression that emphasizes concern for the other person.

[0076] Instead of expecting a straightforward answer, adjust your approach to include clearer statements such as, "Don't worry about what's around you, be confident! Things will get better."

[0077] Additionally, roles are added to the answer based on the inquirer's preferences. These roles include image, voice, and personality. For example, if a response is expected to be from an older, more experienced person, such a role is added. If a response is expected to be from a woman, a female role is added.

[0078] The Interpersonal Adjustment Department 50 is used when the inquiry concerns interpersonal relationships. It features an Interpersonal Adjustment AI 51. This AI has learned information for adjusting relationships among multiple people. The inquiries primarily concern eliminating or reducing conflicts with the target individual, or improving or repairing existing relationships.

[0079] Regarding the relationship between oneself, the target, and those around them, the AI ​​learns from various examples to generate the best answer.

[0080] The Importance Learning Department 60 identifies important information in the content from inquirers and reports it to Manager 82. The Importance Learning Department 60 has an Important Project AI 61. When the inquirer is a company employee, and the inquired content is determined to have an impact on the company's operations or departmental status, the Important Project AI 61 reports information to Manager 82 regularly and irregularly. The Important Project AI 61 assesses importance by using machine learning to determine importance through examples and answers to operational and organizational questions.

[0081] Mobile terminals 70 are owned by various personnel or individuals, and include wearable terminals, smart speakers, and IoT devices. Furthermore, as mentioned above, this system can also connect to a desktop PC, so more broadly, it includes the desktop PC. A display unit 71 is provided to display queries and responses, and an application 72 is built-in to communicate with the AI ​​response unit 10 according to the system. By enhancing the processing power of the AI ​​response unit 10, the number of mobile terminals 70 can be increased arbitrarily. Therefore, all company employees can own one. In this way, all information from the entire workforce enters the AI ​​response unit 10, and the real-time nature of the information is improved.

[0082] Internet 80 and internal company information 81 are the information providers outside the system. Through constant or periodic connections, the response content of AI Response Department 10 can be updated sequentially.

[0083] Manager 82 is responsible for the application of AI Response Department 10. Within the company, all employees can have mobile devices 70, and the General Manager, acting as Manager 82, manages them. Manager 82 receives reports from AI Response Department 10, enabling them to efficiently grasp the overall situation of the workforce in a short time.

[0084] (Explanation of machine learning)

[0085] according to Figure 2 Taking general information AI31 as an example, this paper explains the actions of AI and the content of AI machine learning.

[0086] Figure 2 (a) is a schematic diagram illustrating an example of AI. The AI ​​consists of an input unit 33, an intermediate unit 34, and an output unit 35. The input unit 33 receives the AI's input, appropriately decomposing and inputting the query data. The intermediate unit 34 is a neural network that has undergone deep learning, performing multi-stage processing corresponding to the query data. It has numerous neurons connected in multiple layers. Learning is achieved by varying the weights between neurons. Based on the learned content, the output unit 35 outputs what is considered the optimal data, representing the data's satisfaction level (inferred to be the satisfaction level the queryer might feel). More precisely, the output with the highest satisfaction level is selected as the output.

[0087] Figure 2 (b) is a schematic diagram illustrating an example of AI learning. The AI ​​consists of an input unit 33, an intermediate unit 34, and an output unit 35. The data consists of a large number of example questions. The example questions consist of queries, information sets, model answers, and model answer satisfaction. Model answers and model answer satisfaction are also referred to as teacher data.

[0088] The learning method involves appropriately decomposing the queries for each example question and inputting them into the input unit 33, so that a provisional answer and a provisional answer satisfaction level are output from the output unit 35. If the provisional answer and its satisfaction level differ from the model answer and its satisfaction level, the differences are fed back to the neural network in the intermediate unit 34. This process is repeated to improve learning accuracy.

[0089] As additional learning occurs, the system obtains the inquirer's satisfaction with the answers to their questions during system operation, thus enabling additional learning to be performed at any time.

[0090] (Explanation of the system for providing responses and other measures)

[0091] according to Figure 3 , Figure 4 This section explains the actions of a response system that answers in response to inquiries. Figure 3 This is a flowchart from asking a question to answering it. Figure 4 The diagram shown is displayed on the display unit 71 of the mobile terminal 70.

[0092] The inquirer inputs a query into the mobile terminal 70. For example, as in C101, they input "The consultation has ended." The mobile terminal 70 sends the query to the AI ​​response unit 10, and the terminal communication unit 11 of the AI ​​response unit 10 receives the content (S101). The query receiving unit 20 of the AI ​​response unit 10 analyzes the query content. Sentence decomposition is performed using the word and sentence analysis 21 (S102). "The consultation has ended" is decomposed into "the consultation" and "ended." Then, contextual analysis is performed (S103).

[0093] In this case, since it is the initial stage of the dialogue, no context-based correction is performed. Frequent dialogue identification 23 determines whether the dialogue is frequent (S104). In this case, since it is a simple report, it is judged to be a frequent dialogue, and a response is extracted (S105). The response is "Received." The response is sent to the importance learning unit 60 (S109). The response is corrected according to the terminal settings. For example, the response is corrected depending on whether the inquirer expects a gentle or straightforward response. In a gentle response, it is "Received. Thank you for your hard work." In a straightforward response, it is "Received."

[0094] Additionally, it's possible to add the respondent's role along with the answer's revision. For example, if the inquirer prefers an older superior as the respondent, image data for role A73 can be added. If the inquirer prefers a female superior as the respondent, image data for role B74 can be added. Furthermore, the role referred to here includes not only image but also voice, personality, etc.

[0095] In this way, by allowing the questioner to choose a preferred role, the questioner can develop a sense of closeness to the respondent.

[0096] The response is sent to the mobile terminal 70 via the terminal communication unit 11 (S108). The response "Received. Thank you for your hard work." (C102) and the character A73 are displayed on the display unit 71 of the mobile terminal 70.

[0097] Next, we will explain the situation where the inquirer asks, "Would it be better to go to store A tomorrow?" (C103). The inquiry is received by the terminal communication unit 11 (S101), broken down into short sentences (S102), and analyzed according to context (S103). Through the frequent dialogue identification 23, it is determined that the answer should be based on the company's internal information AI 32 (S104), and the inquiry is sent to the learning unit 30.

[0098] In the learning unit 30, the AI ​​selects internal company information AI 32, inputs a query into the AI, and obtains an answer as the AI's output (S106). For example, the answer might be, "Sales at store A are decreasing. Please check." The answer generation unit 40, based on the terminal's settings, might modify the answer to, for example, "Sales at store A are declining. Go check the situation." (S107)

[0099] The corrected answer is sent to the mobile terminal 70 via the terminal communication unit 11 (S108). The answer (C104) and the character A73 are displayed on the display unit 71 of the mobile terminal 70.

[0100] Frequent conversations include simple reports and questions about basic information. Basic information refers to information that can be found through online searches, such as weather or traffic congestion information. It also includes information easily accessible by accessing related resources.

[0101] When the question is "What is tomorrow's schedule?" (C106), the information can be confirmed by accessing the company's internal system, thus indicating a frequent conversation. In response, display unit 71 shows the answer: "Morning, video conference with Company X. Afternoon, introduction to Company Y." (C107) and character B74.

[0102] Thus, according to the present invention, it is possible to provide accurate answers to queries through a large amount of information and machine learning.

[0103] (Explanation of the interpersonal adjustment system)

[0104] according to Figure 5 , Figure 6 This section explains the actions of the interpersonal adjustment system.

[0105] The inquiries also include questions and discussions about interpersonal relationships, such as not wanting to meet Mr. ○○ or wanting to meet Mr. △△. While advice can be given to a single inquirer, it can also be pushed forward when the recipient is a user of the same system. In this case, the system will initiate a conversation with the recipient even if they don't ask any questions, or the conversation will be added to their answers to other inquiries.

[0106] Improving interpersonal relationships is extremely beneficial for businesses, as it can increase operational efficiency, and for individuals, it can lead to a more comfortable life. Furthermore, interpersonal relationships are generally sensitive issues, and unnecessary intervention by a third party can often complicate matters. However, conversations initiated by AI systems can address issues without preconceived notions.

[0107] The following describes the handling of a scenario where inquirer A sends an inquiry stating, "I hope to do 〇〇's work. I do not wish to speak with Mr. B." (C201). The process involves receiving the inquiry (S201), breaking it down into short sentences (S202), analyzing the context (S203), and using frequent dialogue identification (S204) to determine if interpersonal connections are involved. If not, a general AI response is obtained (S205). If so, interpersonal AI is used for processing (S206).

[0108] The Interpersonal Adjustment AI51 in the Interpersonal Adjustment Department 50 identifies the target individuals. The characteristics of the target individuals are confirmed. Expectations for the target individuals are determined. For example, in business matters, this might include a reluctance to join the same team. In corporate matters, adjustments to business content are made (S206).

[0109] When adjusting business content, confirm the business content, capabilities, and performance of A and B as internal company parameters. Next, confirm the workload of each group within the company. Then, develop a personnel adjustment plan based on this. Regarding this inquiry, develop a plan to separate the business groups of A and B (C202).

[0110] Create answers that correspond to the plan.

[0111] It can also create multiple plans, calculate the effectiveness, and select the plan with the highest effectiveness. In this way, the precision of the adjustments is increased.

[0112] Based on the settings of terminal A and the settings of the target's (B) terminal, modify the prepared response. For example, if the response to A is "○○ business until the 10th. B leaves." and the response to B is "Conducting △△ business", then modify the response to A to "Mr. A, △△ is great. (^^) Please conduct ○○ business until the 10th. Please leave the business adjustment with Mr. B to me." (C203).

[0113] The revised version for B is: “Mr. B, □□ was also great today. Let’s do △△ business together this week.” (C204)

[0114] A reply is sent to the recipient (B) (S208). The message displayed on B's terminal reads, "Mr. B, □□ was great today too. Let's do △△ business together this week." This is a gentle approach, reducing any sense of unease.

[0115] A reply is sent to terminal A (S209). Terminal A displays the message: "Mr. A, △△ is great. (^^) Please proceed with the 〇〇 service until the 10th. Please leave the service adjustment with Mr. B to me." This notifies the terminal of the adjustment status.

[0116] In this way, even when the inquiry involves content related to interpersonal relationships, adjustments can be made without human intervention by contacting the person being inquired about.

[0117] (Explanation of the Important Notice System)

[0118] according to Figure 7 , Figure 8 This section explains the actions of the important matters notification system.

[0119] In enterprise-oriented systems, it is more convenient to report important information that involves a large number of inquiries and reports to managers.

[0120] Therefore, this describes a system that identifies and reports important matters in parallel with the system for responding to inquiries.

[0121] For example, if an inquirer sends a message such as "Company ○○ seems to be pursuing business △△." (C301), the AI ​​response unit 10 receives the inquiry (S201), performs sentence decomposition (S302), and context analysis (S303). Data and destination AIs are sent in parallel to the importance learning unit 60. The importance learning unit 60's important item AI 61 confirms the importance of the △△ business and its relationship with Company ○○ based on internal company parameters, using these as judgment factors. Additionally, the trend of the △△ business is confirmed as a general parameter. Based on this information, the important item AI 61 determines whether it is considered an important matter (S304, C302). If it is determined to be an important matter, the item and its importance are identified. Furthermore, a timeline is set based on the urgency of the important matter, indicating whether it requires immediate judgment, judgment within the week, or judgment within the year.

[0122] Additionally, prepare a report for manager 82 (C304) (S305). Then send the report to manager 82 (S306).

[0123] Inform the inquirer that the information will be reported as important (C303). This is because: as an inquirer, being reported arbitrarily may be unpleasant, and if it is important information, it will motivate them to investigate further details.

[0124] This allows for the fluent response to inquiries while simultaneously extracting all the important information relevant to the company, making it an ideal choice for companies to respond swiftly.

[0125] Example 2

[0126] (An illustration of an example of selecting answers from multiple AIs)

[0127] according to Figure 9 , Figure 10 Examples of selecting answers from multiple AIs are provided.

[0128] As an AI, the system was described based on AIs with larger information bases, such as general information AI31 and internal company information AI32, but it was also able to list or select answers from AI groups with smaller information bases.

[0129] For example, information can be weighted, such as in cases where advice from Steve Jobs' quotes is preferred, to build an AI system that is more relevant to the person asking the question.

[0130] In this embodiment, AIs such as Common Sense AI, Entrepreneur AI, Steve Jobs AI, and Jiakang AI are prepared, and the weights of the AIs are changed.

[0131] The weights were set as follows: Common Sense AI 80%, Entrepreneur AI 10%, Jobs AI 10%, and Jiakang AI 0% (C402).

[0132] The terminal sends a message such as "I lack confidence in my work" as an inquiry. The AI ​​response unit 10 receives the inquiry (S401), performs sentence breakdown (S402) and context analysis (S403), and inputs the inquiry content to each AI (S404).

[0133] The AI ​​group uses common-sense AI, entrepreneur AI, and Jobs AI.

[0134] Each AI outputs an answer and a recommendation score (S405). Common Sense AI's answer is "Reassess the job content.", with a recommendation score of 5 out of 100. The recommendation score reflects the degree of confidence in the answer. A recommendation score of 5 indicates a lack of confidence.

[0135] The AI's response to the entrepreneur was, "Even our predecessors had those moments." Its recommendation rate was 50 out of 100. A recommendation rate of 50 indicates a certain level of confidence.

[0136] The AI's answer to the Jobs story was "Have the courage to trust your heart and intuition." It received a 90% recommendation out of 100. A 90% recommendation indicates a high level of self-confidence (C403).

[0137] The recommendation scores and weights of each answer are multiplied to calculate the answer ranking (S406). The answer rankings are: Common Sense AI 4, Entrepreneur AI 5, and Jobs AI 9 (C404). Therefore, the answer from Jobs AI, which has the highest ranking, is selected (S407, C405).

[0138] Send a response (S408, C406).

[0139] By adopting such a structure, we can obtain answers that match the user's preferences and ideas.

[0140] Additionally, it can display recommendation levels, answer priority, and list the answers from various AIs. This allows the questioner to choose answers and broaden the range of responses.

[0141] Furthermore, it can also take the following form: by combining the thinking and personality of multiple AIs such as Common Sense AI, Entrepreneur AI, Jobs AI, and Ieyasu AI, a comprehensive AI persona with various personalities is pre-generated, and this comprehensive AI persona provides answers. By adopting this form, more accurate answers can be obtained from the fusion of the thinking of multiple AIs. In this case, the preferred form is that the manager 82 and the user can pre-select from the AI ​​group, decide which AIs to be combined, and appropriately determine the combination allocation of the selected AIs.

[0142] Example 3

[0143] (Explanation of an example where there are multiple respondents)

[0144] While AI can provide accurate answers to inquiries, the responses can vary depending on the content of the question and the inquiry's situation, sometimes leading to unacceptable responses. Therefore, having two respondents derive the AI's answer based on a pre-arranged agreement clarifies the process and increases the inquiry's sense of trust.

[0145] As an example, let's explain a scenario where an inquiry such as "I can't get the order" (C501) is received. The AI ​​response unit 10 receives the inquiry from the mobile terminal 70 (S501). It outputs a general information AI response (S502).

[0146] Here, we discuss the AI's estimation of the process from inquiry to response (S503).

[0147] Discussion AI is an AI that constructs a discussion process based on the content of the inquiry, general information, the AI's response, and general information to derive the answer.

[0148] The discussion about AI will be transformed into a dialogue between two superiors (S504). Then, the dialogue content between the two superiors will be sent to the mobile terminal 70 (S505).

[0149] Regarding the inquiry about "not getting orders," the answer is "strengthen trust with customers towards the end of the period." However, the inquirer sometimes finds this answer unconvincing. But by explaining the process leading up to the answer through a dialogue between two people, the inquirer's level of acceptance increases. In this example, the inquirer's understanding deepens through the dialogue from C502 to C505.

[0150] Thus, according to this embodiment, the inquirer can learn about the process rather than just a simple answer, and therefore can be more easily convinced, increasing their trust in the AI ​​response unit 10 and improving business efficiency.

[0151] Industrial availability

[0152] The AI-based answering system of this invention can be understood as having significant industrial applicability for improving the efficiency of business operations by providing AI-based responses to inquiries. Furthermore, this system can function as a facilitator in meetings, facilitating quick and appropriate meetings when many participants are consistently negative or silent. In addition to its effective use in enterprises, this system can also be effectively utilized in all settings requiring human interaction, such as schools, government agencies, group activities, and political environments.

[0153] Furthermore, the AI-powered response system of this invention can also be used for multilingual responses based on simultaneous interpretation and translation functions. This functionality greatly facilitates smooth business operations for overseas deployments with time differences, overseas enterprises, and individuals located overseas. Therefore, this invention has significant industrial applicability.

[0154] Furthermore, the AI ​​response system of this invention enables accurate instructions and reliable communication for the hearing impaired and the elderly through text-based interaction. It also features a function that converts text and voice, thus greatly benefiting the visually impaired. In addition to text and voice, it can be equipped with sign language, Braille, and other communication methods depending on the user. Therefore, according to the AI ​​response system of this invention, healthy individuals, disabled individuals, and the elderly can achieve barrier-free communication. Moreover, in a super-aging society where life expectancy is often considered to be 100 years, and with calls for work style reforms, it also benefits businesses by promoting barrier-free and diversified employment. Therefore, its industrial applicability is considerable.

[0155] Explanation of reference numerals in the attached figures

[0156] 1…AI Response System; 10…AI Response Department; 11…Terminal Communication Department; 12…Information Storage Department; 13…External Communication Department; 20…Inquiry Handling Department; 21…Word and Sentence Analysis; 22…Contextual Understanding; 23…Frequent Dialogue Identification; 30…Learning Department; 31…General Information AI; 32…Internal Company Information AI; 33…Input Department; 34…Intermediary Department; 35…Output Department; 40…Answer Generation Department; 50…Interpersonal Adjustment Department; 51…Interpersonal Adjustment AI; 60…Importance Learning Department; 61…Important Project AI; 70…Mobile Terminal; 71…Display Department; 72…Application; 73…Role A; 74…Role B; 80…Internet; 81…Internal Company Information; 82…Manager.

Claims

1. An AI-powered answering system, characterized in that, The device includes a learning unit and an answer generation unit. The learning unit performs machine learning based on queries, general information, answers, and a satisfaction index regarding the answers. The answer generation unit, based on the learning results of the learning unit, generates answers to queries from multiple mobile terminals and sends them to each mobile terminal. It has an interpersonal adjustment department and an interpersonal adjustment system. When the inquiry concerns interpersonal relationships, the interpersonal adjustment department adjusts interpersonal relationships and business content. The interpersonal adjustment system notifies the inquirer and relevant parties of the answers regarding the adjustment of interpersonal relationships. Notifications to relevant parties are sent separately, with content presented differently from that of the inquirer, and are either sent unilaterally by the interpersonal adjustment system or added to the response to other inquiries. The answer generation unit can modify the expression of the answer based on information related to the inquirer and relevant parties. The general information includes at least one of the following: biographies, quotes from successful people, advisors, self-help books, professional books, comedy studies, humor studies, current affairs materials, Mandarin, arithmetic, science, society, English, laws, departmental orders, ISO, JIS standards and specifications, and intellectual property information, wherein the intellectual property information is patents, trademarks, and copyrights.

2. The AI ​​response system according to claim 1, characterized in that, In addition to general information, machine learning is also performed based on internal company information.

3. The AI ​​response system according to claim 1 or 2, characterized in that, Inquiry includes at least one of the following: report, contact, discussion, opinion, hope, question, worry, dream, likes and dislikes, and self-introduction.

4. The AI ​​response system according to claim 2, characterized in that, Company internal information includes at least one of the following: corporate philosophy, corporate culture, internal regulations, policies, history, contracts, specifications, process documents, stories of past successful employees, customer information, business information, travel reports of employees on business trips, minutes of internal and external company meetings, and employee competency information, which includes qualifications, experience, and performance records.

5. The AI ​​response system according to claim 1 or 2, characterized in that, The response shall include at least one of the following: instructions, suggestions, teachings, guidance, praise, criticism, encouragement, content on improving the comprehensive abilities of individuals, content on improving the comprehensive abilities of members of society, communication of affairs, and support for the production of information.

6. The AI ​​response system according to claim 1 or 2, characterized in that, Administrators or users can select a portion of the general information.

7. The AI ​​response system according to claim 1 or 2, characterized in that, The respondent's appearance and the way they express their answers can be selected from multiple characters.

8. The AI ​​response system according to claim 1 or 2, characterized in that, The response generation department generates a response based on the weighting of each piece of information.

9. The AI ​​response system according to claim 1 or 2, characterized in that, The response generation department generates responses in the form of multi-person dialogues.

10. The AI ​​response system according to claim 1 or 2, characterized in that, It also features an important notification system that targets inquiries and notifies them based on their importance. This system has an importance learning unit that uses machine learning to perform machine learning on inquiries, internal company information, general information, and an importance index for the inquiries. Based on the learning results of this importance learning unit, it selects the most important inquiries from multiple mobile terminals and notifies the manager.

Images