Inquiry response system

Abstract

To automatically respond to an inquiry from a user by a chatbot and seamlessly take over to a responder for a manned response.SOLUTION: An inquiry response system 1 which responds in chat to an inquiry made online from a user in chat includes: an inquiry receiving unit 10 which receives an inquiry from the user; a responder management unit 20 which allocates a responder to the inquiry received by the inquiry receiving unit 10; and an inquiry processing unit 30 which responds to the inquiry by a chatbot 31 corresponding to the responder allocated by the responder management unit 20, and takes over to a manned response in which the allocated responder answers the inquiry in chat when it is determined that the inquiry cannot be handled by the chatbot 31.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to a technology for responding to online inquiries from users, and particularly to a technology effective when applied to an inquiry response system that combines automatic response by a chatbot and human response.

Background Art

[0002] In help desks and support centers that receive and answer questions and inquiries from users via email, chat, voice, etc., depending on the characteristics of products and services, characteristics of the sales region, etc., there are cases where 24-hour, 365-day response is required, and the burden on human operators for response is large. In such a situation, technologies for automatically responding to all or part of the human response using so-called chatbots have been studied.

[0003] For example, Japanese Unexamined Patent Application Publication No. 2020-204876 (Patent Document 1) describes a network system including a first terminal used by a questioner, a second terminal used by an answerer, and a server for automatically chatting with the questioner by a chatbot and taking over the chat with the questioner by at least one answerer when a predetermined condition is satisfied.

[0004] Also, for example, Japanese Unexamined Patent Application Publication No. 2019-128625 (Patent Document 2) describes a communication support device that creates master data used for learning the AI (artificial intelligence) engine of a chatbot from the exchange of messages in a chat function room.

Prior Art Documents

Patent Documents

[0005]

Patent Document 1

Patent Document 2

Summary of the Invention

Problems to be Solved by the Invention

[0006] According to the prior art, for example, while a chatbot automatically responds to inquiries from users, appropriate teacher data is extracted from the conversation and further used for AI learning to improve the accuracy of the chatbot or construct appropriate FAQs (Frequently Asked Questions). In addition, when it becomes difficult for the chatbot to provide an automatic response, a mechanism for transferring to a human response can be realized.

[0007] By automatically responding with a chatbot, the same response can be given to the same type of inquiry, and it can be said that a uniform response can be provided to all users by excluding the individuality of each responder as compared to a human response by an operator or the like. On the other hand, there may be cases where it is more appropriate for the user that the response content varies depending on the experience, skill, personality, and way of thinking of each responder, or the characteristics of the responder are revealed.

[0008] Also, when considering transferring from an automatic response by a chatbot to a human response as in Patent Document 1, if a responder takes over after the chatbot has given a homogenized response, there may be cases where the responder takes over the conversation made with characteristics different from their own, and it is feared that a seamless transfer cannot be achieved, resulting in a situation where it is difficult for both the user who made the inquiry and the responder to handle.

[0009] Therefore, an object of the present invention is to provide an inquiry response system that enables a seamless transfer to a responder when a chatbot automatically responds to an inquiry from a user and then transfers to a human response.

[0010] The above and other objects and novel features of the present invention will become apparent from the description of this specification and the accompanying drawings.

Means for Solving the Problems

[0011] Among the inventions disclosed in the present application, the outline of typical ones will be briefly described as follows.

[0012] The inquiry response system, which is a typical embodiment of the present invention, is an inquiry response system that responds to an online inquiry by chat from a user by chat, and includes an inquiry reception unit that receives an inquiry from the user, a responder management unit that assigns a responder to the inquiry received by the inquiry reception unit, and an inquiry processing unit that corresponds to the inquiry by a chatbot corresponding to the responder assigned by the responder management unit, and when it is determined that the chatbot cannot handle the response, takes over to a manned response in which the assigned responder answers by chat.

Advantages of the Invention

[0013] Among the inventions disclosed in the present application, the effects obtained by typical ones will be briefly described as follows.

[0014] That is, according to a typical embodiment of the present invention, even when taking over from a chatbot response to a manned response by a responder for an inquiry from a user, it is possible to seamlessly take over for both the user and the responder, and it is possible to reduce the burden on the responder.

Brief Description of the Drawings

[0015]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Embodiments for Carrying Out the Invention

[0016] Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In all the drawings for explaining the embodiments, the same parts are generally denoted by the same reference numerals, and the repeated explanations thereof are omitted. On the other hand, for the parts described with reference numerals in a certain drawing, they will not be shown again in the explanation of other drawings, but may be referred to with the same reference numerals.

[0017] (Embodiment 1) <Overview> The inquiry response system according to Embodiment 1 of the present invention automatically responds to questions and inquiries from users by chat or voice through a chatbot, and when it becomes difficult to respond with the chatbot, it is taken over by a human response by a responder. In this mechanism, the chatbot is not implemented as a homogeneous one that excludes humanity, but is associated with each responder individually, and each chatbot corresponds to the user as a "doppelganger" that has individually learned the characteristics of the corresponding responder. As a result, each responder first automatically responds to the chatbot of their own "doppelganger", and when taking over the human response from the chatbot, they can seamlessly take over the human response without changing the characteristics and features in responding to the user by taking over themselves.

[0018] <System Configuration> FIG. 1 is a diagram showing an overview of a configuration example of the inquiry response system according to Embodiment 1 of the present invention. The inquiry response system 1 of the present embodiment is an information processing system that receives inquiries by chat from a user via a user terminal 3 such as a smartphone or a PC through a network 2 such as the Internet, and performs an automatic response by a chatbot or a human response that responds to the input answer by a responder by chat.

[0019] The inquiry response system 1 shown in FIG. 1 may be configured using an information processing terminal such as a PC, or may be configured by a virtual server or the like constructed on a server device or a cloud computing service. Then, by a CPU (Central Processing Unit) (not shown), an OS (Operating System), a DBMS (DataBase Management System), middleware such as a Web server program, and software operating thereon are expanded from a recording device such as an HDD (Hard Disk Drive) onto the memory, and various functions related to inquiry response by chat are realized.

[0020] The inquiry response system 1 has each part such as an inquiry reception part 10, a responder management part 20, and an inquiry processing part 30 implemented by software, for example. The inquiry processing part 30 further has each part such as a chatbot 31 and a chat processing part 32. Also, the inquiry response system 1 has each data store such as a responder master 41, a responder status 42, a common FAQ 51, an individual FAQ 52, and a response history 61 implemented by a database, a file, etc., and further has one or more individual response models 70 which are learned models used when the chatbot 31 answers by AI.

[0021] The inquiry reception part 10 receives, for example, an inquiry in a chat from an inquiry Web page or the like displayed on the user terminal 3 by a Web server program (not shown), determines a responder by the responder management part 20 described later, and has a function of performing an online response by chat by the inquiry processing part 30 described later when a responder is determined. When there is no available responder and a responder cannot be determined, it may have a function of adding the inquiry to a queue waiting for response and managing it, or notifying the user that the response is impossible.

[0022] The responder management part 20, for example, for each responder registered in the responder master 41, based on the status information recorded in the responder status 42, acquires the information of the responder currently logged in to the inquiry response system 1, and has a function of acquiring the information of an available responder, that is, the information of a responder not in the middle of chat correspondence. Further, it has a function of determining a responder who responds to the user's inquiry from the acquired available responders. The responder connects to the inquiry response system 1 via the network 2 or a VPN (Virtual Private Network) (not shown) using an information processing terminal such as a PC (not shown) used by the responder himself / herself, logs in, and waits for an inquiry from the user.

[0023] Note that the respondent master 41 holds, for each respondent, in addition to identification information such as name and nickname, attribute information such as skill level, area of expertise, available time zone and schedule, priority at the time of respondent allocation, and information for specifying the individual response model 70 to be corresponded. Also, the respondent status 42 holds, for example, information on whether or not the respondent is in a chat correspondence for each respondent logged in to the inquiry response system 1. When in a chat correspondence, information on whether it is being corresponded by the chatbot 31 as the "avatar" of the respondent himself / herself described above or whether the respondent himself / herself is directly chatting and corresponding may be held.

[0024] The inquiry processing unit 30 has a function of corresponding, for example, to an inquiry in a chat from a user received by the inquiry reception unit 10 by answering in the chat. As described above, in this embodiment, for an inquiry from a user, initially, it is automatically corresponded by the chatbot 31 as the "avatar" of the respondent himself / herself determined by the respondent management unit 20. On the other hand, when the inquiry processing unit 30 (or the chatbot 31) determines that it is difficult or impossible to perform further automatic correspondence based on the response content of the chatbot 31 (for example, "Sorry, I can't answer") or the output result, etc., it is taken over to a manned correspondence where the respondent himself / herself directly answers by chat.

[0025] Note that the chat processing unit 32 has basic chat functions such as providing a chat screen for the user and the respondent, forming a chat room, and recording chat history in the correspondence history 61, and can be implemented by appropriately using an existing chat engine, application, etc., or in cooperation with these.

[0026] Chatbot 31 has a function of generating a response text for an inquiry from a user using, for example, an individual response model 70 corresponding to a responder determined by a responder management unit 20 by an AI engine (not shown) and outputting it to a chat processing unit 32. Since the individual response model 70 is a model individually learned in advance by an AI engine (not shown) for the characteristics of the corresponding responder, it can be expected to give a response similar to or close to the case where the responder actually responds. As teacher data for learning the characteristics of the responder, for example, the content of the chat when the responder responded in a manned response recorded in the correspondence history 61 can be used.

[0027] The response by the chatbot 31 is extracted and responded, for example, from the FAQs registered in the common FAQ 51 or the individual FAQ 52. For example, the common FAQ 51 can be configured to register basic FAQs common to each responder, and the individual FAQ 52 can be configured to register unique FAQs for each responder. When extracting the FAQ, it may be extracted by pattern matching using simple keywords, or it may be extracted by an AI engine.

[0028] Note that the FAQs registered in the common FAQ 51 and the individual FAQ 52 may be manually registered by the responder or the like, or may be automatically extracted by an AI engine, various tools, etc. based on the history information accumulated in the correspondence history 61. Here, even in the case of a response configuration using an AI model for each responder, it is assumed that it is difficult to form an AI model for each responder from the beginning. A general AI model may be formed and each responder may perform additional learning while using the system to form an AI model for each responder.

[0029] While the chatbot 31 (and the individual response model 70) is automatically responding, it is assumed that the corresponding responder is waiting while displaying the chat screen of the chat room where the chatbot 31 (and the individual response model 70) is communicating with the user on an information processing terminal such as a PC (not shown) used by the responder himself / herself.

[0030] <Automatic Response and Human Response by Chatbot> FIG. 2 is a diagram schematically showing an overview of an example of an aspect of automatic response by the chatbot 31 and human response by responders in Embodiment 1 of the present invention. On the left side of FIG. 2, user terminals 3a to 3c of three users who make inquiries by chat are shown. On the right side, three responders 4a to 4c who are logged in to the inquiry response system 1 are shown, and in the center, chatbots 31a to 31c as "avatars" corresponding to each of the responders 4a to 4c are shown. The chatbots 31a to 31c are actually realized as instances generated in such a form that, for example, the chatbot 31 of the inquiry response system 1 uses individual learning models 70a to 70c corresponding to each of the responders 4a to 4c.

[0031] In the example of FIG. 2, solid arrows indicate that the chatbot 31a automatically responds to an inquiry from the user terminal 3a, and the chatbot 31c automatically responds to an inquiry from the user terminal 3c. On the other hand, solid arrows indicate that the chatbot 31a (which may be another instance using the same individual learning model 70a) automatically responds to an inquiry from the user terminal 3b. In this way, it is also possible to respond to inquiries from a plurality of users in parallel by the chatbot corresponding to a specific responder. In this case, it is desirable to set an upper limit value for the number of inquiries that can be responded to in parallel.

[0032] Also, in the example of FIG. 2, solid arrows indicate that when it becomes difficult for the chatbots 31a and 31c to make an automatic response, they are taken over by the responders 4a and 4c respectively. On the other hand, for example, the chatbot 31a can be taken over by another responder (responder 4b in the figure) or another chatbot (chatbot 31b in the figure) as indicated by the dotted arrow, instead of being taken over by the responder 4a corresponding to itself, based on the content of the inquiry from the user and the status of the responder 4a.

[0033] FIG. 3 is a diagram showing an overview of an example of automatic response by the chatbot 31 in Embodiment 1 of the present invention. In the upper and lower parts of the figure, examples of chat screens related to respective inquiries displayed on the user terminals 3a and 3b are shown. In the example of FIG. 3, for any of the chatbots 31a and 31b corresponding to respective inquiries, for inquiries of general and basic content, by obtaining an answer from the common FAQ 51 in which basic FAQs common to each respondent are registered, it shows that for similar inquiries (in the example in the figure, "Can this product do ○○?"), similar answers (in the example in the figure, "Yes, it can do ○○.") are given.

[0034] FIG. 4 is a diagram showing an overview of another example of automatic response by the chatbot 31 in Embodiment 1 of the present invention. Similar to FIG. 3 described above, in the upper and lower parts of the figure, examples of chat screens related to respective inquiries displayed on the user terminals 3a and 3b are shown. In the example of FIG. 4, the chatbots 31a and 31b corresponding to respective inquiries respectively obtain individual answers from the individual FAQs 52a and 52b in which FAQs unique to the respondents they correspond to are registered, so that for similar inquiries (in the example in the figure, "I want to do ▲▲, but what should I do?"), different answers corresponding to the respondents (in the example in the figure, "Please do ■■ and then △△." and "If there is ○○, you can do it by □□.") are given.

[0035] FIG. 5 is a diagram showing an overview of an example of handover from automatic response by the chatbot 31 to manned response by a responder in Embodiment 1 of the present invention. FIG. 5 shows an example of a chat screen related to an inquiry displayed on the user terminal 3a. In the example of FIG. 5, when it is determined that it is no longer possible to perform automatic response using the individual FAQ 52a by the chatbot 31a for an inquiry from the user (in the example in the figure, "I cannot do ●● with □□."), that is, when the FAQ registered in the individual FAQ 52a cannot provide an answer, or when no corresponding answer can be extracted by keyword pattern matching, a message indicating handover to manned response (in the example in the figure, "I will replace the person in charge. Please wait.") is given as an answer, and it is shown that the transferred responder 4a answers in place of the chatbot 31a (in the example in the figure, "I am ××× in charge. How is it displayed on ▲▲?").

[0036] <Flow of processing> FIG. 6 is a flowchart showing an overview of an example of the flow of processing until a chatbot 31 that answers an inquiry from a user in Embodiment 1 of the present invention is assigned. When the inquiry reception unit 10 of the inquiry response system 1 receives an inquiry from the user terminal 3, it acquires information on the currently logged-in responder via the responder management unit 20 (S01). Then, it determines whether there is an available responder among the currently logged-in responders, that is, whether there is an available responder who is not in the middle of manned response (S02).

[0037] If there is no available responder (No in step S02), the inquiry is added to the queue waiting for response (S03), and the process returns to step S01 and repeats until an available responder appears. If the user does not wish to wait for an answer, or if the number of inquiries remaining in the queue exceeds a predetermined threshold, or if there is not even one responder logged in to the inquiry response system 1 in the first place, etc., instead of adding it to the queue waiting for response, it may be possible to answer the user that it is currently not possible to respond or that it will be responded to at a later date and end the process. In that case, for example, an invitation email may be sent to potential responders, and they may be contacted again after logging in to respond to the user.

[0038] If there is a respondent in the idle state (Yes in step S02), a corresponding respondent is determined from among the respondents in the idle state based on a predetermined standard or condition (S04). The criteria or conditions for determining the respondent are not particularly limited. For example, it may be preferentially determined from respondents with a long waiting state after login or respondents with a schedule that has some leeway, or it may be determined based on the priority set for each respondent. It may also be determined by matching the content of the inquiry with the skill level and area of expertise of the respondent.

[0039] When the respondent is determined, after the inquiry processing unit 30 identifies the individual response model 70 corresponding to the respondent, the chatbot 31 (or its instance) is activated in a form that uses this (S05). By the above series of processes, the assignment of the respondent is completed, and the automatic response by the chatbot 31 to the inquiry is started.

[0040] As described above, according to the inquiry response system 1 which is the first embodiment of the present invention, in response to an inquiry from a user, while automatically responding by the chatbot 31 as a "doppelganger" individually associated with the respondent, when the response becomes difficult, it is possible to seamlessly transfer to a manned response by the respondent who is the "owner" of the "doppelganger" without changing the characteristics and features in the response to the user.

[0041] (Second Embodiment) The inquiry response system which is the second embodiment of the present invention has the same system configuration as the inquiry response system 1 of the above-described first embodiment. For example, when the bases where the respondents are located are dispersed in a plurality of regions with a time difference such as overseas, it switches which base's respondent is preferentially assigned depending on the time zone when the inquiry is received. For example, if the time when the inquiry is received falls between 8:00 and 20:00 in Japanese time, the respondents belonging to the Japanese base are preferentially assigned, and in other cases, the respondents belonging to the US base are preferentially assigned.

[0042] Such priority assignment settings can be changed by setting the location and time zone. Here, the explanation was given in terms of locations, but instead of setting by location unit, it may be configured to set the priority time zone for each person in charge of multiple locations and execute processing for each person in charge. Also, in this embodiment, it is assumed that responses are made in one language (English) at multiple overseas locations. In that case, information for FAQs and AI learning also needs to be stored and learned in that language. Furthermore, some respondents are capable of handling multiple languages. In that case, FAQs and AI learning for each language are required. And when the user explicitly specifies a language (including when selecting a language on the website), or when the language is switched according to the language input by the user to the chatbot, it may be configured such that the respondent and the chatbot that match that language operate. Note that respondents capable of handling multiple languages can also set the language to be prioritized or their native language. When such a setting is made, it is more desirable to configure the system to respond to users who have specified a certain language with a respondent and a chatbot that prioritize that language or consider it their native language.

[0043] <Flow of processing> FIG. 7 is a flowchart showing an overview of an example of the flow of processing until a chatbot 31 that responds to an inquiry from a user in Embodiment 2 of the present invention is assigned. When the inquiry reception unit 10 of the inquiry response system 1 receives an inquiry from the user terminal 3, it acquires the time at that point (S11), and via the respondent management unit 20, acquires information on one or more respondents (hereinafter sometimes referred to as "priority respondents") who belong to one or more locations that should be preferentially assigned at that time among the currently logged-in respondents (S12). Information on which location each respondent belongs to is registered in advance in the respondent master 41. Then, it is determined whether there is an available priority respondent among the currently logged-in priority respondents, that is, whether there is a priority respondent in an available state where they are not engaged in person-to-person correspondence (S13).

[0044] If there is no priority responder in the idle state (No in step S13), information on responders other than priority responders among the logged-in responders (hereinafter sometimes referred to as "non-priority responders") is obtained via the responder management unit 20 (S14). Then, it is determined whether there is an available non-priority responder among the logged-in non-priority responders, that is, whether there is a non-priority responder in an idle state where they are not in a manned response (S15).

[0045] If there is also no available non-priority responder (No in step S15), the inquiry is added to the queue waiting for a response by the priority responder (S16), and the process returns to step S11 and is repeated until an available responder appears. If the user does not wish to wait for a response, or if the number of inquiries remaining in the queue is more than a predetermined threshold, or if there are no responders logged in to the inquiry response system 1 at all, etc., instead of adding it to the queue waiting for a response by the priority responder, it may be possible to respond to the user that it is currently not possible to handle and that it will be handled at a later date, etc., and then end the process. In that case, for example, an invitation email may be sent to potential priority responders, and after they log in, they may be contacted back to the user for handling.

[0046] If there is an available responder (Yes in step S13 or S15), a responder to handle the inquiry is determined from among the available responders based on a predetermined criterion or condition (S17). Similar to the above-described Embodiment 1, the criterion or condition for determining the responder is not particularly limited. For example, it may be preferentially determined from responders with a long standby state after logging in or responders with a schedule that allows for spare time, or it may be determined based on the priority level set for each responder. It may also be determined by matching the inquiry content with the skill level and area of expertise of the responder.

[0047] Once the responder is determined, the inquiry processing unit 30 identifies the individual response model 70 corresponding to the responder, and then starts the chatbot 31 (or its instance) in a form that uses this (S18). Through the above series of processes, the assignment of the responder is completed, and the automatic response by the chatbot 31 to the inquiry is started.

[0048] In addition, for example, when an inquiry from a user is assigned to a non-priority responder, it may happen that the inquiry is in a language of a region different from the language of the base where the non-priority responder belongs. In this case, for example, by multilingualizing the content of the chatbot 31, the common FAQ 51, and the individual FAQ 52, it is possible to respond in the language used by the user at least in the automatic response by the chatbot 31.

[0049] As described above, according to the inquiry response system 1 which is the second embodiment of the present invention, when the bases where the responders are located are dispersed in a plurality of regions with time differences such as overseas, it is possible to switch which base's responder to preferentially assign depending on the time zone when the inquiry is received. Further, even when an assignment is made to a non-priority responder (usually at night for the non-priority responder), it is possible to reduce the response load on the non-priority responder by first automatically responding with the chatbot 31.

[0050] (Embodiment 3) The inquiry response system which is the third embodiment of the present invention applies the same system configuration as the inquiry response system 1 of the above-described first embodiment, instead of applying it to the response to an inquiry from a user at a help desk, a support center, etc. as described so far. For example, it is applied to the automatic response to a real-time question in a chat function that is performed in parallel during an online seminar using a web conferencing system such as Zoom (registered trademark).

[0051] FIG. 8 is a diagram showing an outline of an example of the response to a user in the third embodiment of the present invention. FIG. 8 shows an example of a screen 5a of a web conferencing system displayed on an information processing terminal such as a PC used by an instructor 4d of an online seminar. The instructor 4d is shown conducting an online seminar with himself / herself reflected in a web camera 5b for each attendee displayed on the screen 5a. In addition, on the right side of the screen 5a, it is shown that a chat is being conducted with the attendees during the lecture by means of a chat function provided in many web conferencing systems.

[0052] At this time, for questions in the chat from the participants made during the lecture, instead of the lecturer 4d answering in the chat in real time, the chat function of the web conferencing system and the inquiry processing unit 30 of the inquiry response system 1 of the present embodiment cooperate (for example, using the chat function of the web conferencing system as the chat processing unit 32), so that the chatbot 31d, which is the "avatar" of the lecturer 4d, answers in real time on behalf of the lecturer 4d. That is, for an inquiry in the chat from a participant who is a user, the respondent management unit 20 of the inquiry response system 1 assigns the lecturer 4d as the respondent, and the chatbot 31d automatically answers.

[0053] As a result, even when the lecturer 4d is actually giving a lecture to each participant in an online seminar, it is possible to answer questions in the chat from the participants in a timely manner based on the characteristics, features, knowledge, etc. of the lecturer 4d. For example, in an online seminar or the like, there is often a question-and-answer time provided at the end, but for general questions, the chatbot 31d can answer in real time during the seminar to anticipate and avoid wasting the precious question-and-answer time on general questions. Note that the operation content may be changed as appropriate according to the situation of the online seminar, such as enabling the chatbot 31d to automatically answer questions in the chat by the participants during the question-and-answer time or after the seminar ends.

[0054] In an online seminar, it is considered that there are many cases where it is meaningful for a participant to attend a seminar by a specific lecturer 4d. Therefore, when the chatbot 31d answers questions in the chat from the participants during the lecture by the lecturer 4d, for example, the icon or name of the chatbot 31d on the chat screen may be made to be something that can be recognized as unique to the lecturer 4d, so as to emphasize the situation of receiving an answer from a specific lecturer 4d.

[0055] In addition, when a question that is difficult for the chatbot 31d to answer sufficiently is asked, instead of transferring the live chat support to the instructor 4d as shown in the examples of Embodiments 1 and 2, for example, the fact that such a question has been asked may be highlighted or notified on the screen. For such a question, the instructor 4d can directly answer after confirming the question content during the Q&A session, for example. Not only for such questions, but also for questions that have been answered and resolved by the chatbot 31d based on the correspondence history 61, etc., it is also possible to take actions such as adding supplementary explanations after confirmation.

[0056] In addition, when the chat function of the web conferencing system is a chat in which all participants attend, when the instructor 4d answers a question in the chat, in order to make it clear which question the answer is for, it is desirable to have a mechanism such as specifying the question targeted by the answer and answering it, or duplicating the targeted question and re-entering / displaying it and answering it.

[0057] Also, for example, when a plurality of instructors conduct lectures or classes in sequence, while the first instructor is giving a lecture or the like, the chatbot 31 of the "avatar" of the first instructor automatically responds, and while the next second instructor is giving a lecture or the like, the chatbot 31 of the "avatar" of the second instructor automatically responds. After the end of the lecture or the like by the first instructor, the first instructor himself / herself may take over and provide live support for questions related to the lecture.

[0058] In any aspect, based on the content of the live chat support by the instructor recorded in the correspondence history 61, the individual answer model 70 related to the instructor can be updated by learning with an AI engine (not shown), or FAQs can be extracted and registered in the common FAQ 51 or the individual FAQ 52, etc., so that it can be made available during the next online seminar.

[0059] As described above, according to the inquiry response system 1 which is the third embodiment of the present invention, for real-time questions in the chat function that are conducted in parallel during an online seminar using a web conferencing system, the chatbot 31 as the "avatar" of the lecturer automatically responds, reducing the burden on the lecturer, making the Q&A time during the seminar more beneficial, and enabling the attendees to more keenly sense the significance of taking the online seminar by the lecturer.

[0060] Here, an example of an online seminar lecturer has been used for the explanation. However, it is also possible to apply this embodiment to the classes of teachers in schools, cram schools, etc. Instead of asking the teacher at the end of the target class when something is not understood in the teacher's explanation, when a student does not understand something, the student can ask questions in real time (each time) to the "avatar" of the teacher of the target class, enabling the student to take the class while fully understanding the content, and enabling the teacher to respond only to new questions at around the end of the class or after the class, making efficient response possible.

[0061] In addition, in classes at cram schools, etc., there is also a video playback system in which recorded classes are played back and used. By introducing the "avatar" function of the system of this embodiment into this system, if a student does not understand something while playing the video, the student can stop the playback and ask questions to the "avatar". If the student still does not understand, it is also possible to individually inquire and respond to the teacher or the consultation window. In particular, when the lecturer is a popular lecturer or a famous lecturer, the "avatar" function can provide an attractive service by enabling the student to obtain answers to questions (it may also be an optional function (separate charge)).

[0062] Furthermore, in the case of classes, there are subjects such as "Mathematics", "English", "Japanese", "Science", and "Social Studies". Within these subjects, there are learning themes (for example, "factorization" and "quadratic functions" in the case of "Mathematics"). It may be configured such that the "avatars" of the instructors for each subject of the target class answer questions, or it can also be configured such that the "avatars" of the instructors for each learning theme of the instructors for each subject of the target class answer questions. To implement this, for the former, the learning model is formed to have an avatar function for each instructor, and for the latter, the learning model is formed to have an avatar function for each instructor and each learning theme. The former "avatar" is more widely learned by subject and can answer almost any question. The latter "avatar" is learned by learning theme and can answer questions with higher accuracy.

[0063] As described above, the invention made by the present inventor has been specifically described based on the embodiments. However, it goes without saying that the present invention is not limited to the above-described embodiments and can be variously modified without departing from the gist thereof. Also, the above-described embodiments have been described in detail for easy understanding of the present invention and are not necessarily limited to those having all the configurations described. Further, it is possible to add, delete, or replace a part of the configuration of the above-described embodiments with other configurations.

[0064] For example, in the above-described embodiments, the interaction between the user who makes an inquiry or asks a question and the inquiry response system 1 has been mainly described with respect to text-based chat. However, for example, the speech of the user or the responder may be converted into text by speech recognition and used as an input to the chat, or the text input to the chat by the user or the responder may be converted into speech (read aloud) and output. Also, by combining such speech recognition and speech conversion functions, it is possible to apply them to oral questions and responses in online seminars, inquiries and responses by phone, etc.

[0065] In addition, some or all of the above-described configurations, functions, processing units, processing means, etc. may be implemented in hardware by designing them, for example, in an integrated circuit. Also, the above-described configurations, functions, etc. may be implemented in software by a processor interpreting and executing a program for realizing each function. Information such as programs, tables, files, etc. for realizing each function can be placed in a recording device such as a memory, a hard disk, an SSD (Solid State Drive), or a recording medium such as an IC card, an SD card, or a DVD.

[0066] In addition, in each of the above figures, control lines and information lines show those considered necessary for explanation, and do not necessarily show all control lines and information lines in implementation. In reality, it may be considered that almost all configurations are interconnected.

Industrial Applicability

[0067] The present invention can be used in an inquiry response system that combines automatic response by a chatbot and human response for online inquiries from users.

Explanation of Signs

[0068] 1... Inquiry response system, 2... Network, 3... User terminal, 3a~3c... User terminals, 4a~4c... Respondents, 4d... Instructor, 5a... Screen, 5b... Web camera, 10... Inquiry reception unit, 20... Respondent management unit, 30... Inquiry processing unit, 31... Chatbot, 31a~31c... Chatbots, 32... Chat processing unit, 41... Respondent master, 42... Respondent status, 51... Common FAQ, 52... Individual FAQ, 52a, 52b... Individual FAQs, 61... Response history, 70... Individual answer model, 70a~70c... Individual answer models

Claims

1. An inquiry response system that responds to online inquiries by chat from a user, comprising: an inquiry reception unit that receives an inquiry from the user; a responder management unit that assigns a responder to the inquiry received by the inquiry reception unit; an inquiry processing unit that corresponds to the inquiry by a chatbot corresponding to the responder assigned by the responder management unit, and when it is determined that the chatbot cannot respond, takes over the response by chat by the assigned responder. An inquiry response system having.

2. In the inquiry response system according to Claim 1, the responder management unit assigns a responder to the inquiry from among responders who are logged in to the inquiry response system and are in an idle state where they are not in a manned response. Inquiry response system.

3. In the inquiry response system according to Claim 2, the responder management unit switches the priority assignment base for each base to which each responder belongs according to the time zone when the inquiry is received. Inquiry response system.

4. In the inquiry response system according to Claim 2, the responder management unit adds the inquiry to a queue waiting for response when there is no idle responder. Inquiry response system.

5. In the inquiry response system according to Claim 1, when the inquiry processing unit determines that the chatbot cannot respond, it takes over the manned response by another responder different from the assigned responder or the response by another chatbot. Inquiry response system.

6. In the inquiry response system according to Claim 1, the user is a participant in an online seminar, the responder management unit assigns a lecturer who is giving a lecture in the online seminar as a responder to the inquiry from the user who is a participant. Inquiry response system.

Images