Response support device, response support method, and response support program

Abstract

To appropriately distribute inquiries to operators. [Solution] An inquiry distribution unit 26 sets each of multiple inquiries from station users as a target inquiry. The inquiry distribution unit 26 determines a priority indicating the degree to which an operator will prioritize responding to the target inquiry based on the urgency identified from the category of the target inquiry and the response waiting time for the target inquiry. The inquiry distribution unit 26 displays a confirmation button in the column of an inquiry with a higher priority than a standard on an inquiry list screen showing columns of multiple inquiries displayed on an operator terminal used by an operator.

Description

Technical Field

[0001] This disclosure relates to a technology for assisting in responding to inquiries at stations.

Background Art

[0002] In recent years, the number of unmanned stations has been increasing. It is necessary to create an environment where various people can use each station without inconvenience. To achieve such an environment, it is desirable to enable remote response to inquiries at the station counter even when there is no station staff at the inquiry station. Also, it is desirable to enable response to inquiries at the station counter even through telework or the like. However, remotely responding to a wide variety of inquiries from multiple stations and multiple counters is a complex and difficult situation. Therefore, there is a need to assist the operator's response to inquiries.

[0003] Patent Document 1 describes that when an inquiry input at a station guidance terminal is a stereotyped inquiry, a stereotyped answer is given, and when it is not a stereotyped inquiry, it is connected to a staff terminal.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0005] There may be a large number of inquiries from multiple stations. The number of operators is limited, and users may have to wait for operator assignment. In this case, it is necessary to appropriately allocate the inquiries to the operators. This disclosure aims to enable appropriate allocation of inquiries to operators.

Means for Solving the Problems

[0006] The response support device relating to this disclosure is Selection information for selecting which of multiple inquiries from station users will be prioritized for an operator, the selection information includes the category of the content of the target inquiry and the response waiting time, which is the elapsed time since the target inquiry was made. An inquiry distribution unit determines a priority indicating the degree to which the operator should prioritize handling the target inquiry, based on the urgency identified from the category included in the selection information for the target inquiry received by the input unit and the response waiting time included in the selection information for the target inquiry, and displays a confirmation button in the column of the inquiry with a priority higher than the standard on the inquiry list screen showing the columns of the multiple inquiries displayed on the operator terminal used by the operator. It is equipped with. [Effects of the Invention]

[0007] This disclosure determines the priority level to which an operator will prioritize an inquiry based on the urgency and response time identified from the category. This makes it possible to appropriately route inquiries to operators. [Brief explanation of the drawing]

[0008] [Figure 1] Configuration diagram of the response support device 10 according to Embodiment 1. [Figure 2] A flowchart illustrating the operation of the response support device 10 according to Embodiment 1. [Figure 3] A diagram illustrating the operation of the response support device 10 according to Embodiment 1. [Figure 4] A flowchart illustrating the operation of the response support device 10 according to Embodiment 2. [Figure 5] A diagram illustrating the operation of the response support device 10 according to Embodiment 2. [Figure 6]Flowchart showing the operation of the response support device 10 according to Embodiment 3. [Figure 7] Explanatory diagram of the operation of the response support device 10 according to Embodiment 3. [Figure 8] Flowchart showing the operation of the response support device 10 according to Embodiment 4. [Figure 9] Explanatory diagram of the operation of the response support device 10 according to Embodiment 4. [Figure 10] Flowchart showing the operation of the response support device 10 according to Embodiment 5. [Figure 11] Explanatory diagram of the operation of the response support device 10 according to Embodiment 5. [Figure 12] Configuration diagram of the response support device 10 according to Modified Example 1. [Figure 13] Flowchart showing the operation of the response support device 10 according to Modified Example 1. [Figure 14] Explanatory diagram of the operation of the response support device 10 according to Modified Example 1. [Figure 15] Configuration diagram of the response support device 10 according to Embodiment 6. [Figure 16] Flowchart showing the operation of the response support device 10 according to Embodiment 6. [Figure 17] Explanatory diagram of the operation of the response support device 10 according to Embodiment 6. [Figure 18] Configuration diagram of the response support device 10 according to Embodiment 7. [Figure 19] Flowchart showing the operation of the response support device 10 according to Embodiment 7. [Figure 20] Diagram showing the list screen of inquiries according to Embodiment 7. [Figure 21] Configuration diagram of the response support device 10 according to Modified Example 3. [Figure 22] Configuration diagram of the response support device 10 according to Embodiment 8. [Figure 23] Flowchart showing the operation of the response support device 10 according to Embodiment 8. [Figure 24] Configuration diagram of the response support device 10 according to Embodiment 9.

Modes for Carrying Out the Invention

[0009] Embodiment 1. ***Explanation of the structure*** Referring to Figure 1, the configuration of the response support device 10 according to Embodiment 1 will be described. The response support device 10 is a computer. The response support device 10 comprises hardware including a processor 11, memory 12, storage 13, and a communication interface 14. The processor 11 is connected to the other hardware via signal lines and controls this other hardware.

[0010] Processor 11 is an IC that performs processing. IC stands for Integrated Circuit. Specific examples of processor 11 include CPU, DSP, and GPU. CPU stands for Central Processing Unit. DSP stands for Digital Signal Processor. GPU stands for Graphics Processing Unit.

[0011] Memory 12 is a storage device that temporarily stores data. Specific examples of memory 12 include SRAM and DRAM. SRAM stands for Static Random Access Memory. DRAM stands for Dynamic Random Access Memory.

[0012] Storage 13 is a storage device for storing data. A concrete example of storage 13 is an HDD. HDD stands for Hard Disk Drive. Alternatively, storage 13 may be a portable recording medium such as an SD® memory card, CompactFlash®, NAND flash, flexible disk, optical disk, compact disk, Blu-ray® disc, or DVD. SD stands for Secure Digital. DVD stands for Digital Versatile Disk.

[0013] Communication interface 14 is an interface for communicating with external devices. Specific examples of communication interface 14 include Ethernet®, USB, and HD. This is an MI (registered trademark) port. USB stands for Universal Serial Bus. HDMI stands for High-Definition Multimedia Interface. The response support device 10 is connected to the operator terminal 51 and the remote counter terminal 52, etc., via the communication interface 14.

[0014] The response support device 10 comprises an input unit 21, a model control unit 22, and an output unit 23 as functional components. The functions of each functional component of the response support device 10 are realized by software. Storage 13 stores programs that implement the functions of each functional component of the response support device 10. These programs are loaded into memory 12 by the processor 11 and executed by the processor 11. This enables the implementation of the functions of each functional component of the response support device 10.

[0015] The response generation model 31 is stored in storage 13. However, the response generation model 31 may also be stored in an external storage device of the response support device 10, rather than in storage 13.

[0016] In Figure 1, only one processor 11 was shown. However, there may be multiple processors 11, and multiple processors 11 may work together to execute programs that implement each function.

[0017] ***Explanation of operation*** The operation of the response support device 10 according to Embodiment 1 will be explained with reference to Figures 2 and 3. The operation procedure of the response support device 10 according to Embodiment 1 corresponds to the response support method according to Embodiment 1. Furthermore, the program that realizes the operation of the response support device 10 according to Embodiment 1 corresponds to the response support program according to Embodiment 1.

[0018] (Step S11: Input Processing) The input unit 21 receives inquiry information 41 from users at the station's ticket window. Specifically, an operator is called from a remote counter terminal 52 installed at the station's ticket window via an operator terminal 51 used by the operator, and the user and the operator converse. The conversation is conducted by voice or text. The text data of the conversation between the user and the operator constitutes the inquiry information 41. If the conversation is conducted by voice, the voice is converted into text to generate the inquiry information 41. Note that the inquiry information 41 may consist only of the user's statements from the conversation. Furthermore, the user may be required to input the content of their inquiry before calling an operator from the remote support terminal 52, etc. In this case, the information representing the inputted inquiry content is the inquiry information 41.

[0019] (Step S12: Model control processing) The model control unit 22 inputs the inquiry information 41 received in step S11 into the response generation model 31, which is a learning model. The response generation model 31 is given the station operations manual 42 in advance. In other words, the response generation model 31 is trained using the station operations manual 42 as training data. Alternatively, the response generation model 31 has access to the station operations manual 42. The station operations manual 42 also contains information on how to handle inquiries. When the model control unit 22 inputs the inquiry information 41, it also inputs instructions to the response generation model 31 to generate a proposed response 43 for the inquiry information 41 and to identify the basis description 44 in the station operations manual 42 that will serve as the basis for the proposed response 43. In other words, the model control unit 22, as a prompt, The user then inputs the inquiry information 41, along with instructions to generate a response proposal 43 and identify the basis for the response 44. The model control unit 22 then obtains the proposed response 43 output by the response generation model 31 and the basis description 44 in the station operations manual 42 that serves as the basis for the proposed response 43.

[0020] The learning model is what is known as generative AI or LLM. AI stands for Artificial Intelligence. LLM is an abbreviation for Large Language Model. The learning model may be constructed using algorithms such as BERT and GPT. BERT is a Bidirectional Encoder. GPT stands for Representations from Transformers. A learning model may be constructed by combining multiple algorithms, including these algorithms. The learning model may be a trainable model. If the learning model is trainable, it will be trained using the operator's responses. If the learning model is a generative AI, it is assumed that the learning model will be stored in an external storage device of the response support device 10.

[0021] (Step S13: Output processing) The output unit 23 outputs the proposed response 43 and the justification statement 44 obtained in step S12. Specifically, the output unit 23 outputs the proposed response 43 and the justification statement 44 to the operator terminal 51 or the like used by the operator for display. This allows the operator to refer to the proposed response 43 and the justification statement 44 and respond to the user. Here, when displaying the basis description 44, instead of displaying an excerpt of the basis description 44 in the station manual 42, the page number of the basis description 44 in the station manual 42 may be displayed. This would allow the operator to check the descriptions before and after the basis description 44, potentially leading to a more appropriate response.

[0022] ***Effects of Embodiment 1*** As described above, the response support device 10 according to Embodiment 1 uses the response generation model 31 to obtain a proposed response 43 corresponding to the inquiry information 41 and the basis description 44 for the proposed response 43 in the station operations manual 42. By using the response generation model 31, it is possible to generate a proposed response 43 even for non-standard inquiries. In addition, since the basis description 44 for the proposed response 43 in the station operations manual 42 is obtained, the operator can provide an appropriate response.

[0023] Embodiment 2. Embodiment 2 differs from Embodiment 1 in that additional information 45 is input. Embodiment 2 will explain this difference, and the same points will not be explained.

[0024] ***Explanation of operation*** The operation of the response support device 10 according to Embodiment 2 will be explained with reference to Figures 4 and 5. The process in step S23 is the same as the process in step S13 in Figure 2.

[0025] (Step S21: Input Processing) The input unit 21 accepts additional information 45 in addition to the inquiry information 41. Here, the input unit 21 may accept the additional information 45 directly, or it may accept information that can identify the additional information 45. More details will be described later.

[0026] (Step S22: Model control processing) The model control unit 22 inputs the additional information 45 received in step S21, in addition to the inquiry information 41, into the response generation model 31. The model control unit 22 then retrieves the proposed response 43 and the justification statement 44 output by the response generation model 31.

[0027] Additional information 45 includes information such as the location of ticket counters, train operation information, train fare information, equipment information, user personal information, ticket information, lost and found information, map information, and event information.

[0028] The counter information indicates the location of the counter to which the user is making an inquiry. The counter location includes the station name and ticket gate name of the station where the counter is located. The counter information can be identified from the remote counter terminal 52 to which the operator terminal 51 is connected. Therefore, if the additional information 45 is counter information, the input unit 21 receives terminal identification information from the remote counter terminal 52 to which the operator terminal 51 is connected. The input unit 21 then identifies the counter information from the terminal identification information. In other words, when an inquiry is entered by a user and an operator responds, the input unit 21 obtains terminal identification information from the remote counter terminal 52 to which the responding operator's terminal 51 is connected. The input unit 21 then identifies the counter information from the remote counter terminal 52's identification information by referring to the pre-defined correspondence information between the remote counter terminal 52's identification information and the counter information. In this way, the input unit 21 obtains the identified counter information as additional information 45 in addition to the inquiry information 41. The response generation model 31 can output station- or ticket gate-specific response proposals 43 and supporting documentation 44, such as the nearest exit, by taking into account counter information.

[0029] In addition to the information provided at the ticket window, the names of the user's destination station and intermediate stations may also be entered as additional information 45. For example, if a user inquires about their destination, the input unit 21 may obtain the user's destination station and intermediate station names as additional information 45. When the input unit 21 recognizes an inquiry that includes the destination station and intermediate station names through speech recognition, it is conceivable that the recognized destination station and intermediate station names will be used as additional information 45. The response generation model 31 can output a route as a suggested response 43 by considering the user's destination station and intermediate station names.

[0030] Train operation information indicates the operational status of trains, such as whether they are running normally, or if there are delays or cancellations. Train operation information can be obtained from the system that manages train operations. Therefore, if the additional information 45 is operation information, the input unit 21 accepts the input of operation information from the system that manages train operations. For example, when the input unit 21 receives an inquiry from a user, it accepts input of the category that the inquiry concerns train operation information. If the input unit 21 receives the category that the inquiry concerns train operation information in addition to the inquiry information 41, it obtains the operation information as additional information 45 from the system that manages train operations. The input unit 21 may also recognize the content of the user's inquiry and, if it recognizes that it is an inquiry concerning train operations, it may also include train operation information as additional information 45. The input unit 21 identifies the information to be obtained and the source of the information to be obtained by referring to information that has been previously indicated according to the content of the inquiry. The method of recognizing the content of the inquiry may be any method, such as using AI. The response generation model 31 can output a proposed response 43 and a justification statement 44 that take into account the operating status by considering train operation information. For example, the response generation model 31 can output a proposed route or boarding time that takes the operating status into account as the proposed response 43.

[0031] Train fare information is information that shows the train fare for each section of travel. Train fare information can be obtained from the fare management system. Therefore, if the additional information 45 is fare information, the input unit 21 accepts the input of fare information from the fare management system. For example, when the input unit 21 receives an inquiry from a user, it provides information regarding train fare information. The input unit accepts input in the category of "inquiry." When the input unit 21 receives the category "inquiry about train fare information" in addition to the inquiry information 41, it retrieves fare information as additional information 45 from the fare management system. The input unit 21 may also recognize the content of the user's inquiry and, if it recognizes that it is an inquiry about train fares, include fare information as additional information 45. The response generation model 31, by considering train fare information, can output a proposed response 43 and justification statement 44 that take into account competing sections. A competing section is a section where multiple railway companies' lines run. In competing sections, fares may be set lower to take into account the fares of other companies. This makes it possible to output an appropriate proposed response 43 even for inquiries regarding fares with such special circumstances.

[0032] Equipment information refers to information managed by the railway company regarding equipment installed at stations that is available to users. Specifically, equipment information is data managed by the railway company regarding equipment such as automatic ticket gates, ticket vending machines, and fare adjustment machines. Equipment information can be obtained from the equipment management system. Therefore, if the additional information 45 is equipment information, the input unit 21 accepts the input of equipment information from the equipment management system. For example, when the input unit 21 receives an inquiry from a user, it accepts input of the category that the inquiry was about not being able to pass through the ticket gate. If the input unit 21 receives the category that the inquiry was about not being able to pass through the ticket gate in addition to the inquiry information 41, it obtains equipment information about the automatic ticket gate from the equipment management system as additional information 45. The input unit 21 may also recognize the content of the user's inquiry and, if it recognizes that the inquiry was about not being able to pass through the ticket gate, obtain equipment information about the automatic ticket gate as additional information 45. The response generation model 31 can output a suggested response 43 and a justification statement 44 that correspond to the status of the device operated by the user, by taking device information into consideration. For example, if a user inquires that they were unable to pass through the ticket gate, the model can output a suggested response 43 and a justification statement 44 that take into account the cause, such as insufficient balance or no entry history.

[0033] The user's personal information includes information such as the user's age and gender. The user's personal information is entered by the user using the remote service terminal 52. The input unit 21 receives the user's personal information from the remote service terminal 52. The input unit 21 may also obtain the user's personal information from the remote service terminal 52 by using the recognition results of the video from the camera installed on the remote service terminal 52, the information on the IC card read by the remote service terminal 52, the results of voice recognition, etc. The response generation model 31 can output response suggestions 43 and supporting statements 44 tailored to characteristics such as age and gender by taking into account the user's personal information. For example, it can output response suggestions 43 and supporting statements 44 that are easy for children to understand.

[0034] Ticket information refers to information about the ticket held by the user. The ticket includes both the basic fare ticket and the express train ticket. Ticket information includes the fare, identification information for the train to be boarded, the seat number, etc. Ticket information is entered by the user using the remote ticket terminal 52. In this case, the information printed on the ticket may be manually entered by the user, or the information on the ticket may be read using a scanner. In addition, if ticket information is set on an IC card, the ticket information may be read from the IC card and entered using the card reader installed in the remote ticket terminal 52. The input unit 21 accepts the input of ticket information from the remote ticket terminal 52. By considering ticket information, the response generation model 31 can output a proposed response 43 and supporting documentation 44 that take into account the contents of the ticket held.

[0035] Lost and found information refers to information about items lost on trains that passengers were riding. This information includes the classification of the lost item and information about the train the passenger was on at the time of the loss. Lost item information is entered by the user using the remote service terminal 52. Therefore, the input unit 21 receives lost item information from the remote service terminal 52. For example, when the input unit 21 receives an inquiry from a user, it accepts input of the category that the inquiry is about a lost item. If the input unit 21 receives the category that the inquiry is about a lost item in addition to the inquiry information 41, it obtains lost item information as additional information 45 from the remote service terminal 52. The input unit 21 may also recognize the content of the user's inquiry and, if it recognizes that the inquiry is about a lost item, obtain lost item information as additional information 45. The response generation model 31 can output a response proposal 43 and supporting documentation 44 that include specific contact information, by taking into account information about the lost item. Specific contact information may include the name of a train station or police station.

[0036] The map information displays a map of the area around the station. Map information can be obtained from an external map management server, etc. Therefore, the input unit 21 accepts map information input from the external map management server, etc. In this case, the input unit 21 may also accept map information input corresponding to the location of the counter indicated by the counter information. For example, when the input unit 21 receives an inquiry from a user, it accepts input of the category that it is an inquiry about the surrounding area. If the input unit 21 receives the category that it is an inquiry about the surrounding area in addition to the inquiry information 41, it obtains map information as additional information 45 from the map management server or the like. The input unit 21 may also recognize the content of the user's inquiry and, if it recognizes that it is an inquiry about the surrounding area, obtain map information as additional information 45. The response generation model 31 can output a response proposal 43 showing a specific route and a statement of justification 44 by taking map information into consideration.

[0037] Event information refers to information about events held around the station. This information includes details such as the event location and time. Event information can be obtained from an external event management server. Therefore, the input unit 21 receives event information from the event management server. The response generation model 31 can output response proposals 43 and justification statements 44, such as directions that avoid congestion, by taking event information into consideration.

[0038] ***Effects of Embodiment 2*** As described above, the response support device 10 according to Embodiment 2 inputs additional information 45 in addition to the inquiry information 41 to the response generation model 31. This makes it possible to obtain a more appropriate response proposal 43 and justification statement 44. As a result, the operator can provide an appropriate response.

[0039] Embodiment 3. Embodiment 3 differs from Embodiments 1 and 2 in that past response history 46 is provided to the response generation model 31. Embodiment 3 explains this difference, while the same points are omitted from the explanation. Embodiment 3 describes a case in which a modification has been made to Embodiment 1. However, it is also possible to modify Embodiment 2.

[0040] ***Explanation of operation*** The operation of the response support device 10 according to Embodiment 3 will be explained with reference to Figures 6 and 7. The processes in steps S31 and S33 are the same as steps S11 and S13 in Figure 2.

[0041] (Step S32: Model control processing) The model control unit 22 inputs the inquiry information 41 to the response generation model 31, similar to Embodiment 1. In Embodiment 3, the response generation model 31 is given past response history 46 in addition to the station manual 42. In other words, the response generation model 31 is trained using the response history 46 as training data. Alternatively, the response generation model 31 can access the response history 46. The response history 46 is a combination of past inquiries from users and the responses given to users in response to those inquiries. In other words, the response history 46 is a combination of the inquiry information 41 and the responses given by operators to users in response to the inquiries in the inquiry information 41, and is stored in the storage 13 in Figure 1. The response may be a draft response 43 output by the response generation model 31. However, the response may also be a modified version of the draft response 43 created by the operator, or a response created by a completely different operator. The model control unit 22 then acquires the proposed response 43 and the justification statement 44 output by the response generation model 31.

[0042] ***Effects of Embodiment 3*** As described above, the response support device 10 according to Embodiment 3 is provided with a response generation model 31 containing past response history 46. This makes it possible to generate a response proposal 43 based on the response content given by the operator to the user if a similar inquiry has been made in the past. For example, if an inquiry is made from the ticket window of a new station, it becomes possible to generate a response proposal 43 based on the response history of similar inquiries made in the past from other similar stations. This makes it possible to obtain a more appropriate response proposal 43 and justification statement 44. As a result, the operator is able to provide an appropriate response.

[0043] Embodiment 4. Embodiment 4 differs from Embodiments 1 to 3 in that the classification 47, which is the premise for generating the proposed response 43, is input into the response generation model 31. Embodiment 4 explains this difference, while the same points are omitted from the explanation. Embodiment 4 describes a case in which a modification has been made to Embodiment 1. However, it is also possible to modify Embodiments 2 and 3.

[0044] ***Explanation of operation*** The operation of the response support device 10 according to Embodiment 4 will be explained with reference to Figures 8 and 9. The process in step S43 is the same as the process in step S13 in Figure 2.

[0045] (Step S41: Input Processing) The input unit 21 accepts the input of classification 47 in addition to the inquiry information 41. Here, the input unit 21 may accept classification 47 directly, or it may accept information that can identify classification 47. More details will be described later.

[0046] (Step S42: Model control processing) The model control unit 22 inputs the classification 47 received in step S41, in addition to the query information 41, into the response generation model 31. The response generation model 31 then generates a draft response 43 and a justification statement 44 based on the classification 47. The model control unit 22 then retrieves the draft response 43 and the justification statement 44 output by the response generation model 31.

[0047] Classification 47 is information that serves as a premise for generating response proposal 43. Classification 47 is an attribute of at least one of the following: the station, the user, or the inquiry information 41.

[0048] The attributes of a station include its size, location, and type. Size is determined by factors such as the number of passengers. Stations can be categorized, for example, into large, medium, and small stations. Furthermore, unmanned stations without station staff may also be included in the classification of station size. Location indicates whether the area is a residential area, a business district, or a tourist area. Type indicates whether it is a transfer station or not. In addition, attributes of the station, such as the number of exits, may be included. The attributes of the station can be identified from the remote counter terminal 52 to which the operator terminal 51 is connected. Therefore, the input unit 21 receives terminal identification information from the remote counter terminal 52 to which the operator terminal 51 is connected. The input unit 21 then identifies the attributes of the station from the terminal identification information.

[0049] For example, when giving directions, the route you should take may differ depending on whether you are in a business district or a tourist area. In a business district, you might simply guide someone along the shortest route, while in a tourist area, you might guide them along a route that passes by shops or points of interest.

[0050] User attributes include walking ability and purpose of use. Walking ability is divided into categories such as normal walking and wheelchair use. Purpose of use is divided into categories such as tourists and business people. User attributes are entered by the user using the remote service terminal 52. Therefore, the input unit 21 receives user attribute input from the remote service terminal 52. Alternatively, the input unit 21 may acquire image data obtained by photographing the user with a camera installed in the remote service terminal 52 and identify user attributes from the image data. Furthermore, the input unit 21 may identify user attributes from the results of identifying the voice input from the user using the remote service terminal 52.

[0051] For example, when providing directions, the route to be guided will differ depending on the user's walking condition. Therefore, by generating a response plan 43 based on the user's walking condition, an appropriate response plan 43 is generated.

[0052] The attributes of inquiry information 41 include inquiries specific to the service counter, inquiries specific to a station, inquiries common to the entire railway line, and general inquiries. The attributes of inquiry information 41 are entered by the user using the remote service terminal 52. For example, before connecting to an operator, the user is asked to select the classification of the inquiry, thereby inputting the attributes of inquiry information 41. The input unit 21 then receives the input of attributes for inquiry information 41 from the remote service terminal 52.

[0053] For example, if an inquiry is common to all stations along the railway line, even if an appropriate response example cannot be obtained at the original contact point, if a response example for a similar inquiry can be obtained at another station along the line, it becomes possible to generate a response proposal 43 based on that response example.

[0054] ***Effects of Embodiment 4*** As described above, the response support device 10 according to Embodiment 4 inputs the classification 47, which is the premise for generating the response proposal 43, into the response generation model 31. Then, the response generation model 31 generates the response proposal 43 based on the classification 47. This makes it possible to obtain a more appropriate response proposal 43 and justification statement 44. As a result, the operator can provide an appropriate response.

[0055] Embodiment 5. Embodiment 5 differs from Embodiments 1 to 4 in that it generates the response proposal 43 by separating the parts for which it is difficult to generate the response proposal 43. Embodiment 5 explains this difference, and omits the explanation of the same points. Embodiment 5 describes a case in which a modification has been made to Embodiment 1. However, implementation It is also possible to modify forms 2 to 4.

[0056] ***Explanation of operation*** The operation of the response support device 10 according to Embodiment 5 will be described with reference to Figures 10 and 11. The processes in steps S51 and S53 are the same as the processes in steps S11 and S13 in Figure 2.

[0057] (Step S52: Model control processing) In addition to the query information 41, the model control unit 22 inputs a classification instruction 48 for the response proposal 43 to the response generation model 31, which corresponds to the difficulty of automatically generating the response proposal 43. The classification instruction 48 is a command that instructs the model to classify and generate the response proposal 43 according to the difficulty of automatically generating the response proposal 43. The response generation model 31 then determines the difficulty of automatic generation for each part of the proposed response 43. The response generation model 31 then generates the proposed response 43, categorized by difficulty. For example, the response generation model 31 generates the proposed response 43 by dividing it into parts with high difficulty and parts with low difficulty. The response generation model 31 also identifies the justification statement 44 that serves as the basis for the proposed response 43. The model control unit 22 then acquires the proposed response 43 and the justification statement 44 output by the response generation model 31.

[0058] This section explains how to determine the level of difficulty. The response generation model 31 can determine the difficulty level based on the likelihood that the information has been changed. Therefore, the model control unit 22 may be instructed to determine the difficulty level based on the likelihood that the information has been changed. For example, suppose an inquiry is made regarding the route to the destination station. In this case, suppose the response generation model 31 uses the response history 46 to generate a proposed response 43 that includes the departure time of the train to be boarded, along with transfer stations, etc. However, if the train schedule has been changed due to the effects of a typhoon or the like, it is highly likely that the train departure time has also been changed. In this case, the response generation model 31 generates a response generation model 31 that separates the transfer stations, etc. from the departure time of the train to be boarded. Segmentation means, for example, highlighting and drawing attention to parts of the information that are likely to have changed. Alternatively, the updated information for parts of the information that are likely to have changed may be included separately in response proposal 43. For example, if there has been a change in the train schedule and the updated temporary schedule is available, the information for the temporary schedule may be included in response proposal 43.

[0059] Furthermore, if there are changes to the equipment or the operating schedule, the response generation model 31 may learn that a change has occurred. This makes it easier for the response generation model 31 to determine whether or not there is a high probability that the information has changed. Also, as described in Embodiment 3, there are cases where a proposed response 43 is generated using the response history 46. In this case, for example, it becomes easier to determine that the part of the proposed response 43 that uses the response history 46 from before the operating schedule was changed is likely to have changed information.

[0060] Furthermore, the response generation model 31 may be trained using training data that explicitly labels parts that are not universal and may change. This makes it easier for the response generation model 31 to determine whether or not the information is likely to have changed.

[0061] Furthermore, the response generation model 31 may determine that the difficulty level is high not only based on the likelihood that the information has changed, but also if sufficient information could not be obtained.

[0062] ***Effects of Embodiment 5*** As described above, the response support device 10 according to Embodiment 5 generates the response proposal 43 by separating the parts that are difficult to generate. This prompts the operator to confirm the content of the response proposal 43, preventing the user from receiving an incorrect response.

[0063] ***Other configurations*** <Example 1> In Embodiment 5, the response generation model 31 determined the difficulty of generating the proposed response 43. However, a separate functional component for determining the difficulty of generating the proposed response 43 may be provided.

[0064] Referring to Figure 12, the configuration of the response support device 10 according to the modified example 1 will be described. The response support device 10 differs from the response support device 10 shown in Figure 1 in that it includes a section control unit 24 as a functional component. The function of the section control unit 24 is implemented by software, just like the other functional components.

[0065] The operation of the response support device 10 according to the modified example 1 will be explained with reference to Figures 13 and 14. The process in step S61 is the same as the process in step S51 in Figure 10.

[0066] (Step S62: Model control processing) The model control unit 22 inputs the inquiry information 41 to the response generation model 31, similar to the first embodiment. The model control unit 22 then retrieves the proposed response 43 and the justification statement 44 output by the response generation model 31. In other words, the response generation model 31 generates a proposed response 43 that is not categorized by difficulty level.

[0067] (Step S63: Classification control processing) The classification control unit 24 classifies the proposed response 43 obtained in step S62 according to the difficulty of automatically generating the proposed response 43. The method for determining the difficulty is the same as when the response generation model 31 makes the determination, and is based on factors such as the likelihood that the information has been changed.

[0068] (Step S64: Output processing) The output unit 23 outputs the basis description 44 obtained in step S62 and the response proposal 43 categorized in step S63.

[0069] Embodiment 6. Embodiment 6 differs from Embodiment 3 in that it uses the response history 46 to train the response generation model 31. Embodiment 6 explains this difference, while omitting explanations of the same points.

[0070] ***Explanation of the structure*** Referring to Figure 15, the configuration of the response support device 10 according to Embodiment 6 will be described. The response support device 10 differs from the response support device 10 shown in Figure 1 in that it includes a learning unit 25 as a functional component. The function of the learning unit 25 is implemented by software, similar to other functional components.

[0071] ***Explanation of operation*** The operation of the response support device 10 according to Embodiment 6 will be described with reference to Figures 16 and 17. The processes from step S71 to step S73 are the same as the processes from step S31 to step S33 in Figure 6.

[0072] (Step S74: Learning process) The learning unit 25 uses the combination of the inquiry information 41 input in step S71 and the response content provided by the operator to the user in response to the proposed response 43 output in step S73 as training data to train the response generation model 31. In other words, the learning unit 25 uses the newly obtained response history 46 as training data to train the response generation model 31. The content of the operator's response to the user is obtained from the conversation between the operator and the user. In this case, if the station operations manual 42 that serves as the basis for the operator's response is identified, the learning unit 25 may also include the relevant manual in the learning data.

[0073] In this example, the response generation model 31 is trained each time an inquiry is received. However, it is also possible to accumulate pairs of inquiry information 41 and response content for a certain period of time as training data and train the response generation model 31 all at once.

[0074] ***Effects of Embodiment 6*** As described above, the response support device 10 according to Embodiment 6 uses the content of the operator's response to the inquiry to train the response generation model 31. As a result, the more responses the operator provides, the better the response generation model 31 can output appropriate response proposals 43 and justification statements 44.

[0075] Embodiment 7. Embodiment 7 differs from Embodiments 1 to 6 in that it selects queries to be handled preferentially. Embodiment 7 explains this difference, while omitting explanations of the same points. Embodiment 7 describes a case in which a modification has been made to Embodiment 1. However, it is also possible to make modifications to Embodiments 2 to 6.

[0076] Multiple inquiries may be received from several stations. Since the number of operators is limited, users may have to wait for an operator to be assigned. In this case, it is necessary to appropriately distribute inquiries to operators. Embodiment 7 describes a method for selecting inquiries that require priority handling and appropriately distributing them to operators.

[0077] ***Explanation of the structure*** Referring to Figure 18, the configuration of the response support device 10 according to Embodiment 7 will be described. The response support device 10 differs from the response support device 10 shown in Figure 1 in that it includes a query distribution unit 26 as a functional component. The function of the query distribution unit 26 is implemented by software, similar to other functional components.

[0078] ***Explanation of operation*** Referring to Figure 19, the operation of the response support device 10 according to Embodiment 7 will be explained. The process shown in Figure 19 is executed as a preliminary step to the process shown in Figure 2.

[0079] (Step S81: Selection information reception processing) The input unit 21 accepts selection information. For each currently occurring inquiry, the selection information includes the language used for the inquiry, the category of the inquiry content, and the response waiting time. The selection information also includes the idle time of each operator. The language used for inquiries and the category of the inquiry content are entered by the user before the inquiry is made via a remote inquiry terminal 52 or similar device installed at the station's ticket window. The language used for inquiries and the category of the inquiry content may also be determined from the recognition results obtained by speech recognition of the user's inquiry. The waiting time for a response is counted by a timer and entered as the elapsed time since the inquiry was made. Idle time is counted and entered using a timer, representing the elapsed time since the operator finished handling an inquiry.

[0080] (Step S82: Urgency determination process) The inquiry distribution unit 26 identifies the urgency of each currently occurring inquiry based on the category of the inquiry content included in the selection information received in step S81. Here, it is assumed that the urgency level is predetermined for each category.

[0081] (Step S83: Priority determination process) The inquiry distribution unit 26 determines the priority of each currently occurring inquiry based on the urgency identified in step S82 and the response waiting time included in the selection information. The priority indicates the degree to which the operator will prioritize handling the inquiry. For example, the inquiry distribution unit 26 converts both the urgency and the response waiting time into evaluation values. Here, the higher the urgency, the higher the evaluation value. Also, the longer the response waiting time, the higher the evaluation value. For each inquiry, the inquiry distribution unit 26 determines the priority such that the higher the sum of the evaluation value calculated from the urgency and the evaluation value calculated from the response waiting time, the higher the priority.

[0082] (Step S84: Operator identification process) The inquiry distribution unit 26 identifies an operator capable of handling each currently occurring inquiry based on the language used in the inquiry and the category of the inquiry content included in the selection information. Here, it is assumed that operators capable of handling each language used in the inquiry are predetermined, and operators capable of handling each category of inquiry content are predetermined. The inquiry distribution unit 26 identifies operators capable of handling both the language used in the inquiry and the category of inquiry content for each currently occurring inquiry. For example, for each station employee, the employee's name and ID, work experience information, and skill information are stored in storage 13 as station employee information. Work experience information indicates categories that the station employee has handled in the past. For example, work experience information may indicate that the employee has experience handling inquiries related to transportation IC cards and tourism. Skill information indicates languages ​​that the employee can handle. The inquiry distribution unit 26 identifies operators who can handle each category of inquiry by referring to the work experience information. The inquiry distribution unit 26 identifies operators who can handle the language used in the inquiry by referring to the skill information.

[0083] (Step S85: Allocation process) The query distribution unit 26 sorts the operators in descending order of idle time. The query distribution unit 26 then assigns each operator to the target operator in the sorted order. The query distribution unit 26 then assigns to the target operator the query with the highest priority determined in step S83, from among the queries identified in step S84 as being handled by the target operator.

[0084] (Step S86: Verification process) The inquiry distribution unit 26 notifies the operator of the inquiry that was distributed in step S85 and prompts them to begin responding to the inquiry. For example, as shown in Figure 20, each operator's operator terminal 51 displays a list of currently occurring inquiries. The inquiry distribution unit 26 notifies an operator of the assigned inquiry by displaying a confirmation button in the column of the inquiry assigned to that operator on the list screen displayed on the operator terminal 51 of that operator. When the operator operates the operator terminal 51 and presses the confirmation button, the response to the inquiry begins.

[0085] In other words, the inquiry distribution unit 26 identifies each inquiry as a target inquiry, identifies operators who can handle the target inquiry, and selects an operator from among the operators who can handle the target inquiry to handle it. During this selection, the inquiry distribution unit 26 identifies operators with longer idle times as target operators, and distributes inquiries with high priority that the target operators can handle to the target operators. Then, the inquiry distribution unit 26 displays a confirmation button in the column of the inquiry that has been distributed to the target operator on the list of inquiries displayed on the operator terminal 51 used by the target operator.

[0086] In this instance, the inquiry distribution unit 26 displayed a confirmation button in the column for the inquiry with the highest priority. However, the inquiry distribution unit 26 may also display confirmation buttons in the columns for one or more inquiries with a priority higher than a pre-set criterion, allowing the operator to select one of these inquiries as the one to be handled.

[0087] ***Effects of Embodiment 7*** As described above, the response support device 10 according to Embodiment 7 selects inquiries that require priority handling and appropriately distributes them to operators. This makes it possible to improve user satisfaction with inquiry handling.

[0088] ***Other configurations*** <Modification 2> In Embodiment 7, in step S85, the inquiry distribution unit 26 distributed the inquiry only to operators identified as available in step S84. However, if there are no available operators, the inquiry distribution unit 26 may distribute the inquiry to operators who were not identified as available.

[0089] In Embodiment 7, the operators capable of handling each category of inquiry are predetermined. In this case, it is conceivable that operators can be designated as capable of handling categories for which they have previously handled a certain amount of experience. As shown in Modification 2, it is also possible to gradually increase the number of categories that operators can handle by sorting the inquiries.

[0090] <Variation 3> For some inquiries, a fixed response 32 may be prepared, allowing for automatic responses without operator intervention. In this case, as shown in Figure 21, the response support device 10 includes a response determination unit 27 as a functional component. The response determination unit 27 recognizes the content of the inquiry information 41 received by the input unit 21. The method for recognizing the content of the inquiry information 41 can be any method, such as using AI. The response determination unit 27 determines whether the recognized content of the inquiry information 41 is an inquiry for which a fixed answer 32 has been prepared. If the content of the inquiry information 41 is an inquiry for which a fixed answer 32 has been prepared, the response determination unit 27 outputs the fixed answer 32 corresponding to the content of the inquiry information 41 to the remote service terminal 52, etc. On the other hand, if the content of the inquiry information 41 is not an inquiry for which a fixed answer 32 has been prepared, the response determination unit 27 causes an operator to handle the inquiry. Specifically, the response determination unit 27 causes the inquiry distribution unit 26 to assign an operator.

[0091] The response determination unit 27 may input the inquiry information 41 to the model control unit 22 and obtain a response proposal 43 if the content of the inquiry information 41 is not an inquiry for which a fixed response 32 has been prepared. If the confidence level of the response proposal 43 is higher than the standard value, the response determination unit 27 outputs the response proposal 43 to the remote service terminal 52 or the like. On the other hand, if the confidence level of the response proposal 43 is below the standard value, the response determination unit 27 assigns an operator to the inquiry distribution unit 26.

[0092] The response determination unit 27 automatically tells the operator the content of the inquiry for which no fixed answer 32 exists. The system may also prompt the user to input the result of whether or not a response was possible. If the response determination unit 27 receives input indicating that an automatic response was possible, it may add the operator's response to the storage 13 as a fixed answer 32 for the inquiry content of the inquiry information 41. This allows the fixed answers 32 to be gradually enriched.

[0093] The response generation model 31 may also allow the operator to input an evaluation of the proposed response 43. By learning from the proposed response 43 and the operator's evaluation, the response generation model 31 can improve the accuracy of calculating the confidence level of the proposed response 43. This makes it possible to improve the accuracy of the response determination unit 27's decision on whether to automatically respond with the proposed response 43 or to have the operator respond.

[0094] <Modification 4> A separate terminal may be provided for each purpose of use. In this case, the inquiry distribution unit 26 determines the priority based on the terminal from which the inquiry originated. For example, a terminal could be provided that can only be used for categories requiring urgent inquiries. In this case, the inquiry distribution unit 26 would not perform an urgency determination based on the category, but would treat inquiries from this terminal as highly urgent. The inquiry distribution unit 26 would then determine the priority so that it gives the highest priority. This makes it possible to determine priority simply and appropriately. In addition, by providing a terminal that can only be used for categories requiring urgent inquiries, it is possible to prevent situations where inquiries are delayed because a terminal is unavailable, even though an inquiry is urgent.

[0095] <Modification 5> In Embodiment 7, the longer the response waiting time, the higher the evaluation value and the higher the priority. Inquiries from terminals where a long response waiting time is expected may be given a higher priority. Specifically, the inquiry distribution unit 26 calculates the priority as described in Embodiment 7, and then adjusts the priority of inquiries from terminals where a long response waiting time is expected to be increased.

[0096] In this process, the inquiry distribution unit 26 predicts the response waiting time for each terminal. Specifically, the inquiry distribution unit 26 predicts the response waiting time for each terminal based on the length of the queue or the number of people waiting. For example, the inquiry distribution unit 26 may define a unit waiting time for each unit length of the queue and predict the response waiting time from the queue length and the unit waiting time. Alternatively, the inquiry distribution unit 26 may define a unit waiting time per person and predict the response waiting time from the number of people waiting and the unit waiting time. Here, the input unit 21 receives video from cameras around the terminal. The inquiry distribution unit 26 identifies the length of the queue or the number of people waiting for processing at the terminal from the video.

[0097] The inquiry distribution unit 26 may predict the response waiting time based on the attributes of the people waiting, in addition to the length of the queue or the number of people waiting. For example, if the people waiting include someone who requires assistance, the inquiry distribution unit 26 may predict the response waiting time by assuming that it will take longer to assist that person than others. Furthermore, if the category or content of the inquiry of the person waiting can be identified, the inquiry distribution unit 26 may predict the response waiting time by taking into account the category or content of the inquiry of the person waiting. For example, a weight may be assigned to each inquiry category or content, and the inquiry distribution unit 26 may predict the response waiting time by multiplying the unit waiting time by the weight.

[0098] Furthermore, when predicting response time based solely on queue length, the queue length is used to... This can be described as determining priority. For example, if the queue is less than 2 meters long, the priority may not be adjusted; if the queue is between 2 meters and 5 meters long, the priority may be increased by one level, and so on. Similarly, when predicting response time based solely on the number of people waiting, it can be said that priority is determined by the number of people waiting. For example, one might consider not adjusting the priority if there are fewer than two people waiting, or increasing the priority by one level if there are two or more but fewer than five people waiting, and so on.

[0099] <Variation 6> The priority of an inquiry may be increased if it is from a user who has a ticket that allows them to receive priority support. Specifically, the inquiry distribution unit 26 calculates the priority as described in Embodiment 7 and then adjusts it to increase the priority if the inquiry is from a user who has a ticket. For example, a ticket system could be offered for a fee that guarantees priority handling of inquiries, ensuring that inquiries from users who purchase such tickets are processed preferentially. The ticket would be recorded on an IC card, and when the IC card with the ticket recorded on it is held over a terminal, the inquiry distribution unit 26 would adjust the priority accordingly. This would allow users in urgent situations to receive priority handling of their inquiries by purchasing a ticket.

[0100] Embodiment 8. Embodiment 8 differs from Embodiments 1 to 7 in that it estimates the emotions of at least one of the users or the operator. Embodiment 8 explains this difference, while omitting explanations of the same points. Embodiment 8 describes a case in which a modification has been made to Embodiment 1. However, it is also possible to modify Embodiments 2 to 7.

[0101] ***Explanation of the structure*** Referring to Figure 22, the configuration of the response support device 10 according to Embodiment 8 will be described. The response support device 10 differs from the response support device 10 shown in Figure 1 in that it includes an emotion estimation unit 28 as a functional component. The function of the emotion estimation unit 28 is implemented by software, similar to other functional components.

[0102] ***Explanation of operation*** Referring to Figure 23, the operation of the response support device 10 according to Embodiment 8 will be described. The process shown in Figure 23 is executed as a subsequent step to the process shown in Figure 2. The process shown in Figure 23 may be executed after a certain period of time, such as one day, has elapsed, targeting all of the processes shown in Figure 2 that were executed during that period. For example, the process shown in Figure 23 may be executed when the operator finishes their work for the day, targeting each of the processes shown in Figure 2 that were executed by that operator on that day.

[0103] (Step S91: Conversation information reception processing) The input unit 21 receives input of conversation information between the user and the operator. The conversation information includes the conversation between the user and the operator, which is the inquiry information 41 received in step S11 of Figure 2, and the conversation consisting of the response made by the operator in step S13 of Figure 2 by referring to the response proposal 43, etc., and the user's response to that response. The input unit 21 converts conversation information into text when the user and the operator are conversing by voice.

[0104] (Step S92: User emotion determination process) The emotion estimation unit 28 determines from the conversation information received in step S91 whether the user's emotions are in a positive or negative state. Specifically, the emotion estimation unit 28 determines whether the user's emotions are positive or negative based on the content of what the user says in the conversation, and, if the conversation is conducted verbally, the tone and speed of the user's voice. In this case, the emotion estimation unit 28 may assign a score to the user's emotions instead of determining whether they are positive or negative. For example, the emotion estimation unit 28 may assign a score such that a higher score indicates a more positive emotion, and a lower score indicates a more negative emotion. In this case, the emotion estimation unit 28 may control the score by referring to modifiers in the words spoken by the user. For example, a higher score may be given when the modifier "very" is attached to a word of gratitude. As a specific example, the word "Thank you very much, that was very helpful" may receive a higher score than the word "Thank you very much, that was very helpful."

[0105] (Step S93: Evaluation output processing) The emotion estimation unit 28 outputs the user's emotion determined in step S92. Specifically, the emotion estimation unit 28 outputs to the operator terminal 51 that the user's emotion was positive if the emotion is positive. The emotion estimation unit 28 may also output the score if the user's emotion has been assigned a score. The emotion estimation unit 28 may convert the score into a star rating representing the user's satisfaction level and output and display it on the operator terminal 51. In this case, the emotion estimation unit 28 may also output and display any words of gratitude expressed by the user on the operator terminal 51. Furthermore, if the emotion estimation unit 28 detects that the user's emotion is negative, it outputs to the storage 13 that the emotion was negative, along with the operator's identification information and the category of the inquiry. The emotion estimation unit 28 may also output the score if the user's emotion has been assigned a score. The information output to the storage 13 can be reviewed by the operator's manager at any time and used for operator training, etc. In other words, categories that each operator struggles to handle can be identified, allowing for guidance on appropriate response methods.

[0106] (Step S94: Operator emotion determination process) The emotion estimation unit 28 determines from the conversation information received in step S91 whether the operator's emotions are in a positive or negative state. Specifically, the emotion estimation unit 28 determines whether the operator's emotions are positive or negative based on the content of the conversation and, if the conversation is conducted verbally, the tone and speed of the operator's voice. In this case, the emotion estimation unit 28 may assign a score to the operator's emotions instead of determining whether they are positive or negative. For example, the emotion estimation unit 28 may assign a score such that a higher score indicates a more positive emotion, and a lower score indicates a more negative emotion.

[0107] (Step S95: Stress output processing) The emotion estimation unit 28 quantifies the operator's stress level and degree of attachment to their work based on the operator's emotions determined in step S94, and outputs this information, along with the operator's identification information, to the storage unit 13. The information output to the storage unit 13 can be reviewed by the operator's manager at any time and used for operator care or staffing purposes.

[0108] ***Effects of Embodiment 8*** As described above, the response support device 10 according to Embodiment 8 notifies the operator of the user's emotions when the user's emotions are positive. For example, the response support device 10 is When a user finishes their work for the day, the operator is notified of the user's positive emotions. This allows the operator to end their workday in a good mood and increase their motivation for the job.

[0109] Furthermore, the response support device 10 according to Embodiment 8 outputs the user's emotions to the storage 13 along with the operator's identification information and the inquiry category if the user's emotions are negative. This allows for the identification of categories that each operator is not good at handling, making it possible to provide operators with guidance on appropriate response methods.

[0110] Furthermore, the response support device 10 according to Embodiment 8 quantifies the operator's emotions, stress levels, and degree of attachment to their work, and outputs these to the storage device 13. This makes it possible to appropriately provide care for operators and manage staffing levels.

[0111] ***Other configurations***

[0112] <Example 7> The emotion estimation unit 28 identifies operators whose estimated positive emotions are more prevalent than the standard for the number of users they have interacted with. The emotion estimation unit 28 then raises the evaluation of the identified operators; in other words, the operators' performance ratings are increased. Furthermore, the emotion estimation unit 28 extracts conversation information from the identified operators where the user's emotion was estimated to be positive, as exemplary response information. The emotion estimation unit 28 makes the extracted exemplary response information available to other operators for reference. For example, the emotion estimation unit 28 stores the exemplary response information so that other operators can refer to it as needed.

[0113] The emotion estimation unit 28 identifies operators whose estimated negative emotions are more prevalent than the evaluation criteria, relative to the number of users they have interacted with. The emotion estimation unit 28 then lowers the evaluation of the identified operators. In other words, the operators' performance ratings are reduced.

[0114] <Differentiation Example 8> In cases where the same user repeatedly makes inquiries, and it is suspected that the operator's mood is negative each time, it may be advisable to register this user and route their inquiries to a specific operator, such as a veteran operator. Specifically, the emotion estimation unit 28 identifies users whose estimated operator emotions are negative more often than a baseline as specific users. The emotion estimation unit 28 then registers these specific users in the storage 13. For example, if a specific user makes an inquiry by holding an IC card over a reader, the emotion estimation unit 28 registers the identification information contained in the IC card as the identification information of that specific user. When an inquiry is made by a user whose identification information is registered in the storage 13, the inquiry distribution unit 26 determines that it is an inquiry from a specific user. The inquiry distribution unit 26 then distributes the inquiry from the specific user to a specific operator. This will enable appropriate responses even to difficult users. As a result, unpleasant experiences for both users and operators can be avoided. Avoiding unpleasant experiences for users will improve user satisfaction with the railway company and lead to higher repeat usage rates. Also, avoiding unpleasant experiences for operators will increase their job satisfaction and lead to lower turnover rates.

[0115] <Modification 9> In Embodiment 6, the response history 46 was used as training data. In addition to this, the learning unit 25 may also include the user's emotions in response to the operator's response as training data. The learning unit 25 can learn the response as correct data if the emotion is positive. On the other hand, the learning unit 25 can learn the response as incorrect data if the emotion is negative.

[0116] The contents described in Embodiments 4 and 5 may be combined. Specifically, the contents described in Embodiment 4 may be combined to include classification 47 in the training data. In other words, the learning unit 25 may train the response generation model 31 using the inquiry information 41, classification 47, the response content, and the user's reaction as training data. Alternatively, the contents described in Embodiment 5 may be combined to train the response generation model 31 using the inquiry information 41, classification instruction 48, the response content, and the user's reaction as training data.

[0117] Embodiment 9. Embodiment 9 differs from Embodiments 1 to 8 in that it evaluates equipment such as a remote counter terminal 52 provided by a railway operator and used by users. Embodiment 9 explains this difference, while omitting explanations of the same points. Embodiment 9 describes a case in which a modification has been made to Embodiment 1. However, it is also possible to make modifications to Embodiments 2 to 8.

[0118] ***Explanation of the structure*** Referring to Figure 24, the configuration of the response support device 10 according to Embodiment 9 will be described. The response support device 10 differs from the response support device 10 shown in Figure 1 in that it includes an equipment evaluation unit 29 as a functional component. The function of the equipment evaluation unit 29 is implemented by software, similar to other functional components.

[0119] ***Explanation of operation*** The equipment evaluation unit 29 records the usage status of equipment such as the remote service terminal 52 from which the inquiry information 41 was input. The usage status includes user attributes, number of uses, usage time, etc. Furthermore, the device evaluation unit 29 evaluates the user's satisfaction with the device they used. The device evaluation unit 29 determines whether the user's evaluation was positive or negative and uses this to determine the user's satisfaction level. Specifically, the device evaluation unit 29 uses the user's voice or camera image to determine whether the user's evaluation was positive or negative. In this case, the device evaluation unit 29 may record keywords included in the user's statements related to the evaluation. Alternatively, the user may be asked to input their evaluation of the device after using it. The equipment evaluation unit 29, in accordance with a request from the administrator of the response support device 10, outputs and displays the usage status and satisfaction level with the equipment on the display device.

[0120] Here, "user voice" refers to the user's voice included in conversation information between a user using the device being evaluated and an operator responding to the user's inquiry. "Camera image" refers to an image of the user during a conversation between the user using the device being evaluated and an operator responding to the user's inquiry.

[0121] The equipment evaluation unit 29 may also output and display keywords included in user statements related to the evaluation. The equipment evaluation unit 29 may display information indicating the equipment to be evaluated and keywords included in user statements related to the evaluation in a manner visible to other users. For example, the equipment evaluation unit 29 may display information indicating the equipment and keywords on a display device installed in a location such as a station platform or passageway. When the equipment evaluation unit 29 displays information on a display device visible while riding an escalator, it may also display information in accordance with the movement of the escalator. The display could also be made to move to make it easier for users to see. Sharing positive user feedback with other users can serve as advertising for the device. Conversely, sharing negative user feedback with other users can let them know that improvements to the device are expected. The device evaluation unit 29 may identify the attributes of the person viewing the display device from images obtained by a camera installed near the display device. The device evaluation unit 29 may then switch the display according to the identified attributes. For example, if the attribute is that of a child, the device evaluation unit 29 may switch to displaying more hiragana characters to make it easier for children to read. If the attribute is that of an elderly person, the device may also display a description of the device and instructions for use to make it easier for them to understand. The device evaluation unit 29 may also provide output in voice or sign language for people who are visually or hearing impaired.

[0122] ***Effects of Embodiment 9*** As described above, the response support device 10 according to Embodiment 9 records usage status and evaluates satisfaction levels, and displays the results as requested. This allows for comparison and display of usage status and satisfaction levels of different terminals in different locations, thereby supporting railway companies in considering which equipment to introduce and which improvements to make. For example, it can support considerations such as insufficient measures for vulnerable road users, poor placement of equipment, or lack of information such as equipment manuals.

[0123] ***Other configurations*** <Variation 10> In Embodiment 9, the evaluation was conducted on equipment provided by the railway operator and used by users. However, the evaluation may also be conducted on services provided by the railway operator and used by users. Examples of such services include free rides for children and book lending services. When evaluating a service, it is possible to identify which service the conversation is about based on the content of the conversation between the user and the operator. For example, the equipment evaluation unit 29 inputs the conversation information between the user and the operator into an evaluation model, which is a learning model, and has the evaluation model identify which service it is about.

[0124] <Variation 11> The device evaluation unit 29 may input the user's statements and camera images into an evaluation model, which is a learning model, and obtain the user's satisfaction level output by the evaluation model. In this case, in addition to evaluating the user's satisfaction level, the device evaluation unit 29 may also prompt the evaluation model to suggest improvement plans to improve the satisfaction level, and obtain the user's satisfaction level and improvement plans output by the evaluation model. Improvement plans are ways to make things better if the evaluation is positive, and ways to address the points raised if the evaluation is negative. Furthermore, the equipment evaluation unit 29 may instruct the evaluation model to extract keywords related to the satisfaction evaluation of statements that are more satisfying than the first criterion. Similarly, the equipment evaluation unit 29 may instruct the evaluation model to extract keywords related to the satisfaction evaluation of statements that are less satisfying than the second criterion, which is set lower than the first criterion. The equipment evaluation unit 29 may then obtain keywords for statements that are more satisfying than the first criterion and keywords for statements that are less satisfying than the second criterion. Furthermore, the evaluation model may include instructions for the equipment evaluation unit 29 to output keywords in bullet points for high evaluations and keywords in bullet points for low evaluations. Alternatively, the evaluation model may include instructions for the equipment evaluation unit 29 to output a summary of the ratio of high to low evaluations and the reasons for each evaluation.

[0125] <Variation 12> In Embodiment 9, the evaluation focused on the device itself. However, the device evaluation unit 29 may also evaluate the user's satisfaction with the automatically outputted response, such as the automated response using the fixed answer 32 described in Modification 3. For example, the device evaluation unit 29 may determine the user's satisfaction with the automated response from the user's voice or camera image after the response. In other words, when a chatbot provides a response, the device evaluation unit 29 may determine the user's satisfaction with this response.

[0126] <Example 13> In the embodiments described above, each functional component was implemented in software. However, in Modification 6, each functional component may be implemented in hardware. The differences between this Modification 6 and the above embodiments will be explained below.

[0127] When each functional component is implemented in hardware, the response support device 10 includes electronic circuits instead of the processor 11, memory 12, and storage 13. The electronic circuits are dedicated circuits that implement the functions of each functional component, as well as the functions of the memory 12 and storage 13.

[0128] Electronic circuits can include single circuits, complex circuits, programmed processors, parallel programmed processors, logic ICs, GAs, ASICs, and FPGAs. GA stands for Gate Array. ASIC stands for Application Specific Integrated Circuit. FPGA stands for Field-Programmable Gate Array. Each functional component may be implemented in a single electronic circuit, or it may be implemented by distributing each functional component across multiple electronic circuits.

[0129] <Example 14> As an alternative modification (7), some of the functional components may be implemented in hardware, while others may be implemented in software.

[0130] The processor 11, memory 12, storage 13, and electronic circuitry are collectively referred to as the processing circuit. In other words, the function of each functional component is realized by the processing circuit.

[0131] Furthermore, the term "part" in the above explanation may be replaced with "circuit," "process," "procedure," "processing," or "processing circuit."

[0132] The various aspects of this disclosure are summarized below as an appendix. (Note 1) A model control unit inputs customer inquiry information at the station ticket window into a response generation model, which is a learning model given the station operations manual, and acquires the response proposal output by the response generation model. An output unit that outputs the proposed response obtained by the model control unit. A response support device equipped with the following features. (Note 2) The model control unit further obtains the basis for the proposed response in the station operations manual, The output unit further outputs the basis description. The response support device described in Appendix 1. (Note 3) The model control unit further inputs window information indicating the location of the window to the response generation model. A response support device as described in Appendix 1 or 2. (Note 4) The model control unit further inputs train operation information to the response generation model. A response support device as described in any one of the items 1 to 3 in the appendix. (Note 5) The model control unit further inputs train fare information to the response generation model. A response support device as described in any one of the items 1 to 4 in the appendix. (Note 6) The model control unit further inputs device information managed by devices installed at the station that are available to the user into the response generation model. A response support device as described in any one of the items 1 to 5 in the appendix. (Note 7) The model control unit further inputs the user's personal information into the response generation model, as described in any one of the items 1 to 6. (Note 8) The model control unit further inputs ticket information about the ticket held by the user into the response generation model. A response support device as described in any one of the items 1 through 7 of the appendix. (Note 9) The model control unit further inputs information about the lost item that the user has lost into the response generation model. A response support device as described in any one of the items 1 through 8 of the appendix. (Note 10) The model control unit further inputs map information of the area around the station to the response generation model. A response support device as described in any one of the items 1 through 9 of the appendix. (Note 11) The model control unit further inputs event information about events taking place around the station to the response generation model. A response support device as described in any one of the items 1 to 10 in the appendix. (Note 12) The response generation model is given a response history, which is a pair of past query information and the response content to that query information. A response support device as described in any one of the items 1 through 11 of the appendix. (Note 13) The model control unit inputs a classification based on at least one attribute of the station, the user, and the inquiry information into the response generation model, and obtains the response proposal generated by the response generation model based on the classification. A response support device as described in any one of the items 1 to 12 in the appendix. (Note 14) The model control unit inputs a classification instruction for the response proposals to the response generation model, corresponding to the difficulty of automatically generating the response proposals, and acquires the response proposals generated by the response generation model that have been classified according to the difficulty. A response support device as described in any one of the items 1 to 13 in the appendix. (Note 15) The response support device further, A classification control unit that classifies the proposed responses obtained by the model control unit according to the difficulty of automatically generating the proposed responses. Equipped with, The output unit outputs the response proposals that have been divided by the division control unit. A response support device as described in any one of the items 1 to 13 in the appendix. (Note 16) The response support device further, A learning unit that trains the response generation model using the response history, which is a pair of inquiry information and the response content in which the operator responds to the user based on the response proposal output by the output unit and the justification description, as training data. A response support device according to any one of the appendices 1 to 15, comprising: (Note 17) The response support device further, The inquiry distribution unit determines the priority of inquiries based on the category to which the content of the inquiry information belongs and the user's waiting time for response, and prioritizes high-priority inquiries for handling by operators. A response support device according to any one of the appendices 1 to 16, comprising: (Note 18) The response support device further, Emotion estimation unit estimates the user's emotions based on the user's response to the operator's response to the inquiry information. A response support device according to any one of the appendices 1 to 17, comprising: (Note 19) The response support device further, The evaluation subject is at least one of the equipment and services provided by the railway operator, and the input unit receives at least the user's statements from conversation information between the user using the equipment and the operator responding to inquiries from the user. A device evaluation unit determines the user's satisfaction level with the evaluation target used by the user based on the user's statements received by the input unit. A response support device as described in any one of the appendices 1 to 18, comprising: (Note 20) The computer inputs customer inquiry information from the station ticket window into a response generation model, which is a learning model given the station operations manual, and obtains the response proposal output by the response generation model. A response support method in which a computer outputs the aforementioned response proposal. (Note 21) A model control process that takes customer inquiry information at the station ticket window, inputs it into a response generation model which is a learning model given the station operations manual, and obtains the response proposal output by the response generation model. Output processing to output the proposed response obtained by the model control processing, A response support program that enables a computer to function as a response support device.

[0133] The embodiments and variations of this disclosure have been described above. Some of these embodiments and variations may be implemented in combination. Alternatively, some or all of them may be implemented in part. However, this disclosure is not limited to the embodiments and variations described above, and various modifications are possible as needed. [Explanation of symbols]

[0134] 10 Response support device, 11 Processor, 12 Memory, 13 Storage, 14 Communication interface, 21 Input unit, 22 Model control unit, 23 Output unit, 24 Classification control unit, 25 Learning unit, 26 Inquiry distribution unit, 27 Response determination unit, 28 Emotion estimation unit, 29 Equipment evaluation unit, 31 Response generation model, 41 Inquiry information, 42 Station operations manual, 43 Draft response, 44 Justification description, 45 Additional information, 46 Response history, 47 Classification, 48 Classification instruction, 51 Operator terminal, 52 Remote counter terminal.

Claims

1. Selection information for selecting which of multiple inquiries from station users will be prioritized for an operator, the selection information includes the category of the content of the target inquiry and the response waiting time, which is the elapsed time since the target inquiry was made. An inquiry distribution unit converts the urgency identified from the category included in the selection information for the target inquiry received by the input unit into an evaluation value, converts the response waiting time included in the selection information for the target inquiry into an evaluation value, determines a priority indicating the degree to which the operator should prioritize handling the target inquiry from the evaluation value converted from the urgency and the evaluation value converted from the response waiting time, and displays a confirmation button in the column of the inquiry with a priority higher than the standard on the inquiry list screen showing the columns of the multiple inquiries displayed on the operator terminal used by the operator, thereby notifying the operator of the inquiry with a priority higher than the standard. A response support device equipped with the following features.

2. Work experience information indicating the categories in which each operator has experience handling tasks is stored in the memory as station staff information. The inquiry distribution unit, for each of the multiple inquiries, refers to the station staff information to identify an operator from the multiple operators who has experience handling the category included in the selection information, and displays the confirmation button in the column of the inquiry list screen displayed on the operator terminal used by the target operator among the multiple operators, for inquiries in the category that the target operator has experience handling and whose priority is higher than the standard. The response support device according to claim 1.

3. The aforementioned selection information includes the language used for the inquiry. Skill information indicating the languages ​​that each operator can communicate is stored in the station staff's memory as station staff information. The inquiry distribution unit, for each of the multiple inquiries, refers to the station staff information to identify an operator from the multiple operators who can handle the language included in the selection information, and displays the confirmation button in the column of the inquiry list screen displayed on the operator terminal used by the target operator among the multiple operators, for inquiries in the language that the target operator can handle and whose priority is higher than the standard. The response support device according to claim 1.

4. The inquiry distribution unit sets the multiple operators to the target operator in order of longest idle time. The response support device according to claim 2.

5. When the aforementioned confirmation button is pressed, the system will begin responding to the inquiry corresponding to the field where the confirmation button is displayed. The response support device according to claim 1.

6. The aforementioned inquiry distribution unit is, The system determines whether the content of the aforementioned inquiry can be automatically responded to with a pre-prepared fixed answer, and if such an automatic response is not possible, it determines the priority, and displays a confirmation button in the column for inquiries with a priority higher than the standard. If, after the operator has responded to the inquiry, the operator has determined that an automated response is possible to the inquiry, the operator's response to the user is registered as the fixed response. The response support device according to claim 1.

7. Terminals are available for inquiries based on the purpose of use. The aforementioned inquiry distribution unit determines the priority based on the terminal from which the target inquiry originated. The response support device according to claim 1.

8. The inquiry distribution unit adjusts the priority of inquiries from terminals where a long response waiting time is expected to be increased. The response support device according to claim 1.

9. The inquiry distribution unit adjusts the priority of the inquiry to a higher level if the inquiry in question is from a user who has a ticket that allows for priority handling. The response support device according to claim 1.

10. The computer receives selection information for selecting which of several inquiries from station users should be prioritized for an operator, and the selection information includes the category of the content of the target inquiry and the response waiting time, which is the elapsed time since the target inquiry was made. A response support method comprising: a computer converting the urgency identified from the category included in the selection information for the target inquiry into an evaluation value; converting the response waiting time included in the selection information for the target inquiry into an evaluation value; determining a priority indicating the degree to which the operator should prioritize handling the target inquiry based on the evaluation value converted from the urgency and the evaluation value converted from the response waiting time; and notifying the operator of inquiries with a priority higher than the standard by displaying a confirmation button in the column of the inquiry with a priority higher than the standard on the inquiry list screen showing the columns of the multiple inquiries displayed on the operator terminal used by the operator.

11. Selection information for selecting which of the multiple inquiries from station users will be prioritized for the operator, and which includes the category of the content of the target inquiry and the response waiting time, which is the elapsed time since the target inquiry was made. The inquiry distribution process involves converting the urgency identified from the category included in the selection information for the target inquiry received through the input process into an evaluation value, converting the response waiting time included in the selection information for the target inquiry into an evaluation value, determining a priority indicating the degree to which the operator should prioritize handling the target inquiry based on the evaluation value converted from the urgency and the evaluation value converted from the response waiting time, and displaying a confirmation button in the column of the inquiry with a priority higher than the standard on the inquiry list screen showing the columns of the multiple inquiries displayed on the operator terminal used by the operator, thereby notifying the operator of the inquiry with a priority higher than the standard. A response support program that enables a computer to function as a response support device.

Images