Response control device and response control method

Abstract

This response control device comprises: a response method determination unit 14 that determines a scenario to be applied to a chatbot and a scheduled handling operator on the basis of the contents of an inquiry from a user, evaluation information about the response ability of an operator, and operating status information about the operator; a scenario response unit 11 that performs a chatbot process according to the scenario determined by the response method determination unit 14; and an operator connection unit 12 that performs a process to connect to the scheduled handling operator after the performing of the chatbot process. An appropriate combination of the scenario and the scheduled handling operator is identified in accordance not only with the operating status of the operator but also with the contents of the inquiry from the user, to execute an appropriate transition from the chatbot to the operator.

Description

Response control device and response control method

[0001] The present invention relates to a response control device and a response control method, and is particularly suitable for application to a response system in which, after a chatbot responds to a user's inquiry, an operator can continue to respond using the information collected by the chatbot.

[0002] In recent years, chatbots that converse with users using voice, text, or option presentations have been increasingly introduced in call center systems, online customer service systems in which operators provide customer service online (on the web) via the Internet, etc. For example, a system has been provided in which a chatbot first responds to simple inquiries from users, and then an operator can take over from the chatbot and respond as needed.

[0003] When the chatbot responds on behalf of an operator, it searches among multiple operators for a standby operator (an operator who is not currently responding to an inquiry) and requests that operator to take over the response. The requested operator takes over the information collected by the chatbot from the user and continues the chatbot's response.

[0004] In the past, most chatbots in this type of response system responded according to a pre-designed, fixed scenario. In this case, when the chatbot finished executing the fixed scenario and tried to hand over to an operator, the transition from the chatbot to the operator could not be carried out smoothly depending on the operator's standby status.

[0005] In response to this, a dialogue system is known that selectively uses scenarios of different lengths depending on the operating status of the operator (either available, responding, or absent) (see, for example, Patent Document 1). In the dialogue system described in Patent Document 1, the connection time required for the operator to be able to connect to the next user is identified depending on the operating status of the operator, and a question scenario is identified in which the time required to execute the question scenario falls within the connection time, and the chatbot asks one or more questions to the user using the identified question scenario.

[0006] According to the system described in Patent Document 1, if an operator is available to respond quickly, the depth of the preliminary questions posed by the chatbot can be reduced (or no preliminary questions can be asked), allowing the user to be connected to an operator quickly. On the other hand, if no operator is available to respond quickly, the time until connecting to an operator can be effectively utilized to ask more in-depth questions using the chatbot, thereby reducing the number of questions asked by the operator after connection. In this way, by determining the type of preliminary questions depending on the operator's operating status, the total user response time can be shortened and user satisfaction can be improved.

[0007] Patent No. 7097460

[0008] However, the dialogue system described in Patent Document 1 only changes the question scenario depending on the operating status of the operator, which causes problems such as the inability to hand over the inquiry to an appropriate operator depending on the content of the inquiry, and the inability to hand over the inquiry to an operator after collecting sufficient information using the chatbot.

[0009] The present invention has been made to solve such problems, and aims to enable appropriate handover from a chatbot to an operator based not only on the operator's operating status but also on the content of the user's inquiry.

[0010] In order to solve the above-mentioned problems, the present invention provides a response system in which a chatbot responds to a user inquiry, and then an operator can continue to respond using information collected by the chatbot.When a user makes an inquiry, the system determines a response method, including a scenario to be applied to the chatbot and an operator who is scheduled to respond after the chatbot responds, based on the content of the inquiry, evaluation information regarding the operator's response ability, and information about the operator's operating status.

[0011] According to the present invention configured as described above, the chatbot scenario and the operator scheduled to respond are selected based on the operating status and response ability of the operator, so that an appropriate combination of scenario and operator scheduled to respond can be identified based not only on the operating status of the operator but also on the content of the inquiry from the user, and the handover from the chatbot to the operator can be carried out appropriately.

[0012] 1 is a diagram showing an example of the overall configuration of a response system to which a response control device of this embodiment is applied; FIG. 2 is a block diagram showing an example of the functional configuration of a response control device according to this embodiment; FIG. 3 is a flowchart showing an example of the operation of a response control device according to this embodiment; FIG. 4 is a block diagram showing an example of the functional configuration of a response control device according to a first modified example; FIG. 5 is a flowchart showing an example of the operation of a response control device according to the first modified example; FIG. 6 is a block diagram showing an example of the functional configuration of a response control device according to a second modified example; and FIG. 7 is a flowchart showing an example of the operation of a response control device according to the second modified example.

[0013] An embodiment of the present invention will be described below with reference to the drawings. Fig. 1 is a diagram showing an example of the overall configuration of a response system to which a response control device of this embodiment is applied. The response system of this embodiment is a system in which a response control device 1 responds to an inquiry from a user sent from a user terminal 2 using a chatbot, and then provides information collected by the chatbot to an operator terminal 3, allowing an operator to continue responding using the collected information. The operator terminal 3 is a system in which a plurality of operator terminals 3 are used by a plurality of operators. -1 , 3 -2, ..., 3 -n Includes:

[0014] The response control device 1 is a server group consisting of one server or multiple servers connected to a user terminal 2 and an operator terminal 3 via a communication network 100 such as the Internet or a mobile phone network, and performs various processes required for the chatbot to respond to inquiries from users and for the operator to respond. Here, the response control device 1 first executes the chatbot's processing according to a scenario, and then executes the processing according to one of the operator terminals 3. -i (i is any one of 1 to n) and provides the information collected by the chatbot.

[0015] The user terminal 2 is configured, for example, by a personal computer, a tablet, a smartphone, etc. The user terminal 2 is connected to the response control device 1 via a communication network 100, which allows the user to interact with the chatbot using text or voice. The user terminal 2 also communicates with the operator terminal 3 via the response control device 1. -i This allows users to communicate with an operator using text, voice or video.

[0016] The operator terminal 3 is configured with, for example, a personal computer, a tablet, a smartphone, etc. The operator terminal 3 is connected to the response control device 1 via a communication network 100, and provides the response control device 1 with information that can identify the operating status of the operator. The operator terminal 3 is also connected to the user terminal 2 via the response control device 1, which allows the operator to communicate with the user using text, audio, or video. The communication network 100 that connects the response control device 1 and the operator terminal 3 may be a LAN (Local Area Network).

[0017] Fig. 2 is a block diagram showing an example of the functional configuration of the response control device 1 according to this embodiment. As shown in Fig. 2, the response control device 1 of this embodiment includes, as its functional configuration, a scenario response unit 11, an operator connection unit 12, an operating status management unit 13, and a response method determination unit 14. The response method determination unit 14 includes, as specific functional configurations, an operator search unit 14A and a scenario identification unit 14B. The response control device 1 of this embodiment also includes, as storage media, a scenario storage unit 21, a collected information storage unit 22, an evaluation information storage unit 23, and an operating status storage unit 24.

[0018] The functional blocks 11 to 14 can be configured using any of hardware, a DSP (Digital Signal Processor), and software. For example, when configured using software, the functional blocks 11 to 14 are actually configured with a computer's CPU, RAM, ROM, etc., and are realized by the operation of a program stored in a storage medium such as RAM, ROM, a hard disk, or a semiconductor memory. Note that the functional blocks 11 to 14 do not necessarily have to be provided on one server, and may be provided separately on multiple servers.

[0019] The scenario storage unit 21 stores multiple scenarios to be applied to the chatbot. A scenario defines the content of a question and answer session to be executed by the chatbot when a user makes an inquiry and before connecting the user to an operator, and includes one or more response sentences (including at least one of a greeting, a question, an answer, a statement, etc.).

[0020] Basically, a scenario is designed for the chatbot to collect information from a user before connecting to an operator. Response sentences included in a scenario can be broadly divided into deep-level responses for collecting information directly related to the content of the user's inquiry and superficial-level responses for collecting information not directly related to the content of the user's inquiry. Deep-level responses include, for example, responses for extracting the user's issues and responses for collecting information necessary for resolving the issues. On the other hand, superficial-level responses include, for example, responses for confirming the field to which the inquiry belongs and responses for collecting formal information such as the user's name, age, and desired response. A response sentence for confirming the field of the inquiry is, for example, a question containing multiple field options, and prompts the user to select one of the options.

[0021] The scenario storage unit 21 stores scenarios including deep response sentences (hereinafter referred to as deep scenarios), scenarios including response sentences for confirming the field of an inquiry (hereinafter referred to as field confirmation scenarios), and scenarios including response sentences for collecting other formal information (hereinafter referred to as surface scenarios). Multiple deep scenarios are stored for each field of inquiry. Each type of scenario is designed to collect one or more pieces of information from a user through one or more response sentences. A scenario may be designed to collect one piece of information per response sentence, or may be designed to collect multiple pieces of information per response sentence.

[0022] The scenario response unit 11 executes dialogue processing by the chatbot based on one of the scenarios identified from multiple scenarios stored in the scenario storage unit 21. When a user makes an inquiry, the scenario response unit 11 executes a field confirmation scenario to confirm the field to which the inquiry belongs. The scenario response unit 11 then executes either a deep scenario corresponding to the confirmed field or a superficial scenario unrelated to the field. The response method determination unit 14 determines which scenario to execute (details on this will be described later).

[0023] The scenario response unit 11 stores information collected from the user terminal 2 by the chatbot executing a scenario in the collection information storage unit 22. That is, the scenario response unit 11 provides one response sentence to the user terminal 2, and stores the information acquired from the user terminal 2 as a reply in the collection information storage unit 22. The scenario response unit 11 repeats the same process for one or more response sentences included in the scenario, thereby sequentially storing the information acquired from the user terminal 2 in the collection information storage unit 22.

[0024] After the scenario response unit 11 has finished executing the scenario, the operator connection unit 12 transmits the collected information stored in the collected information storage unit 22 during the execution of the scenario to the operator terminal 3. -i and then the user terminal 2 and the operator terminal 3 -i A plurality of operator terminals 3 are connected. -1 , 3 -2 , ..., 3 -n Which operator terminal 3 -i The response method determination unit 14 determines whether to connect the user of the user terminal 2 connected by the operator connection unit 12 to the operator terminal 3 (details of this will be described later). -i The operator can communicate using text, voice or video.

[0025] The evaluation information storage unit 23 stores evaluation information regarding the response ability of an operator for each operator in advance. An operator's response ability indicates in which field of inquiries the operator has sufficient information gathering ability. Information gathering ability includes the ability to extract issues and the ability to gather information necessary for resolving issues. The evaluation information is information indicating the field in which the operator is evaluated as having sufficient information gathering ability. This evaluation is performed manually, and the evaluation results are stored in the evaluation information storage unit 23 as evaluation information. If there are multiple fields in which an operator is evaluated as having sufficient information gathering ability, information indicating those multiple fields is stored in the evaluation information storage unit 23 as evaluation information for that operator.

[0026] The operating status storage unit 24 stores operating status information indicating the operating status of an operator regarding a dialogue with a user. The operating status of an operator is, for example, either "in service," which is set when the operator is in dialogue with a user, or "not in service," which is set when the operator is not in dialogue with a user. Note that the classification of the operating status is not limited to this. For example, the operating status may be either "in service," which is set when the operator is in dialogue with a user, "in post-processing," which is set when post-processing is being performed after the dialogue with the user has ended, or "standby," which is set when the operator is not in dialogue with a user or performing post-processing and can immediately start responding.

[0027] The operation status management unit 13 sequentially updates and manages the operation status information stored in the operation status storage unit 24 for each operator. For example, the operation status management unit 13 acquires information that can identify the operation status of the operator from the operator terminal 3, and updates the operation status information in the operation status storage unit 24 based on the acquired information. For example, a user interface that allows the operator to select an operation status classification through operation on the operator terminal 3 can be displayed on the operator terminal 3, and classification information selected through operation on the user interface can be transmitted from the operator terminal 3 to the operation status management unit 13.

[0028] As another example, the operating status management unit 13 may monitor the operating status of the operator connection unit 12, and recognize the operating status of an operator corresponding to an operator terminal 3 connected to the user terminal 2 by the operator connection unit 12 as "in service", and recognize the operating status of an operator corresponding to an operator terminal 3 not connected to the user terminal 2 as "not in service". Furthermore, the operating status of an operator corresponding to an operator terminal 3 within a predetermined time after the connection with the user terminal 2 was disconnected may be recognized as "in post-processing", and the operating status of an operator corresponding to an operator terminal 3 for which a predetermined time or more has passed since the connection was disconnected may be recognized as "on standby".

[0029] When a user makes an inquiry to the chatbot, the response method determination unit 14 determines a response method including a scenario to be applied to the chatbot and an operator (hereinafter referred to as the operator scheduled to respond) who will respond after the chatbot responds, based on the content of the inquiry (the field confirmed by executing the field confirmation scenario), the operator's evaluation information stored in the evaluation information storage unit 23, and the operator's operating status information stored in the operating status storage unit 24.

[0030] The response method determination unit 14 executes the processing of the operator search unit 14A and the scenario identification unit 14B to determine the response method. When a user makes an inquiry, the operator search unit 14A searches for an available operator as a scheduled responding operator based on the operating status information of the operators stored in the operating status storage unit 24. An available operator means, for example, an operator who can respond as soon as possible after the completion of the chatbot scenario execution. For example, the operator search unit 14A searches for an operator whose operating status is set to "not responding" or "on standby" as a available responding operator. Note that this search may be performed after the field of the inquiry is confirmed by the scenario response unit 11 executing a field confirmation scenario.

[0031] Here, if multiple operators that fit the above classification are found, the operator search unit 14A selects one of them. For example, it selects randomly. On the other hand, if no matching operators are found, the operator search unit 14A searches for an operator whose operating status is set to "in service" or "in post-processing" as a scheduled operator. In this case, after the scenario response unit 11 has finished executing the chatbot scenario, the operator connection unit 12 waits until it can confirm that the operating status of the scheduled operator has changed to "not in service" or "on standby", and then connects the operator terminal 3 of the scheduled operator to the chatbot. -i Make a connection to.

[0032] The scenario identification unit 14B identifies a scenario to be applied to the chatbot based on the content of the inquiry (the field confirmed by executing the field confirmation scenario) and the evaluation information of the scheduled handling operator searched for by the operator search unit 14A. Here, the scenario identification unit 14B determines whether the scheduled handling operator can respond to the content of the inquiry based on the evaluation information of the scheduled handling operator, and identifies a superficial scenario if it is determined that the operator can respond, whereas it identifies a deep scenario if it is determined that the operator cannot respond.

[0033] That is, the scenario identification unit 14B determines whether the field of the inquiry confirmed by executing the field confirmation scenario matches the field indicated in the evaluation information of the scheduled handling operator stored in the evaluation information storage unit 23. If the fields match, the scenario identification unit 14B identifies a superficial scenario, assuming that the scheduled handling operator has the ability to respond to the content of the user's inquiry (has the ability to collect necessary and sufficient information). Because the time required to execute a superficial scenario is shorter than that of a deep scenario, it is possible to shorten the dialogue in the chatbot as much as possible and connect to an operator with sufficient response ability.

[0034] If the fields match, the execution of the superficial scenario is also skipped and the operator terminal 3 of the operator who is scheduled to handle the case is used. -i The operator terminal 3 may be connected to the user terminal 2 by the operator connection unit 12. -i Through dialogue with users, operators not only extract issues and collect information necessary for problem solving, but also collect formal information.

[0035] On the other hand, if the fields do not match, the scenario identification unit 14B identifies a deep scenario, assuming that the operator scheduled to handle the inquiry does not have the ability to adequately respond to the content of the user's inquiry. As described above, a deep scenario is a scenario that includes response sentences for extracting the issues the user has, response sentences for collecting information necessary for resolving the issues, etc. The information collected by executing this deep scenario is provided to the operator scheduled to handle the inquiry, so even an operator with low information gathering ability in the relevant field can take over from the chatbot and respond.

[0036] 3 is a flowchart showing an example of the operation of the response control device 1 according to the present embodiment configured as described above. The flowchart shown in FIG. 3 starts when an inquiry is made from the user terminal 2 to the chatbot of the response control device 1.

[0037] First, the operator search unit 14A searches for available operators as scheduled operators based on the operating status information of the operators stored in the operating status storage unit 24 (step S1). Next, the scenario response unit 11 executes a field confirmation scenario to confirm the field to which the inquiry belongs (step S2). That is, the scenario response unit 11 sends a response sentence to the user terminal 2 to confirm the field of the inquiry, and obtains information indicating the field from the user terminal 2 as a reply. Note that the processing of step S1 and the processing of step S2 may be performed in reverse order.

[0038] Next, the scenario identification unit 14B determines whether the operator scheduled to respond can respond to the content of the user's inquiry based on the field of the inquiry confirmed by executing the field confirmation scenario and the evaluation information of the operator scheduled to respond stored in the evaluation information storage unit 23 (fields evaluated as having sufficient information gathering ability) (step S3).

[0039] If it is determined that the operator is available to respond, the scenario identification unit 14B identifies a superficial scenario (step S4). On the other hand, if it is determined that the operator is not available to respond, the scenario identification unit 14B identifies a deep scenario (step S5). Once the scenario identification unit 14B identifies a scenario, the scenario response unit 11 executes the identified scenario (steps S6 to S9).

[0040] That is, the scenario response unit 11 transmits the response sentences included in the identified scenario to the user terminal 2 (step S6), acquires answer information from the user terminal 2 as a reply (step S7), and stores the acquired answer information in the collection information storage unit 22 (step S8).The scenario response unit 11 then determines whether or not all of the response sentences included in the identified scenario have been transmitted to the user terminal 2 (step S9), and returns to step S6 if any unsent response sentences remain.

[0041] When all of the response sentences included in the identified scenario have been sent to the user terminal 2, that is, when the execution of the scenario has been completed, the operator connection unit 12 connects the operator terminal 3 of the operator scheduled to handle the scenario searched for in step S1 to the user terminal 2. -i (Step S10). Here, it is determined whether the operating status of the operator scheduled to handle the call is "not handling" or "on standby." If the operating status is not "not handling" or "on standby," the determination in Step S10 is repeated until it is.

[0042] The operator connection unit 12 is connected to the operator terminal 3 of the operator who is scheduled to handle the call. -i When it is determined that the operator terminal 3 of the operator scheduled to handle the scenario is in a state where the operator terminal 3 of the operator scheduled to handle the scenario is in a connectable state, the collected information stored in the collected information storage unit 22 during the execution of the scenario is -i (Step S11), and the user terminal 2 and the operator terminal 3 -i (Step S12). This completes the process of the flowchart shown in FIG. 3. -i The connection with the user is terminated by the user or operator after the interaction is completed.

[0043] Although the example in which the scenario identified by the scenario identification unit 14B in step S4 or step S5 is executed to the end has been described, the present invention is not limited to this. For example, a loop process may be formed that returns from step S9 to step S1, and each time a response sentence is sent to the user terminal 2 and a response is obtained, the search for the operator scheduled to handle the situation and the identification of the scenario may be re-executed. Here, the process of step S2 is omitted from the second and subsequent loop processes.

[0044] As explained in detail above, in this embodiment, in a response system in which a chatbot responds to an inquiry from a user, and then an operator can continue to respond using information collected by the chatbot, when a user makes an inquiry, a response method is determined based on the content of the inquiry, evaluation information regarding the operator's response ability, and information about the operator's operating status, including a scenario to be applied to the chatbot and an operator who is scheduled to respond after the chatbot responds.

[0045] According to this embodiment, which is configured in this manner, unlike conventional technology in which long and short scenarios are selected based on the operating status of the operator, the chatbot scenario and the operator scheduled to respond are selected based on the operating status and response ability of the operator.Therefore, an appropriate combination of scenario and operator scheduled to respond can be identified based not only on the operating status of the operator but also on the content of the inquiry from the user, allowing for appropriate handover from the chatbot to the operator.

[0046] (First Modification) In the above embodiment, even when it is determined that an operator searched for by the operator search unit 14A as a scheduled responding operator is unable to respond to the content of the user inquiry, an example has been described in which the scheduled responding operator continues to respond on behalf of the chatbot after executing a deep scenario. However, this is not limited to this. For example, when it is determined that the scheduled responding operator is unable to respond to the content of the user inquiry, another available operator (who may not be an operator who is immediately available depending on the operating status) may be searched for again as a scheduled responding operator.

[0047] Fig. 4 is a block diagram showing an example of the functional configuration of a response control device 1' according to the first modified example. In Fig. 4, components assigned the same reference numerals as those shown in Fig. 2 have the same functions, and therefore redundant explanations will be omitted here. As shown in Fig. 4, the response control device 1' according to the first modified example includes a response method determination unit 14' instead of the response method determination unit 14. The response method determination unit 14' includes an operator search unit 14A' and a scenario identification unit 14B' instead of the operator search unit 14A and the scenario identification unit 14B.

[0048] When an inquiry is received from a user, the operator search unit 14A' first searches for an available operator as a scheduled operator based on the operating status information of the operators stored in the operating status storage unit 24. This is the same as the operator search unit 14A described in the above embodiment. That is, the operator search unit 14A' searches for an operator whose operating status is set to "not operating" or "on standby" as a scheduled operator.

[0049] The scenario identification unit 14B' determines whether the operator scheduled to respond can respond to the content of the inquiry based on the content of the inquiry (the field confirmed by executing the field confirmation scenario) and the evaluation information of the operator scheduled to respond searched by the operator search unit 14A', and if it is determined that the operator can respond, it identifies a scenario with a small number of responses, while if it is determined that the operator cannot respond, it identifies a scenario with a large number of responses.

[0050] Here, scenarios with a small number of responses are basically superficial scenarios, but among them, scenarios with a particularly small number of response sentences are identified. For example, a superficial scenario designed to collect one piece of information per response sentence is a scenario with a small number of responses. Note that even deep scenarios with a small number of responses can be identified here. The purpose of identifying scenarios with a small number of responses is to enable the chatbot to switch to an operator as quickly as possible.

[0051] On the other hand, a scenario with a large number of responses is basically a deep scenario, but it does not necessarily have to be a deep scenario. In the first variant, another operator who can respond to the user's inquiry is searched again as the scheduled operator, so it is not necessary for the chatbot to collect deep information. For example, the scenario identified by the scenario identification unit 14B' when it is determined that the scheduled operator is not able to respond to the content of the inquiry may be a superficial scenario that includes many response sentences.

[0052] When the scenario identification unit 14B' determines that the scheduled operator is not able to respond to the inquiry, the operator search unit 14A' re-searches for an operator who can respond to the content of the inquiry as the scheduled operator based on the content of the inquiry (the field confirmed by executing the field confirmation scenario) and the evaluation information of the operator stored in the evaluation information storage unit 23.

[0053] As described above, in this case, the scenario identification unit 14B' identifies a scenario with a large number of responses. The scheduled responding operators who are re-searched from the perspective of response capability may be able to respond to inquiries from users, but may not be able to respond immediately from the perspective of their operating status. The reason for identifying a scenario with a large number of responses is to buy time until the searched scheduled responding operators become available to respond.

[0054] Fig. 5 is a flowchart showing an example of the operation of the response control device 1' according to the first modified example configured as described above. The flowchart shown in Fig. 5 starts when an inquiry is made to the chatbot of the response control device 1' from the user terminal 2. In Fig. 5, the processes assigned the same step numbers as those shown in Fig. 3 are the same, so duplicated explanations will be omitted here.

[0055] In step S3 of Figure 5, if the scenario identification unit 14B' determines that the operator scheduled to respond found in step S1 is able to respond to inquiries from users, the scenario identification unit 14B' identifies a scenario with a low number of responses (step S21).

[0056] On the other hand, if the scenario identification unit 14B' determines that the scheduled operator searched for in step S1 is not able to respond to the user's inquiry, the scenario identification unit 14B' identifies a scenario with a large number of responses (step S22). Furthermore, the operator search unit 14A' searches again for an operator who can respond to the content of the inquiry as a scheduled operator based on the operator evaluation information stored in the evaluation information storage unit 23 (step S23).

[0057] After the processing of step S21 or step S23 is executed, the processing of steps S6 to S12 is executed, and the processing of the flowchart shown in Fig. 5 is completed. Note that a loop process may be formed by returning from step S9 to step S1, and each time a response sentence is sent to the user terminal 2 and a reply is obtained, the search for a scheduled operator and the identification of a scenario may be re-executed.

[0058] (Second Modification) In the above embodiment, when a user makes an inquiry, first a search is made for an available operator from the perspective of the operator's operating status, and then it is determined whether the available operator can respond to the content of the user's inquiry. However, the present invention is not limited to this. For example, when a user makes an inquiry, first a search is made for an operator who can respond to the inquiry from the perspective of the operator's response ability, and then a scenario is identified from the perspective of the operating status of the searched available operator.

[0059] Figure 6 is a block diagram showing an example of the functional configuration of a response control device 1" relating to the second modified example. In this Figure 6, components given the same symbols as those shown in Figure 2 have the same functions, so duplicate explanations will be omitted here. As shown in Figure 6, the response control device 1" relating to the second modified example is equipped with a response method determination unit 14" instead of the response method determination unit 14. The response method determination unit 14" is equipped with an operator search unit 14A", a first response method determination unit 14C and a second response method determination unit 14D instead of the operator search unit 14A and the scenario determination unit 14B.

[0060] When an inquiry is made by a user, the operator search unit 14A" searches for an operator who can respond to the content of the inquiry based on the content of the inquiry and the evaluation information of the operator stored in the evaluation information storage unit 23. In other words, the operator search unit 14A" searches for an operator whose field of the inquiry confirmed by executing the field confirmation scenario matches the field indicated by the evaluation information stored in the evaluation information storage unit 23, as an operator to be scheduled to respond. Here, if multiple applicable operators are found, the operator search unit 14A" selects one of them. For example, it may select one operator at random.

[0061] When the operator search unit 14A″ searches for an operator who can respond, the first response method identification unit 14C determines whether or not the searched operator can respond within a predetermined time based on the operating status information stored in the operating status storage unit 24 for the searched operator, and if the operator can respond, identifies the searched operator as an operator who is scheduled to respond, and also identifies a scenario with a small number of responses.

[0062] For example, if the operating status of an operator is set to "not responding" or "on standby," the first response method identification unit 14C determines that the operator can respond within a predetermined time. Alternatively, the operating status management unit 13 may not only manage the operating status of each agent as a classification as described above, but also manage the estimated remaining time required for the operator to become available, and may determine that the operator can respond if the estimated remaining time is less than or equal to a predetermined time.

[0063] The method of managing the estimated remaining time required until the customer can respond is arbitrary. For example, the information collected by the chatbot is transmitted to the operator terminal 3 by the operator connection unit 12. -i When the message is sent to the operator terminal 3, a user interface that allows the operator to input an expected time required for the response is displayed on the operator terminal 3, and the expected time required is input by the operator through an operation on the user interface. -i to the operation status management unit 13. The operator checks the chatbot's collected information sent by the operator connection unit 12 and inputs the estimated time required to complete a series of responses between the operator and the user. The operation status management unit 13 then successively updates the estimated remaining time by successively subtracting the elapsed time from the estimated required time.

[0064] As another example, multiple combinations of information collected by the chatbot and the actual time required for an operator to respond are prepared as learning data, and a prediction model is generated using the learning data so that a predicted time required is output when collected information is input. The operating status management unit 13 may obtain the predicted time required using the prediction model, and successively subtract the elapsed time from the obtained predicted time required, thereby successively updating the predicted remaining time until an operator becomes available to respond.

[0065] If the operator search unit 14A" does not find an operator who can respond, or if it is determined that the operator who was searched for by the operator search unit 14A" and is able to respond cannot respond within a specified time, the second response method identification unit 14D searches for another operator who can respond as an operator who is scheduled to respond based on the operating status information of the operator stored in the operating status memory unit 24, and identifies a deep scenario or a scenario with a large number of responses. For example, the second response method identification unit 14D identifies a deep scenario if no operator who can respond is found, and identifies a scenario with a large number of responses if it is determined that the operator who can respond cannot respond within a specified time.

[0066] In the case of the second modified example, the operator search unit 14A" searches for operators who are scheduled to respond from the perspective of their response ability, without taking into account the operating status of the operators, so it is basically not the case that no operators who can respond are found. However, if operating status categories such as "away from desk," "on break," and "on vacation" are added and managed, and operators with these categories are excluded from the search, it is possible that no operators who can respond are found.

[0067] If no operator capable of responding to the inquiry is found, an operator who is not sufficiently capable of responding to the inquiry will respond instead of the chatbot. Therefore, in this case, a deep scenario is identified and chatbot processing is executed. On the other hand, if an operator capable of responding to the inquiry is found but it is determined that the operator cannot respond within the specified time, a scenario with a large number of responses is identified and chatbot processing is executed to buy time until the operator becomes available.

[0068] Here, an example has been described in which, when the operator search unit 14A" finds multiple operators who can respond to the contents of an inquiry, one of them is randomly selected; however, the present invention is not limited to this. For example, an operator who can respond within a specified time may be selected based on the operating status information of the multiple searched operators. Then, when an operator who can respond within the specified time can be selected, the first response method identification unit 14C identifies the selected operator as the operator scheduled to respond, and also identifies a scenario with a small number of responses.

[0069] FIG. 7 is a flowchart showing an example of the operation of the response control device 1" according to the second modified example configured as described above. The flowchart shown in FIG. 7 starts when an inquiry is made from the user terminal 2 to the chatbot of the response control device 1". In FIG. 7, the processes assigned the same step numbers as those shown in FIG. 3 are the same, so duplicate explanations will be omitted here.

[0070] First, the scenario response unit 11 executes a field confirmation scenario to confirm the field to which the content of the inquiry belongs (step S31). Next, the operator search unit 14A" searches for an operator who can respond to the content of the inquiry based on the content of the user's inquiry (the field of the inquiry confirmed by executing the field confirmation scenario) and the evaluation information of the operator stored in the evaluation information storage unit 23 (step S32).

[0071] Here, the first response method identification unit 14C determines whether or not an operator who can respond has been searched for by the operator search unit 14A" (step S33). If an operator who can respond has been searched for, the first response method identification unit 14C determines whether or not the searched operator can respond within a predetermined time based on the operating status information stored in the operating status storage unit 24 for the searched operator (step S34). Here, if it is determined that the searched operator can respond within the predetermined time, the first response method identification unit 14C identifies the operator as an operator scheduled to respond, and identifies a scenario with a small number of responses (step S35).

[0072] If the first response method identification unit 14C determines in step S33 that no operator available to respond was found, or if the first response method identification unit 14C determines in step S34 that the searched operator available to respond will not be able to respond within the specified time, the second response method identification unit 14D searches for another operator available to respond as a scheduled operator based on the operating status information of the operator stored in the operating status memory unit 24 (step S36), and identifies a deep scenario or a scenario with a large number of responses (step S37).

[0073] After the processing of step S35 or step S37 is executed, the processing of steps S6 to S12 is executed, and the processing of the flowchart shown in Fig. 7 is terminated. Note that a loop process may be formed that returns from step S9 to step S31, and each time a response sentence is sent to the user terminal 2 and a reply is obtained, the search for a scheduled operator and the identification of a scenario may be re-executed.

[0074] In the above embodiment and each modified example, an example has been described in which evaluation information regarding an agent's response ability is information indicating which fields of inquiries the agent has sufficient information gathering ability for, i.e., information indicating the fields in which the agent is evaluated as having sufficient information gathering ability. However, the present invention is not limited to this. For example, evaluation information indicating whether the agent has high information gathering ability regardless of the field may be used. For example, the results of a human evaluation of an agent from the perspectives of the agent's ability to understand what the other party is saying, their ability to ask questions, the amount of experience in actual responses, the amount of knowledge, etc. may be used as evaluation information. When such evaluation information is used, it is possible to omit first executing a field confirmation scenario when a user makes an inquiry.

[0075] In addition, in the above embodiment, the estimated remaining time required for the identified scheduled responding operator to become available to respond to a conversation with the user may be managed, and a scenario having a required time close to the estimated remaining time may be identified.

[0076] Furthermore, the above-described embodiment and each of the modifications are merely examples of specific embodiments of the present invention, and the technical scope of the present invention should not be construed as being limited by these. In other words, the present invention can be embodied in various forms without departing from the gist or main features thereof.

[0077] 1, 1', 1" Response control device 2 User terminal 3 Operator terminal 11 Scenario response unit 12 Operator connection unit 13 Operation status management unit 14, 14', 14" Response method determination unit 14A, 14A', 14" Operator search unit 14B, 14B' Scenario identification unit 14C First response method identification unit 14D Second response method identification unit 21 Scenario storage unit 22 Collected information storage unit 23 Evaluation information storage unit 24 Operation status storage unit

Claims

1. A response control device applied to a response system in which, after a chatbot responds to an inquiry from a user, an operator can continue to respond using information collected by the chatbot, an evaluation information storage unit that stores evaluation information regarding the response ability of the operator for each of the operators in advance; an operation status management unit that sequentially updates and manages operation status information indicating an operation status of the operator for each of the operators; a response method determination unit that, when an inquiry is made from the user, determines a response method including a scenario to be applied to the chatbot and an operator scheduled to respond after a response by the chatbot based on the content of the inquiry, evaluation information of the operator, and operating status information of the operator, The response method determination unit is an operator search unit that searches for an available operator as a scheduled responding operator based on the operating status information of the operator when an inquiry is received from the user; a scenario specification unit that specifies a scenario to be applied to the chatbot based on the content of the inquiry and the evaluation information of the scheduled responding operator searched for by the operator search unit. A response control device comprising:

2. The response control device described in claim 1, characterized in that the scenario identification unit determines whether the operator scheduled to respond is able to respond to the content of the inquiry based on evaluation information of the operator scheduled to respond, and if it is determined that the operator is able to respond, identifies a superficial scenario for collecting superficial information regarding the inquiry from the user, while if it is determined that the operator is not able to respond, identifies a deep scenario for collecting deep information regarding the inquiry from the user.

3. The scenario identification unit determines whether the scheduled operator can respond to the content of the inquiry based on evaluation information of the scheduled operator, and if it is determined that the scheduled operator can respond, identifies a scenario with a small number of responses, whereas if it is determined that the scheduled operator cannot respond, identifies a scenario with a large number of responses; When the scenario identification unit determines that the operator is not capable of responding, the operator search unit searches again for an operator capable of responding to the content of the inquiry as the scheduled operator based on the evaluation information of the operator stored in the evaluation information storage unit.

2. The response control device according to claim 1,

4. A response control device applied to a response system in which, after a chatbot responds to an inquiry from a user, an operator can continue to respond using information collected by the chatbot, an evaluation information storage unit that stores evaluation information regarding the response ability of the operator for each of the operators in advance; an operation status management unit that sequentially updates and manages operation status information indicating an operation status of the operator for each of the operators; a response method determination unit that, when an inquiry is made from the user, determines a response method including a scenario to be applied to the chatbot and an operator scheduled to respond after a response by the chatbot based on the content of the inquiry, evaluation information of the operator, and operating status information of the operator, The response method determination unit is an operator search unit that searches for an operator who can respond to the content of an inquiry from the user based on the content of the inquiry and evaluation information of the operator; a first response method identification unit that, when the operator search unit has searched for the operator capable of responding, determines whether the searched operator is capable of responding within a predetermined time based on operating status information of the searched operator, and, if the operator is capable of responding, identifies the searched operator as an operator scheduled to respond, and also identifies a scenario with a small number of responses; and a second response method specification unit that, when the operator search unit does not search for an operator capable of responding, or when it is determined that the operator searched for by the operator search unit is not capable of responding within a predetermined time, searches for another operator capable of responding as the scheduled operator based on operating status information of the operator, and specifies a deep scenario for collecting deep information regarding the inquiry from the user or a scenario with a large number of responses. A response control device comprising:

5. A method for controlling a response in a response system in which a chatbot responds to an inquiry from a user and then an operator can continue to respond using information collected by the chatbot, comprising: A first step in which a response method determination unit of a response control device, when receiving an inquiry from the user, determines a response method including a scenario to be applied to the chatbot and an operator scheduled to respond after the response by the chatbot based on the content of the inquiry, evaluation information on the response ability of the operator, and operating status information indicating the operating status of the operator; A second step in which a scenario response unit of the response control device executes processing of the chatbot according to the scenario determined by the response method determination unit; and a third step in which an operator connection unit of the response control device executes a process of connecting to the operator determined by the response method determination unit after the scenario response unit executes the process of the chatbot, In the first step, the response method determination unit an operator search process for searching for an available operator as a scheduled operator based on the operating status information of the operator when an inquiry is received from the user; A scenario specification process is executed to specify a scenario to be applied to the chatbot based on the content of the inquiry and the evaluation information of the scheduled responding operator searched for by the operator search process. A response control method comprising:

6. A method for controlling responses in a response system in which, after a chatbot responds to an inquiry from a user, an operator can continue to respond using information collected by the chatbot, comprising: A first step in which a response method determination unit of a response control device, when receiving an inquiry from the user, determines a response method including a scenario to be applied to the chatbot and an operator scheduled to respond after the response by the chatbot based on the content of the inquiry, evaluation information on the response ability of the operator, and operating status information indicating the operating status of the operator; A second step in which a scenario response unit of the response control device executes processing of the chatbot according to the scenario determined by the response method determination unit; and a third step in which an operator connection unit of the response control device executes a process of connecting to the operator determined by the response method determination unit after the scenario response unit executes the process of the chatbot, In the first step, the response method determination unit an operator search process for searching for an operator who can respond to the content of an inquiry from the user based on the content of the inquiry and evaluation information of the operator; a first response method specification process for, when an operator capable of responding is searched for by the operator search process, judging whether or not the searched operator can respond within a predetermined time based on operating status information of the searched operator, and, if the operator can respond, specifying the searched operator as an operator scheduled to respond, and specifying a scenario with a small number of responses; When the operator search process does not find an operator capable of responding, or when the first response method specification process determines that the operator capable of responding searched for by the operator search process is not capable of responding within a predetermined time, a second response method specification process is executed to search for another operator capable of responding as the scheduled responding operator based on the operating status information of the operator, and to specify a deep scenario for collecting deep information regarding the inquiry from the user or a scenario with a large number of responses. A response control method comprising: