Storage medium for response time manager, response time management device, and chat system

Abstract

Provided is a storage medium for a response time management program, a response time management device, and a chat system. The storage medium includes: a response information transmitting unit that transmits, in response to transmission of a question from a user terminal (500), answer information created by an operator based on information obtained by accessing an information source (200, 300, 350) to the user terminal (500) from an operator terminal (10); an elapsed time measuring unit (118) that measures an elapsed time from reception of the question by the user terminal (500) to transmission of the answer information to the user terminal (500); a reference response time determining unit (114) that determines a reference response time from reception of the question by the user terminal (500) to transmission of the answer information to the user terminal (500) based on an amount of information of the answer information transmitted from the operator terminal (10) to the user terminal (500); and a response time management unit that temporarily holds transmission of the answer information when it is detected that the operator has performed an operation of transmitting the answer information to the user terminal (500) before the elapsed time reaches the reference response time, and transmits the answer information after the reference response time is reached.

Description

Technical Field

[0001] This invention relates to a storage medium, a response time management device, and a chat system suitable for use with human chat services. Background Technology

[0002] A system is known to accept input from a user at a communication terminal and output a response to the input via a network. Patent Document 1 discloses a technique for a response device to output a response to a string input by a user via a communication terminal. The response device accepts a string input by the user, performs a first response process, and determines a response to the accepted string. At this time, a sub-processing unit performs a second response process in parallel with the first response process, determining a response to the received string. Then, the response from the first response process is output, and if the information related to the user meets predetermined conditions, the response from the second response process is output in addition to the response from the first response process, thereby enabling a response suitable for the user.

[0003] In human chat services, human operators receiving user inquiries need to handle diverse questions from various users. Therefore, human operators can leverage generative AI systems to obtain the necessary information to respond to users. Using generative AI systems, a large amount of information can be obtained in a shorter time. However, when human operators reply to users with a large amount of information in too short a time, users may feel as if they are receiving a response from an unmanned system, potentially leading to decreased customer satisfaction. This situation can also occur when human operators, due to their extensive knowledge, immediately send a large amount of information to users in response to the inquiries received.

[0004] Existing technical documents

[0005] Patent documents Patent document 1: Japanese Patent Application Publication No. 2019-159880 (JP2019-159880A). Summary of the Invention

[0006] One technical solution of the present invention provides a storage medium that can be read by a computer. This storage medium stores a response time management program executable by a computer equipped with a processor and a storage unit. The response time management program manages the time elapsed from when an operator's terminal receives a query from a customer's terminal, and the operator accesses an information source to obtain a response based on the original response obtained from the information source, until the operator sends the response from the operator's terminal to the user terminal. The response time management program causes the computer to perform the following processing: detect the elapsed time from when the operator's terminal receives the query from the user terminal to when it sends a response; determine a baseline response time based on the amount of information in the response sent from the operator's terminal to the user terminal; and if the operator sends a response before the baseline response time has elapsed, temporarily hold off on sending the response until the baseline response time has elapsed.

[0007] Another technical solution of the present invention provides a response time management device that manages the time from when an operator terminal receives a question from a customer-operated user terminal, the operator accesses an information source to obtain a response based on the information obtained from the information source (i.e., the original response), until the operator terminal sends the response to the user terminal. The device comprises: an elapsed time detection unit that detects the elapsed time from when the operator terminal receives the question from the user terminal until it sends the response to the user terminal; a reference response time determination unit that determines a reference value (i.e., a reference response time) based on the amount of information in the response sent from the operator terminal to the user terminal; and a response information sending unit that, if it detects that the operator has sent the response to the user terminal before the elapsed time reaches the reference response time, temporarily holds the sending of the response to the user terminal, and sends the response after the reference response time has been reached.

[0008] Another technical solution of the present invention provides a human chat system comprising a user terminal operated by a customer, an operator terminal operated by an operator, and a response time management device. The response time management device manages the time elapsed from when the operator terminal receives a question from a customer via the user terminal, when the operator accesses an information source to obtain a response based on the original response obtained from the information source, until the response is sent from the operator terminal to the user terminal. The response time management device comprises: an elapsed time detection unit that detects the elapsed time from when the operator terminal receives the question from the user terminal until it sends a response to the user terminal; a reference response time determination unit that determines a reference value, i.e., a reference response time, based on the amount of information in the response sent from the operator terminal to the user terminal; and a response information sending unit that, if it detects that the operator has sent a response to the user terminal before the elapsed time reaches the reference response time, temporarily holds the sending of the response information and sends it only after the reference response time has elapsed. Attached Figure Description

[0009] The objectives, features, and advantages of the present invention are further illustrated by the following description of embodiments in conjunction with the accompanying drawings.

[0010] Figure 1 This is a concept diagram illustrating the general structure of a human chat system; Figure 2 This is a block diagram illustrating the functional modules of the computer that constitutes the response time management device and the constituent elements connected to the response time management device via a network; Figure 3 This is a flowchart illustrating the general processing steps of the response time management process executed by the CPU of the response time management device; Figure 4 This is another flowchart illustrating the general processing steps of the response time management process performed by the CPU of the response time management device. Detailed Implementation

[0011] The following is for reference Figures 1-4 Embodiments of the present invention will be described below. In all the accompanying drawings described below, common components are labeled with the same reference numerals, and repeated descriptions are omitted. It should be noted that not all components shown in the following embodiments are necessarily essential components of the present invention.

[0012] Figure 1This is a conceptual diagram illustrating the general structure of a human chat system. Operator terminal 10, operated by a human chat operator (hereinafter referred to as operator), is connected via network NW to external LLM 200, internal generative AI system (In-house Generative AI) 300, internal database (In-house DB) 350, customer database 400, and user terminal 500. Server 100, constituting a response time management device, is connected to both operator terminal 10 and network NW. Server 100 and operator terminal 10 may also be connected via network NW without a direct connection. Furthermore, the response time management function of server 100 described below can also be included in operator terminal 10. The internal generative AI system 300 and internal database 350 are connected to a common search source 310. The internal generative AI system 300 and internal database 350 (internal DB) may also be connected to separate search sources instead of the common search source 310.

[0013] External LLM 200 can employ, for example, LLMs (Large Language Models) provided by third parties. An LLM is a large-scale artificial intelligence model for Natural Language Processing (NLP) built using massive datasets of text and deep learning techniques. By learning from large amounts of text data (web pages, books, articles, etc.), an LLM can understand patterns in human language and effectively perform Natural Language Generation (NLG) tasks.

[0014] The following example illustrates the use of LLM as an external resource for providing human chat services to users. However, other external services that leverage AI, such as ChatGPT, Google Gemini (registered trademark), Cerence (registered trademark), and Claude 3 (registered trademark), can also be used. Additionally, depending on the performance requirements of human chat, other natural language processing methods, including SLM (Small Language Model), can be used.

[0015] The internal generative AI system 300 is a system that stores and builds information about the company that provides human chat services as a search source 310. This includes a wealth of documents such as the goods or services sold or provided by the company and their accompanying manuals, as well as information related to past FAQs (Frequently Asked Questions). This system can improve the accuracy of AI-generated answers and reduce illusions.

[0016] In addition, for questions raised by users driving passenger cars and other vehicles, especially those provided by our company, we can not only provide general answers, but also provide more specific responses, such as "Please press the switch with the green symbol in the lower left corner of the multiple turn signals on the right side of the steering wheel spokes." Furthermore, when receiving inquiries related to services such as vehicles provided by the company, or websites that sell goods or services on the Internet, namely EC (Electronic Commerce) websites, the company can provide users with more detailed and comprehensive information.

[0017] Furthermore, since it is our own system, it is advantageous that even with increased transactions between operators and AI during human chat sessions, cost increases can be mitigated.

[0018] The customer database 400 is a database that stores login information such as passwords, linked to the user ID entered by the user when logging into the live chat service.

[0019] User terminal 500 is a user-operated information processing device that provides human-operated chat services via connection to operator terminal 10. User terminal 500 can use, for example, smartphones, tablets, PDAs (Personal Digital Assistants), personal computers, etc. Additionally, user terminal 500 can also use IVI (In-vehicle Infortainment) devices connected to a wireless communication network and NW network.

[0020] Network NW is capable of using the Internet. It should be noted that the internal generative AI system 300, customer database 400, and server 100 can be connected via a closed-circuit network service or using VPN technology, without using the Internet. Alternatively, the internal generative AI system 300 and customer database 400 can be included within server 100. To enhance information security, strong encryption technology is preferably used to protect customer database 400.

[0021] A customer sends a query to an operator terminal 10 via a user terminal 500. Upon receiving the query, the operator accesses one or more of the following as needed: an external LLM 200, an internal generative AI system 300, or an internal database 350, to obtain an answer corresponding to the query from the customer. Hereinafter, the external LLM 200, internal generative AI system 300, or internal database 350 accessed by the operator can be simply referred to as the information source. Additionally, the answer obtained from the information source is sometimes called the original answer. Based on the original answer, the operator creates a response message for the customer and outputs this response message to the user terminal 500, which is the source of the query.

[0022] The above-described customer-operator interaction is explained in more detail. The operator sees the query information sent from user terminal 500 to operator terminal 10 and, as needed, can communicate with the customer via operator terminal 10 and user terminal 500 to confirm and organize the customer's intent and main point of the query. If the operator determines that they can answer the customer's query without accessing the aforementioned information source, they can operate operator terminal 10 to prepare a response.

[0023] On the other hand, when it is determined that access to an information source is required, the operator operates the operator terminal 10 to prepare a prompt for a question posed to the external LLM 200 or the internal generative AI system 300 and output it to the information source, or to prepare a search command for the internal database 350 and output it to the information source. The operator obtains the original answer corresponding to the prompt or search command from the information source accessed by the operator. The operator processes the original answer as needed, such as by editing, to create a response. When it is determined that no processing of the original answer is required, the operator can use the original answer as the response. The operator uses the operator terminal 10 to send the response to the user terminal 500, which is the source of the question.

[0024] Sometimes, customer inquiries are multifaceted or complex, leading to lengthy responses between customers and operators. Conversely, when accessing external LLM 200 or internal generative AI systems 300, even complex customer inquiries can result in a large volume of responses being sent to user terminal 500 within a short period. The time elapsed from when user terminal 500 outputs a question to when operator terminal 10 sends a response to that question will be referred to as the response time. To manage this response time in the human chat service, server 100 functions as a response time management device.

[0025] Server 100 includes a CPU (Central Processing Unit) 101, internal memory 102, I / O 104, storage device 105, and a data bus 103 capable of exchanging data between these elements.

[0026] A response time management program for managing response time can be stored in storage device 105, read from storage device 105 when server 100 starts, loaded into main memory 102, and executed by CPU 101. Alternatively, main memory 102 may include ROM (Read-Only Memory), flash memory, etc., and a chat program stored in these ROMs or flash memories can be loaded into main memory 102 and executed by CPU 101. Additionally, programs read from an optical disc drive or card reader (not shown) or downloaded via network NW can also be loaded into main memory 102 and executed by CPU 101.

[0027] It should be noted that by distributing all or part of the hardware components across multiple computers and interconnecting them via a network (NW), a server 100 can be virtually constructed. Thus, the concept of server 100 includes not only the devices housed in a single casing or enclosure, but also a virtualized computer system.

[0028] Figure 2 This is a block diagram illustrating a response time management device implemented by executing a response time management program through the CPU 101 of server 100, and also showing other elements connected to server 100 via network NW.

[0029] As described above, the operator, based on the query information sent from the user terminal 500 to the operator terminal 10, accesses one or more of the following information sources: external LLM 200, internal generative AI system 300, and internal database 350, to obtain the original answer. Then, the original answer is edited or processed as needed to create a response message. The operator operates the operator terminal 10 to send the response message to the user terminal 500 that sent the query information. The response message input unit 110 inputs the response message sent from the operator terminal 10 and temporarily stores it.

[0030] The response information quantity determination unit 112 determines the response information quantity based on the amount of response information input to the response information input unit 110. For example, the response information quantity determination unit 112 can determine the response information quantity based on the text length of the response information and the number of key points (critical points) contained in the response. The response information quantity determination unit 112 is configured to, instead of the above, or in addition to the above, also analyze the query information sent from the user terminal 500 to the operator terminal 10 and determine the level of information quantity of the response information corresponding to that query information.

[0031] The baseline response time determination unit 114 determines the baseline response time based on the determination result of the response information content of the response information content determination unit 112. Here, the baseline response time is determined based on the information content of the response information sent from the operator terminal 10 to the user terminal 500, which is the source of the query information. For example, if a customer asks a question and the response to that question has little information content, even if the response is close to an immediate one, the customer will rarely feel uncomfortable. On the other hand, even if the response to a customer's question has a lot of information content, if it is close to an immediate one, even if there is a human conversation between the customer and the operator, the customer will feel that they are responding to a machine, which may reduce customer satisfaction.

[0032] Therefore, the baseline response time determination unit 114 is configured to determine the baseline response time in a manner that increases with the amount of response information. The relationship between the amount of response information and the baseline response time can be linear or non-linear. By setting a longer baseline response time when the amount of response information is large, as explained below, it is possible to reduce the sense of disharmony felt by customers using human chat due to the short response time compared to the amount of response information, thereby improving customer satisfaction.

[0033] The response issuance detection unit 116 detects when the operator operates the operator terminal 10 to issue a response message to the user terminal 500. The elapsed time measurement unit 118 measures and manages the time elapsed since the user terminal 500 outputs a query message to the operator terminal 10 (hereinafter referred to as the elapsed time after query acceptance). Specifically, the elapsed time measurement unit 118 monitors the elapsed time after query acceptance, starting from the moment the query message is output to the operator terminal 10 via the user terminal 500, and determines whether the reference response time determined by the reference response time determination unit 114 has been reached. Then, when it is determined that the elapsed time after query acceptance has reached the reference response time, the elapsed time measurement unit 118 issues a response message issuance instruction to the response message issuance unit 120.

[0034] Upon receiving the instruction from the time measurement unit 118, the response information issuing unit 120 sends the response information temporarily stored in the response information input unit 110 to the user terminal 500 that issued the question to the operator terminal 10.

[0035] However, it is also envisioned that in cases where the query information sent from the user terminal 500 is complex, or where there are many key points to answer, a relatively long time may be required for the operator to prepare the response. In such cases, the elapsed time measurement unit 118 is configured to monitor the elapsed time after the query is received, independent of the reference response time determined by the reference response time determination unit 114. When a predetermined time has elapsed since the user terminal 500 sent the query information to the operator terminal 10, a message such as "Preparing a response, but it will take some time. Please wait a little longer" can be sent to the user terminal 500. The predetermined time can be 15 seconds, 20 seconds, or 30 seconds.

[0036] As explained above, the response information determination unit 112 can be configured to analyze the query information sent from the user terminal 500 to the operator terminal 10 and determine the level of response information corresponding to that query information. With this configuration, when a query information is sent from the user terminal 500 to the operator terminal 10, the remaining time before the response information is sent from the operator terminal 10 to the user terminal 500 can be estimated. Therefore, based on the time elapsed since the query was received, as detected by the elapsed time measurement unit 118, an estimate of the remaining waiting time can be sent to the user terminal 500. The following will refer to... Figure 4 Explain it.

[0037] Figure 3This is a flowchart illustrating the general processing steps of the response time management process executed by the CPU 101 of server 100. The response time management process begins when a customer-operated user terminal 500 connects to an operator-operated operator terminal 10 to start a live chat session. This process can be repeated even when multiple responses are made within a single live chat session. Figure 3 The processing is shown. It should be noted that, in Figure 3 Before the processing shown, customer login and server 100 user authentication are performed as needed to start a human chat session.

[0038] In S300, CPU 101 determines whether a query message has been sent from the user terminal 500 operated by the customer to the operator terminal 10. During the period when the determination is negative, the processing of S300 is repeated. When a query message is detected from the user terminal 500 to the operator terminal 10 (S300: Yes), the processing enters S302, and CPU 101 starts a timer to count the time elapsed since the query was received.

[0039] In S304, CPU 101 determines whether the elapsed time since the question was received has reached the specified time. If the determination in S304 is affirmative (S304: Yes), the process proceeds to S306, sending a delayed response message to the user terminal 500 that sent the question. The delayed response message could be, for example, "Preparing to answer your question. Please wait a moment." The specified time could be 15 seconds, 20 seconds, or 30 seconds. During periods when user questions are complex, have many key points to answer, and require time for the operator to prepare the response, repeating S304 and S306 allows the customer to recognize that a response is being prepared.

[0040] The above illustrates an example of issuing a response delay message in S306 when a predetermined time has elapsed since the question was accepted in S304. However, alternatives or other methods can be used, as described below. That is, suppose the operator accesses external information sources such as LLM 200, internal generative AI 300, and internal database 350 to prepare response information, and embeds a link in the obtained original response. If the CPU 101 detects in S304 that the operator has pressed the link, the aforementioned response delay message can also be issued to the user terminal 500 in S306. The operator pressed the link because more time is needed to prepare the response information. Furthermore, not limited to pressing a link, if the CPU 101 detects in S304 that the operator, having obtained the original response, has performed a further information search operation, a response delay message can also be issued to the user terminal 500 in S306.

[0041] In S308, CPU 101 determines whether the operator has sent a response message from operator terminal 10 to user terminal 500. If the determination is negative (S308: No), it returns to S304. If the determination in S308 is positive (S308: Yes), processing proceeds to S310, where the baseline response time is determined based on the amount of information in the response message sent by the operator through operator terminal 10. It should be noted that the response message is not immediately sent to user terminal 500 but is temporarily stored in server 100.

[0042] In S312, CPU 101 determines whether the time elapsed since the query was accepted has reached the baseline response time. During the period when the determination is negative, the determination action in S312 is repeated. When the determination in S312 is positive (S312: Yes), CPU 101 sends the response information to user terminal 500 in S314.

[0043] use Figure 3 The response time management system, by responding to customers via live chat with a response time commensurate with the amount of information in the reply, can suppress the feeling of interacting with a person and improve customer satisfaction when using live chat. Furthermore, when preparing a reply takes time, it can prevent customers from mistakenly believing the chat session has been interrupted and leaving the conversation, or from becoming anxious due to prolonged silence, thereby improving customer satisfaction.

[0044] Reference Figure 4 Another example of the time management process of the present invention is illustrated. In Figure 4 In each of the processing steps shown, for the... Figure 3 Processing steps that are identical to those shown are labeled with the same reference numerals and their descriptions are omitted, in order to be consistent with... Figure 3 The explanation will focus on the differences in the processing methods shown. Figure 3 The baseline response time of S310 shown in the processing steps is determined by the processing time in... Figure 4 The process, as S400, occurs between S302 and S304, which is consistent with... Figure 3 The processing steps shown are different. Furthermore, the response information determination unit 112 is configured to analyze the question information sent from the user terminal 500 to the operator terminal 10 and determine the level of response information corresponding to the question information.

[0045] When a query is sent from the user terminal 500 operated by the customer to the operator terminal 10 operated by the operator, and the determination in S300 is affirmative (S300: Yes), after the time elapsed since the query was received begins in S302, the CPU 101 determines the baseline response time in S400. That is, based on the query message sent from the user terminal 500 to the operator terminal 10, the CPU estimates the remaining time until the operator terminal 10 sends a response message back to the user terminal 500, and determines the baseline response time.

[0046] In S304, the CPU 101 determines whether the elapsed time since the question was accepted has reached the specified time. If the determination in S304 is affirmative (S304: Yes), the process proceeds to S306, where a message indicating a delayed response is sent to the user terminal 500 that sent the question. At this time, based on the aforementioned baseline response time and elapsed time since the question was accepted, the remaining time until the operator terminal 10 sends a response to the user terminal 500 can be predicted. Therefore, an example of the delayed response message could be "Preparing to answer your question. An answer is expected in approximately 30 seconds." Alternatively, or in addition to the above, a bar graph or hourglass-like graphic display indicating the remaining waiting time can be shown on the user terminal 500's display screen. As time passes and approaches the baseline response time, the bar graph shortens or the sand in the hourglass falls from top to bottom. Furthermore, an audio signal indicating a shortening remaining waiting time can be emitted via the user terminal 500's speaker.

[0047] In S308, CPU 101 determines whether the operator has sent a response message from operator terminal 10 to user terminal 500. If the determination is negative (S308: No), it returns to S304. This is consistent with... Figure 3 The example shown is the same. When the determination in S308 is affirmative (S308: Yes), the process proceeds to S312. The CPU 101 determines whether the elapsed time since the query was accepted has reached the baseline response time. During the period when the determination is negative (S312: No), the determination action in S312 is repeated. When the determination in S312 is affirmative (S312: Yes), the CPU 101 sends the response information to the user terminal 500 in S314.

[0048] exist Figure 4In the response time management process shown, a link can be embedded in the original response received when the operator is preparing to answer, and the operator pressing the link can be detected in S304. In this case, a response delay message can also be sent to the user terminal 500 in S306. That is, if the time required for the operator to collect new information and prepare a response is longer than expected, a message conveying this meaning can also be sent to the user terminal 500. In S304, either the determination of whether the time elapsed after the question was accepted has reached the specified time and the determination of whether the operator pressed the link as described above can be performed, or both determinations can be performed.

[0049] exist Figure 4 In the flowchart, the following is omitted Figure 3 The flowchart shows S310, but the processing in S310 can also be omitted. In this way, although the baseline response time is determined in S400 based on the query information output from the user terminal 500, the baseline response time can be corrected if the amount of information in the response information sent from the operator terminal 10 differs from the amount of information predicted when determining the baseline response time in S400. As a result, when sending the response information to the user terminal 500 in S314, messages such as "Answer given earlier than expected" or "The response took longer than expected. We apologize" can also be added.

[0050] According to reference Figure 4 The response time management process described above, when a customer receives a response from an operator via live chat, not only responds with a response time corresponding to the amount of information in the response, but also suppresses any negative feelings associated with interacting with a person, thus improving customer satisfaction when using live chat. Furthermore, when preparing to send a response takes time, the estimated time until the response is sent can be communicated to the customer. This prevents customers from mistakenly believing the chat session has been interrupted and leaving the conversation due to prolonged waiting times, or from becoming anxious due to prolonged silence, thereby improving customer satisfaction.

[0051] In reference Figure 4 The above description illustrates an example where CPU 101, based on the query information sent from user terminal 500 to operator terminal 10, estimates the remaining time until operator terminal 10 sends a response information back to user terminal 500 in S400 to determine the baseline response time. Alternatively, or in addition to this example, the baseline response time can also be determined as follows: that is, the baseline response time can be adjusted based on whether the user is impatient or easygoing.

[0052] Regarding the waiting time from when a user sends a query from user terminal 500 to when they receive a response, a shorter baseline response time can be set for impatient users who prefer shorter response times. Alternatively, a longer baseline response time can be set for users who are more easygoing and not concerned about waiting time. For example, CPU 101 can obtain response speed preference information that determines a user's sensitivity to the length of the waiting time from when they send a query from user terminal 500 to when they receive a response, and adjust the baseline response time based on this response speed preference information.

[0053] As a method to obtain information on response speed preference, for example, in a customer database 400 ( Figure 1 The system stores response speed preference information and can read out the response speed preference information corresponding to the identified user based on the login information when using the live chat service. Regarding response speed preference information, customers can input it themselves during user registration before using the live chat service. For example, response speed preference information can be set and stored based on the user's answer to a question such as "Do you prefer a short waiting time from asking a question to receiving a reply?"

[0054] Alternatively, at the start of a chat session, or at an appropriate point during the session, such as after a user sends a question, the operator can ask the customer, "Would you prefer a short waiting time between your question and a response?" and set a response speed preference based on the user's reply. With this setting, users who normally respond slowly can adjust their response speed preference to be shorter in urgent situations, such as when a question indicates an anomaly has been detected in their vehicle. Furthermore, when an operator receives a question from a customer, if the question is text-based, the operator determines the urgency based on the wording; if it is voice-based, the operator determines the urgency based on the customer's tone and the content of the question. The operator can also set a response speed preference in these situations. Alternatively, information containing the communication content between the user terminal 500 and the operator terminal 10 can be sent to the external LLM 200 or the internal generative AI system 300. The external LLM 200 or the internal generative AI system 300 can determine whether the customer wants to answer in a short time, and the CPU 101 sets the response speed preference information based on the determination result.

[0055] According to this embodiment, the following effects can be achieved.

[0056] (1) The response time management program responds to the operator terminal 10 operated by the operator and the user terminal 500 operated by the customer. Figure 2The response time management program manages the time from when the operator terminal 10 sends the response information to the user terminal 500 after receiving a customer's query information via the user terminal 500, to when the operator accesses information sources 200, 300, and 350 to obtain the corresponding response information and creates a response information based on the information obtained from information sources 200, 300, and 350 (i.e., the original response), until the operation of sending the response information from the operator terminal 10 to the user terminal 500. Figure 2 The following processing is performed: The time elapsed from when operator terminal 10 receives the query information from user terminal 500 until it sends the response information back to user terminal 500 is detected. Figure 3 S312); Based on the amount of information in the response information sent from the operator terminal 10 to the user terminal 500, i.e., the amount of response information, a reference value, i.e., the reference response time, is determined as the time from when the operator terminal 10 receives the query information from the user terminal 500 until it sends a response information to the user terminal 500. Figure 3 S310); and in the case where it is detected that the operator has performed an operation to send a response message to the user terminal before the elapsed time reaches the reference response time ( Figure 3 S308: Yes), temporarily hold back the issuance of response information, and issue response information only after the baseline response time has been reached. Figure 3 (S314).

[0057] Therefore, when customers receive responses from operators via human chat, responding with a response time that corresponds to the amount of information in the response can suppress the negative feeling of responding to a person and improve customer satisfaction when using human chat.

[0058] (2) In the response time management procedure, the operator accesses information sources 200, 300, and 350 to obtain response information sent to the user terminal 500, and creates response information based on the information obtained from information sources 200, 300, and 350, i.e., the original response. The computer 100 also executes the following: if it detects that the operator who obtained the original response has performed an operation to obtain further information, it sends a message to the user terminal 500 informing the customer that it is preparing to send response information.

[0059] Therefore, when preparing to send a reply takes time, it can prevent customers from mistakenly believing that the chat session has been interrupted and leaving the conversation due to long waiting times, or from becoming anxious due to prolonged silence, thereby improving customer satisfaction.

[0060] (3) In the response time management procedure, the computer is also made to: when it is detected that the time elapsed before sending a response message to the user terminal 500 has reached the baseline response time, send a message to the user terminal 500 informing the customer that a response message is being prepared.

[0061] Therefore, even when the preparation time for sending a reply is long, it can prevent customers from mistakenly believing that the chat session has been interrupted and leaving the conversation, or from becoming anxious due to prolonged silence, thereby improving customer satisfaction.

[0062] (4) In the response time management procedure, the baseline response time is determined by estimating the amount of response information based on the query information sent from the user terminal 500 to the operator terminal 10.

[0063] Therefore, when preparing to send a reply takes time, customers can be informed of the estimated time until the reply is sent. This helps prevent customers from mistakenly believing the chat has been interrupted and leaving the conversation due to long waiting times, or from becoming anxious due to prolonged silence, thereby improving customer satisfaction.

[0064] (5) In the response time management procedure, the computer 100 further performs: obtaining response speed preference information that determines the length of the waiting time from when the customer asks a question to when the operator terminal 10 sends a response; and adjusting the baseline response time according to the response speed preference information.

[0065] Therefore, the timing of the response can be adjusted according to the customer's preference regarding waiting time, such as whether they prefer a shorter response or a more relaxed pace of communication, thus enabling a more relaxed and stress-free response tailored to the customer's personality.

[0066] (6) The response time management device responds to the operator terminal 10 operated by the operator and the user terminal 500 operated by the customer. Figure 2The response time management device includes: a time detection unit 118 that detects the elapsed time from when the operator terminal 10 receives a question from the user terminal 500 until it sends a response from the operator terminal 10 to the user terminal 500; and a response time management device that manages the time elapsed from when the operator terminal 10 receives the question from the user terminal 500 until it sends a response from the user terminal 500. The reference response time determination unit 114 determines a reference value, i.e., a reference response time, based on the amount of information in the response information sent from the operator terminal 10 to the user terminal 500, i.e., the amount of response information, as the reference value. The response information sending unit 120 temporarily holds the sending of response information to the user terminal 500 when it detects that the operator has performed the operation of sending response information to the user terminal 500 before the reference response time has elapsed, and then sends the response information after the reference response time has elapsed.

[0067] Therefore, when customers receive responses from operators via human chat, responding with a response time that corresponds to the amount of information in the response can suppress the negative feeling of responding to a person and improve customer satisfaction when using human chat.

[0068] (7) The human chat system has 500 user terminals for customer operation. Figure 2The operator terminal 10 and the response time management device 100 are operated by the operator terminal 10. The response time management device 100 manages the time from when the operator terminal 10 receives a question from the user terminal 500 through the user terminal 500, when the operator accesses information sources 200, 300, and 350 to obtain a response based on the information obtained from the information sources 200, 300, and 350 (i.e., the original response), until the operation of sending the response from the operator terminal 10 to the user terminal 500 is performed. The response time management device 100 includes: a time detection unit 118 that detects when the operator terminal 10 receives a question from the user terminal 500 through the user terminal 500. The system includes: a reference response time unit 114, which determines a reference value, i.e., a reference response time, based on the amount of information in the response information sent from the operator terminal 10 to the user terminal 500, i.e., the amount of response information; and a response information sending unit 120, which temporarily holds the sending of the response information after the reference response time has elapsed, if it detects that the operator has sent the response information to the user terminal 500 before the reference response time has elapsed.

[0069] Therefore, when customers receive responses from operators via human chat, responding with a response time that corresponds to the amount of information in the response can suppress the negative feeling of responding to a person and improve customer satisfaction when using human chat.

[0070] The above description is merely an example, and the present invention is not limited to the above embodiments and modifications as long as it does not impair the characteristics of the invention. One or more of the embodiments and modifications can be combined in any way, and modifications can also be combined with each other.

[0071] By employing this invention, when a customer receives a response from an operator via human chat, responding with a response time corresponding to the amount of information in the response can suppress the feeling of responding to a person and improve customer satisfaction when using human chat.

[0072] The present invention has been described above in conjunction with preferred embodiments, but those skilled in the art should understand that various modifications and changes can be made without departing from the scope of the claims.

Claims

1. A storage medium readable by a computer, storing a response time management program executable by a computer having a processor and a storage unit, the response time management program managing the time from when an operator's terminal receives a query from a customer's terminal via an operator's terminal, the operator accesses an information source to obtain an answer based on information from the information source (i.e., an original answer), until the operation of sending the answer from the operator's terminal to the user terminal is performed. The response time management program causes the computer to perform the following processes: The time elapsed from when the operator terminal receives the question information from the user terminal to when it sends a response information to the user terminal is detected. Based on the amount of information in the response information sent from the operator terminal to the user terminal (i.e., the response information volume), a baseline response time is determined, which is the time from when the operator terminal receives the query information from the user terminal until it sends the response information back to the user terminal; and If it is detected that the operator sent the response information to the user terminal before the elapsed time reaches the baseline response time, the sending of the response information is temporarily suspended, and the response information is sent again after the baseline response time is reached.

2. The storage medium according to claim 1, characterized in that, The response time management program also causes the computer to perform: If the operator who received the original response performs an operation to obtain further information, a message is sent to the user terminal informing the customer that the operator is preparing to issue the response information.

3. The storage medium according to claim 1, characterized in that, The response time management program also causes the computer to perform: If, before sending the response information to the user terminal, the system detects that the elapsed time has reached the baseline response time, it sends a message to the user terminal informing the customer that it is preparing to send the response information.

4. The storage medium according to any one of claims 1 to 3, characterized in that, The response time management program. In determining the baseline response time, the amount of response information is estimated based on the query information sent from the user terminal to the operator terminal.

5. The storage medium according to any one of claims 1 to 3, characterized in that, The response time management program also causes the computer to perform: Obtain response speed preference information that can determine the length of the waiting time from when the customer sends the question to when the operator terminal sends the response; as well as The baseline response time is adjusted based on the response speed preference information.

6. A response time management device, which manages the time from when an operator's terminal receives a question from a customer's user terminal via the user terminal, the operator accesses an information source to obtain an answer based on information from the information source (i.e., an original answer), until the operation of sending the answer from the operator's terminal to the user terminal is performed, characterized in that, have: The time detection unit (118) detects the elapsed time from when the operator terminal receives the question information from the user terminal to when it sends the answer information to the user terminal. The reference response time determination unit (114) determines a reference value, i.e., a reference response time, based on the amount of information in the response information sent from the operator terminal to the user terminal, i.e., the amount of response information. as well as The response information issuing unit (120) temporarily holds back sending the response information to the user terminal when it detects that the operator has performed the operation of sending the response information to the user terminal before the elapsed time reaches the reference response time, and then sends the response information after the reference response time is reached.

7. A chat system, comprising: The user terminal operated by the customer; Operator terminal operated by the operator; and A response time management device manages the time from when an operator's terminal receives a query from a customer's terminal via an operator's terminal, the operator accesses an information source to obtain a response based on the original response obtained from the information source, until the response is sent from the operator's terminal to the user terminal. The chat system is characterized by the following features: The response time management device includes: The time detection unit (118) detects the elapsed time from when the operator terminal receives the question information from the user terminal to when it sends the answer information to the user terminal. The reference response time determination unit (114) determines a reference value, i.e., a reference response time, based on the amount of information in the response information sent from the operator terminal to the user terminal, i.e., the amount of response information; and the reference response time is determined by the operator terminal from the moment it receives the query information from the user terminal until it sends the response information back to the user terminal. The response information issuing unit (120) temporarily holds back the issuance of the response information when it detects that the operator has issued the response information to the user terminal before the elapsed time reaches the reference response time, and then issues the response information after the reference response time has been reached.

Images