JP7804866B2 - Operator support system, operator support device, and operator support method - Google Patents

Description

The present invention relates to an operator support system, an operator support device, and an operator support method that support the work of operators who create customer interaction records based on the content of customer inquiries made over the phone.

At a call center, after an operator has finished answering a customer inquiry over the phone, they create a customer interaction record (ticket) that breaks down the customer's inquiry into categories according to a specified format. Creating this customer interaction record is time-consuming, so it is desirable to support the operator's work and reduce their burden.

A known technology for supporting such operator work and reducing their burden is to use AI to extract the customer's needs from the voice recognition results of the conversation between the customer and the operator, search for FAQs (questions and answers) corresponding to those needs, and present the found FAQs to the operator (see Patent Document 1).

Patent No. 6675788

Conventional technology allows operators to smoothly handle customer inquiries over the phone by presenting them with answers that correspond to the customer's needs, and the use of AI can reduce the operator's workload. However, conventional technology does not take into consideration item extraction, which extracts information corresponding to each item in a customer interaction record from the voice recognition results, and is therefore unable to reduce the operator's workload in creating customer interaction records.

On the other hand, it is possible to use generation AI to perform item extraction processing, which extracts information corresponding to each item in a customer interaction record from the voice recognition results. However, conversations between customers and operators may include casual conversation or other topics, and in such cases, the accuracy of the item extraction processing using generation AI may decrease, resulting in a problem of reduced quality in the customer interaction record.

The primary objective of the present invention is to provide an operator support system, operator support device, and operator support method that can support operators who create customer interaction records, thereby reducing the operator's workload and improving the quality of customer interaction records.

The operator support system of the present invention is an operator support system in which one or more processors execute processes to support the work of an operator who creates customer interaction records, and the processor instructs a speech recognition unit to execute a speech recognition process in predetermined time units , which converts the voices of the customer and the operator during customer interactions into text using speech recognition, obtains the speech recognition results for the predetermined time units , and in response to an operator's operation specifying a target range for the speech recognition results, instructs a text analysis unit to execute an item extraction process, which extracts a draft of the registration information for each item of the customer interaction record based on the speech recognition results included in the target range, obtains the item extraction results, displays a screen on the operator terminal that presents the item extraction results to the operator, and instructs a management unit to register the item extraction results as the customer interaction record, in response to an operator's operation after confirming the item extraction results displayed on the screen.

Furthermore, the operator support device of the present invention is an operator support device in which a processor executes a process to support the work of an operator who creates customer interaction records, and the processor instructs a speech recognition unit to execute a speech recognition process in predetermined time units , which converts the voices of the customer and the operator during customer interaction into text using speech recognition, thereby obtaining speech recognition results for the predetermined time units , and instructs a text analysis unit to execute an item extraction process to extract text corresponding to each item of the customer interaction record based on the speech recognition results included in the target range, in response to an operator's operation to specify a target range for the speech recognition results, thereby obtaining the item extraction results, instructs a GUI control unit to display a screen on an operator terminal that presents the item extraction results to the operator, and instructs a management unit to register the item extraction results as the customer interaction record, in response to an operator's operation to confirm the item extraction results displayed on the screen.

Furthermore, the operator support method of the present invention is an operator support method that causes one or more processors to execute processing to support the work of an operator who creates customer interaction records, and is configured to instruct a speech recognition unit to execute speech recognition processing in predetermined time units to convert the speech of the customer and operator during customer interaction into text using speech recognition, obtain speech recognition results for the predetermined time units , instruct a text analysis unit to execute item extraction processing to extract text corresponding to each item of the customer interaction record based on the speech recognition results included in the target range in response to an operator's operation to specify a target range for the speech recognition results, obtain the item extraction results, display a screen on the operator terminal that presents the item extraction results to the operator, and instruct a management unit to register the item extraction results as the customer interaction record in response to an operator's operation after confirming the item extraction results displayed on the screen.

This invention allows unnecessary parts, such as casual conversation during a call, to be removed from the speech recognition results before the item extraction process is performed, thereby improving the accuracy of the item extraction results with simple operations by the operator. This reduces the operator's workload and improves the quality of the customer interaction records created based on the item extraction results.

Overall configuration of an operator support system according to the present embodiment. A block diagram showing the schematic configuration of a main control unit and a GUI control unit according to the present embodiment. A sequence diagram showing the procedure of processing performed in the operator support system according to the present embodiment. A sequence diagram showing the procedure of processing performed in the operator support system according to the present embodiment. FIG. 10 is an explanatory diagram showing an example of a prompt input to a text analysis unit and an answer output from the text analysis unit according to the present embodiment. FIG. 10 is an explanatory diagram showing a call screen displayed on an operator terminal according to the embodiment; FIG. 10 is an explanatory diagram showing a confirmation screen displayed on an operator terminal according to the embodiment; FIG. 10 is an explanatory diagram showing a confirmation screen displayed on an operator terminal according to the embodiment; FIG. 10 is an explanatory diagram showing a registration screen displayed on an operator terminal according to the embodiment; FIG. 10 is an explanatory diagram showing a registration screen displayed on an operator terminal according to the embodiment; FIG. 10 is an explanatory diagram showing a call screen displayed on an operator terminal according to a first modification; FIG. 10 is a sequence diagram showing a procedure of processing performed in the operator support system according to the second modification; FIG. 10 is an explanatory diagram showing a call screen displayed on an operator terminal according to a second modification; FIG. 11 is an explanatory diagram showing a call screen displayed on an operator terminal according to a third modified example. FIG. 13 is an explanatory diagram showing a call screen displayed on an operator terminal according to a fourth modified example. FIG. 13 is an explanatory diagram showing a call screen displayed on an operator terminal according to a fifth modified example. FIG. 13 is an explanatory diagram showing a call screen displayed on an operator terminal according to a sixth modified example. FIG. 13 is a sequence diagram showing the procedure of processing performed in the operator support system according to the sixth modification. FIG. 13 is an explanatory diagram showing a confirmation screen displayed on an operator terminal according to a seventh modification; FIG. 13 is a sequence diagram showing the procedure of processing performed in the operator support system according to the seventh modification.

A first invention made to solve the above problem is an operator support system in which one or more processors execute processes to support the work of an operator who creates customer interaction records, the processor instructing a speech recognition unit to execute a speech recognition process in predetermined time units , which converts the voices of the customer and the operator during customer interactions into text using speech recognition, obtaining speech recognition results for the predetermined time units , and instructing a text analysis unit to execute an item extraction process to extract a draft of registration information for each item of the customer interaction record based on the speech recognition results included in the target range, in response to an operator's operation to specify a target range for the speech recognition results, obtaining the item extraction results, displaying a screen on an operator terminal that presents the item extraction results to the operator, and instructing a management unit to register the item extraction results as the customer interaction record, in response to an operator's operation to confirm the item extraction results displayed on the screen.

This allows the item extraction process to be performed while excluding unnecessary parts, such as casual conversations, from the speech recognition results. This allows the operator to easily perform simple operations to improve the accuracy of the item extraction results. This reduces the operator's workload and improves the quality of customer interaction records created based on the item extraction results. Furthermore, because the speech recognition process is performed even during a call, it reduces the waiting time until the speech recognition results containing the entire call are presented to the user after the call ends.

In addition, in a second invention, when a single inquiry call includes multiple topics, the processor instructs the text analysis unit to perform the item extraction process for each of the multiple topics in response to an operator's operation to separately select a range related to each of the multiple topics, and obtains the item extraction results for each of the multiple topics.

This means that even if a single inquiry call covers multiple topics, the operator can create separate customer interaction records for each of the multiple topics with a simple operation.

In addition, a third invention is configured such that the processor displays a screen on the operator terminal that presents the speech recognition result to an operator, and in response to an operation by the operator that specifies the target range for the speech recognition result displayed on the screen, instructs the text analysis unit to execute the item extraction process based on the speech recognition result included in the target range, thereby obtaining the item extraction result.

This allows the operator to visually check the speech recognition results and appropriately specify the target range. The screen on which the operator specifies the target range for the speech recognition results may be the screen displayed after the call ends, or it may be the screen displayed during the call.

In addition, a fourth invention is configured such that the processor, in response to an operator's operation to specify the start and end points of the target range during a call, instructs the text analysis unit to execute the item extraction process based on the speech recognition result for the speech included in the target range, and obtains the item extraction result.

This allows the operator to specify the start and end points of the target range during a call. In this case, an operation unit (e.g., a button) for specifying the start and end points of the target range may be displayed on the screen of the operator terminal, and the operator may specify the start and end points by operating the operation unit.

In addition, a fifth invention is configured such that the processor instructs the call unit to guide the customer to explain the outline of the inquiry before the operator responds, acquires the audio of the inquiry spoken by the customer, instructs the speech recognition unit to perform the speech recognition process on the audio of the inquiry, acquires the text of the inquiry as the speech recognition result, instructs the text analysis unit to analyze the text of the inquiry, acquires tag information and a summary of the inquiry, selects an operator that matches the inquiry based on the tag information, instructs the call unit to connect the customer's call to the selected operator, and displays a screen on the operator terminal that presents the summary to the operator.

This allows the appropriate operator to answer the call based on the customer's inquiry, resulting in smoother response to customer inquiries and increased customer satisfaction.

A sixth aspect of the present invention is an operator support system in which one or more processors execute a process for supporting an operator's work of creating a customer interaction record, the processor instructing a speech recognition unit to execute a speech recognition process for converting the speech of a customer and an operator into text by speech recognition, and acquiring a speech recognition result, and instructing a text analysis unit to execute an item extraction process for extracting a draft of registration information for each item of the customer interaction record based on the speech recognition result included in the target range in response to an operation by an operator to specify a target range for the speech recognition result, and acquiring the item extraction result, and displaying a screen on an operator terminal to present the item extraction result to the operator, and the operator's operation after confirming the item extraction result displayed on the screen. and instructs the management unit to register the item extraction result as the customer interaction record in accordance with the operation, and further causes a screen for presenting the speech recognition result to be displayed on the operator terminal during the call, and acquires the corrected speech recognition result in accordance with a user operation to correct the speech recognition result displayed on the screen, and instructs the text analysis unit to execute the item extraction process based on the corrected speech recognition result to acquire the item extraction result, and furthermore, when a user operation to switch to an edit mode is performed on the screen for presenting the speech recognition result to the operator, the system switches to the edit mode, accepts the user operation to correct the speech recognition result, and temporarily stops reflecting the speech recognition result on the screen.

This allows the operator to correct inappropriate parts of the speech recognition results before the end of the call, so that the execution of the item extraction process can be instructed early and the customer interaction record can be created quickly.In addition, while the operator is manually correcting the speech recognition results, the corrections made by the operator can be prevented from being overwritten and erased by the next speech recognition result output.

Further, a seventh invention is an operator support system in which one or more processors execute a process to support the work of an operator who creates customer interaction records, wherein the processor instructs a speech recognition unit to execute a speech recognition process that converts the speech of the customer and the operator into text by speech recognition, thereby obtaining a speech recognition result; in response to an operator operation that specifies a target range for the speech recognition result, instructs a text analysis unit to execute an item extraction process that extracts a draft of registration information for each item of the customer interaction record based on the speech recognition results included in the target range, thereby obtaining the item extraction result; causes an operator terminal to display a screen that presents the item extraction result to the operator; instructs a management unit to register the item extraction result as the customer interaction record, according to an operator operation that confirms the item extraction result displayed on the screen; further causes the operator terminal to display a screen that allows the operator to input notes during a call, thereby obtaining the notes input by the operator; and instructs the text analysis unit to execute the item extraction process based on the speech recognition result and the notes, thereby obtaining the item extraction result that reflects the contents of the notes.

This allows you to obtain item extraction results that reflect the contents of your notes, thereby improving the quality of customer interaction records registered based on the item extraction results.

An eighth invention is an operator support system in which one or more processors execute processes to support the work of an operator who creates customer interaction records, wherein the processor instructs a speech recognition unit to execute a speech recognition process that converts the speech of the customer and the operator into text by speech recognition, thereby obtaining a speech recognition result; in response to an operator operation that specifies a target range for the speech recognition result, instructs a text analysis unit to execute an item extraction process that extracts a draft of registration information for each item of the customer interaction record based on the speech recognition results included in the target range, thereby obtaining the item extraction result; displays a screen on an operator terminal that presents the item extraction result to the operator; instructs a management unit to register the item extraction result as the customer interaction record, according to an operator operation that confirms the item extraction result displayed on the screen; and further instructs the text analysis unit to execute a similar case search process that searches for past cases similar to the content of the current customer's inquiry, thereby obtaining information on similar cases as a search result, and displaying the information on the similar cases on the operator terminal.

This allows operators to view past cases similar to the customer's current inquiry. This allows for smoother phone responses to customer inquiries and increases customer satisfaction.

A ninth invention is an operator support system in which one or more processors execute a process to support the work of an operator who creates customer interaction records, wherein the processor instructs a speech recognition unit to execute a speech recognition process to convert the speech of a customer and an operator into text by speech recognition, thereby obtaining a speech recognition result; in response to an operator operation to specify a target range for the speech recognition result, instructs a text analysis unit to execute an item extraction process to extract a draft of registration information for each item of the customer interaction record based on the speech recognition results included in the target range, thereby obtaining the item extraction result; displays a screen on an operator terminal to present the item extraction result to the operator; instructs a management unit to register the item extraction result as the customer interaction record, according to an operator operation after confirming the item extraction result displayed on the screen; further instructs the text analysis unit to execute a process to proofread the speech recognition result, thereby obtaining a proofread speech recognition result; and instructs the text analysis unit to execute the item extraction process on the proofread speech recognition result, thereby obtaining the item extraction result.

This allows the item extraction process to be performed based on the calibrated speech recognition results, thereby improving the accuracy of customer interaction records. The calibration process may be performed in response to an operator's instruction to perform it, or it may be performed independently of an operator's instruction.

Further, a tenth invention is an operator support device in which a processor executes a process to support the work of an operator who creates customer interaction records, wherein the processor instructs a speech recognition unit to execute a speech recognition process in predetermined time units , which converts the speech of the customer and the operator during customer interaction into text using speech recognition, thereby obtaining speech recognition results for the predetermined time units , and instructs a text analysis unit to execute an item extraction process to extract text corresponding to each item of the customer interaction record based on the speech recognition results included in the target range in response to an operator's operation to specify a target range for the speech recognition results, thereby obtaining the item extraction results, instructs a GUI control unit to display a screen on an operator terminal that presents the item extraction results to the operator, and instructs a management unit to register the item extraction results as the customer interaction record in response to an operator's operation to confirm the item extraction results displayed on the screen.

According to this, as in the first invention, unnecessary parts such as chatter during a call can be excluded from the speech recognition results when performing the item extraction process, so the accuracy of the item extraction results can be improved with a simple operation by the operator. This reduces the operator's workload and improves the quality of the customer interaction record created based on the item extraction results. Furthermore, because the speech recognition process is performed even during a call, the waiting time until the speech recognition results containing the entire call are presented to the user after the call ends can be shortened.

An eleventh invention is an operator support method that causes one or more processors to execute a process to support the work of an operator who creates customer interaction records, the method comprising the steps of instructing a speech recognition unit to execute a speech recognition process that converts the speech of a customer and an operator during customer interaction into text using speech recognition in predetermined time units , obtaining speech recognition results for the predetermined time units , instructing a text analysis unit to execute an item extraction process that extracts text corresponding to each item of the customer interaction record based on the speech recognition results included in the target range in response to an operator's operation to specify a target range for the speech recognition results, obtaining the item extraction results, displaying a screen on an operator terminal that presents the item extraction results to the operator, and instructing a management unit to register the item extraction results as the customer interaction record in response to an operator's operation to confirm the item extraction results displayed on the screen.

According to this, as in the first invention, unnecessary parts such as chatter during a call can be excluded from the speech recognition results when performing the item extraction process, so the accuracy of the item extraction results can be improved with a simple operation by the operator. This reduces the operator's workload and improves the quality of the customer interaction record created based on the item extraction results. Furthermore, because the speech recognition process is performed even during a call, the waiting time until the speech recognition results containing the entire call are presented to the user after the call ends can be shortened.

Embodiments of the present invention will be described below with reference to the drawings.

Figure 1 is a diagram showing the overall configuration of the operator support system according to this embodiment.

The operator support system supports the work of call center operators who create customer interaction records (tickets) based on the inquiries made by customers over the phone. Customer interaction records are records in which the content of customer inquiries is divided into items according to a specified format.

The operator support system comprises a call unit 1, a voice interception unit 2, a main control unit 3 (operator support device), a voice recognition unit 4, a text analysis unit 5, a management unit 6, a GUI control unit 7, and an operator terminal 8.

The call unit 1 relays audio transmission and reception between the telephone 9 used by the customer and the operator terminal 8 used by the operator, enabling calls between the customer and the operator. The call unit 1 may be implemented using a known cloud service, or may be implemented using a telephone system that builds "PBX (main unit)" functionality on the cloud and allows business phone functions to be used over the Internet.

The audio interception unit 2 intercepts (collects) the audio of the call between the customer and the operator conducted by the call unit 1. The audio interception unit 2 may be implemented using a known cloud service.

The voice recognition unit 4 converts the voice of the call acquired by the voice interception unit 2 into text using voice recognition. The voice recognition unit 4 may be implemented using a known cloud service.

The text analysis unit 5 performs an item extraction process on the speech recognition results output from the speech recognition unit 4. In the item extraction process, a character string that serves as a draft (original proposal) of the registration information for each item in the customer interaction record is extracted from the speech recognition results using generation AI. The text analysis unit 5 also performs a summary generation process on the speech recognition results. In the summary generation process, a summary of the inquiry content is generated using generation AI. The text analysis unit 5 may be implemented using a cloud service such as Google Bard (registered trademark), Microsoft Bing AI (registered trademark), or ChatGPT.

The management unit 6 manages various data such as customer interaction records using a database. The management unit 6 may be implemented, for example, as a tool for managing customer contact history known as a CRM tool, or as a general relational database. It may also be implemented as a known incident management tool or a cloud service that includes web server functionality.

The main control unit 3 controls each part of the operator support system, generates registration information for each item in the customer interaction record from the audio of the call between the customer and the operator, and performs the necessary processing to register that registration information in the database of the management unit 6.

The GUI control unit 7 generates display information (e.g., HTML data) for various screens to be displayed on the operator terminal 8, and displays the various screens on the operator terminal 8. The GUI control unit 7 also acquires operation information corresponding to the operator's operations performed on the operator terminal 8. The GUI control unit 7 may be realized, for example, by a CRM tool or incident management tool, which is a software package including web server functionality, or by a cloud service including web server functionality.

The operator terminal 8 is configured as a PC, smartphone, tablet terminal, or the like. The operator terminal 8 is equipped with a display device 11 such as an LCD panel, an input device 12 such as a touch panel, keyboard, or mouse, a microphone 13, and a speaker 14. The microphone 13 and speaker 14 may also be headsets.

At the operator terminal 8, the operator can use the microphone 13 and speaker 14 to talk to the customer. Furthermore, the operator terminal 8 uses an application such as a browser to display various screens on the display device 11 for the operator to view, based on display information sent from the GUI control unit 7. Furthermore, the operator terminal 8 detects screen operations performed by the operator using the input device 12, and sends the operation information to the GUI control unit 7.

Next, we will explain the general configuration of the main control unit 3 and GUI control unit 7 according to this embodiment. Figure 2 is a block diagram showing the general configuration of the main control unit 3 and GUI control unit 7.

The main control unit 3 includes a communication unit 21, a memory unit 22, and a processor 23.

The communication unit 21 communicates with the voice interception unit 2, voice recognition unit 4, text analysis unit 5, management unit 6, and GUI control unit 7.

The memory unit 22 stores programs executed by the processor 23, etc.

The processor 23 performs various processes by executing programs stored in the storage unit 22. In this embodiment, the processor 23 performs processes such as voice recognition control, item extraction control, and screen control.

In the voice recognition control process, the processor 23 instructs the voice recognition unit 4 to execute a voice recognition process that converts the voice of the conversation between the customer and the operator into text through voice recognition, and obtains the voice recognition results.

In the item extraction control process, the processor 23 instructs the text analysis unit 5 to execute an item extraction process that uses the generation AI to extract a draft (original plan) of the registration information for each item in the customer interaction record based on the speech recognition results, thereby obtaining the item extraction results.

In the screen control process, the processor 23 instructs the GUI control unit 7 to display on the operator terminal 8 a call screen 51 (see Figure 6) that presents the operator with the voice recognition results of the call voice during the call, a confirmation screen 61 (see Figures 7 and 8) that allows the operator to confirm the voice recognition results after the call ends, and a registration screen 71 (see Figures 9 and 10) that allows the operator to register each item in the customer interaction record.

The GUI control unit 7 includes a communication unit 31, a memory unit 32, and a processor 33.

The communication unit 31 communicates with the main control unit 3.

The memory unit 32 stores programs executed by the processor 33, etc.

The processor 33 performs various processes by executing programs stored in the storage unit 32. In this embodiment, the processor 33 performs processes such as generating screen information and acquiring operation information.

In the screen information generation process, the processor 33 generates display information for the screen to be displayed on the operator terminal 8. In this embodiment, display information for the call screen 51 (see Figure 6), confirmation screen 61 (see Figures 7 and 8), and registration screen 71 (see Figures 9 and 10) is generated.

In the operation information acquisition process, the processor 33 acquires operation information generated by detecting an operator's operation on the operator terminal 8.

Next, we will explain the processing procedures performed by the operator support system according to this embodiment. Figures 3 and 4 are sequence diagrams showing the processing procedures performed by the operator support system.

First, the GUI control unit 7 displays a login screen on the operator terminal 8. The operator performs a login operation on the operator terminal 8, and when the login is successful, the GUI control unit 7 displays a call screen 51 (see Figure 6) on the operator terminal 8. At this time, login information such as the operator ID is sent to the main control unit 3, and then sent to the management unit 6 where it is registered in the database.

Next, when the call unit 1 detects an incoming call from a customer making an inquiry, it assigns a contact ID to the call and sends the contact ID, the customer's telephone number, and the operator ID to the main control unit 3. The main control unit 3 then sends the contact ID, telephone number, and operator ID to the management unit 6. The management unit 6 then registers the contact ID, telephone number, and operator ID in a database.

In addition, when the call unit 1 detects an incoming call from a customer making an inquiry and the contact ID, customer telephone number, and operator ID are registered in the management unit 6 database, a new customer interaction record (ticket) may also be registered. In this case, items other than the contact ID, telephone number, and operator ID may be left blank, and items related to these items, such as the store name that is apparent from the customer's telephone number or the operator name that is apparent from the operator ID, may also be registered.

Furthermore, when a new customer interaction record (ticket) is registered in the management unit 6, the operator may be notified of this. Specifically, when the main control unit 3 receives a response from the management unit 6 indicating that the new customer interaction record has been successfully registered, it may request external software to send a notification that the new customer interaction record has been registered to the operator's contact information (operator terminal 8) associated with the operator ID. Examples of this external software include email and communication tools with business chat functions.

In addition, the notification that a new customer interaction record has been registered may include access information for displaying the customer interaction record viewing screen on the operator terminal 8. An example of access information is a URL. The operator can view the contents of the newly registered customer interaction record by performing an operation to open the customer interaction record viewing screen on the operator terminal 8.

Next, when a call between the customer and the operator is initiated in the call unit 1, the voice interception unit 2 collects the voice of the call from the call unit 1. Next, the voice recognition unit 4 converts the voice of the call into text using voice recognition and sends the voice recognition results to the main control unit 3. Next, the main control unit 3 sends the voice recognition results to the GUI control unit 7. Next, the GUI control unit 7 displays the voice recognition results on the call screen 51. At this time, voice recognition is performed on the voice of the call in units of a predetermined time (for example, 30 seconds), and the voice recognition results for the predetermined time units are displayed sequentially on the call screen 51.

Next, when the call unit 1 detects the end of the call between the customer and the operator (end of call), it notifies the main control unit 3 of the end of the call. Next, the main control unit 3 instructs the GUI control unit 7 to transition to a confirmation screen 61 (see Figures 7 and 8), and the GUI control unit 7 displays the confirmation screen 61 on the operator terminal 8.

Next, as shown in FIG. 4, when the operator specifies the target range of the item extraction process on the confirmation screen 61 displayed on the operator terminal 8 and performs an operation to instruct execution of the item extraction process, the GUI control unit 7 sends an instruction to execute the item extraction process to the main control unit 3. At this time, information on the target range specified by the operator is also sent to the main control unit 3. Next, the main control unit 3 instructs the GUI control unit 7 to transition to a registration screen 71 (see FIGS. 9 and 10), and the GUI control unit 7 displays the registration screen 71 on the operator terminal 8.

Next, the main control unit 3 obtains the customer number and customer name based on the customer's telephone number and sends the customer number and customer name to the GUI control unit 7. The GUI control unit 7 then displays the customer number and customer name on the registration screen 71.

The main control unit 3 also sends a prompt to the text analysis unit 5 instructing it to perform item extraction processing. At this time, the speech recognition results included in the target range of the item extraction processing are extracted, and the speech recognition results included in that target range are added to the prompt. Next, the text analysis unit 5 performs item extraction processing based on the speech recognition results included in the target range, and sends the item-specific extraction results to the main control unit 3. Next, the main control unit 3 sends the item extraction results to the GUI control unit 7. Next, the GUI control unit 7 reflects the item extraction results on the registration screen 71.

Next, when the operator performs an operation to instruct registration on the registration screen 71 displayed on the operator terminal 8, the GUI control unit 7 sends a registration instruction to the main control unit 3. Next, the main control unit 3 sends the item extraction results to the management unit 6 as registration information. Next, the management unit 6 registers the item extraction results as registration information in the database. At this time, if the operator performs an operation on the operator terminal 8 to modify the item extraction results displayed on the registration screen 71, the modified item extraction results are sent to the management unit 6 as registration information and registered in the database.

Next, we will explain the prompts input to the text analysis unit 5 according to this embodiment and the answers output from the text analysis unit 5. Figure 5 is an explanatory diagram showing an example of a prompt and its answer.

The main control unit 3 instructs the text analysis unit 5 to execute an item extraction process based on the speech recognition results output from the speech recognition unit 4. At this time, for example, a prompt (instruction sentence) such as that shown in Figure 5(A) is generated by the main control unit 3, and this prompt is input to the text analysis unit 5, which then outputs the answer shown in Figure 5(B).

In the example shown in Figure 5(A), the prompt contains questions about the customer representative, reception category, causing device, inquiry content, failure content, power indicator, recovery possibility, recovery method, and result summary. In response to this prompt, a response is output containing item extraction results for each item specified in the prompt, as shown in Figure 5(B).

The example shown in Figure 5(A) is a prompt related to the first of two topics included in one call, and the speech recognition results related to the first topic specified by the operator on the confirmation screen 61 (see Figure 7) are added to the prompt. For the second topic, a prompt is generated separately and input to the text analysis unit 5.

Next, we will explain the call screen 51 displayed on the operator terminal 8 according to this embodiment. Figure 6 is an explanatory diagram showing the call screen 51.

When login is successful, the operator terminal 8 displays the call screen 51 shown in Figure 6. The call screen 51 has a call start button 52. When the operator is notified of an incoming call with an inquiry from a customer and operates the call start button 52, a call between the customer and the operator begins. When the call between the customer and the operator begins, the voice recognition unit 4 performs a voice recognition process to convert the voice of the call into text.

The call screen 51 also has a voice recognition result display unit 53. The voice recognition result display unit 53 displays the voice recognition results output from the voice recognition unit 4. In this embodiment, voice recognition is performed on the voice of the call in predetermined time units, and the voice recognition results for the predetermined time units are displayed sequentially on the call screen 51. This allows the status of voice recognition by the voice recognition unit 4 to be displayed on the call screen 51, allowing the operator to confirm that the voice recognition process is being performed normally.

Next, we will explain the confirmation screen 61 displayed on the operator terminal 8 according to this embodiment. Figures 7 and 8 are explanatory diagrams showing the confirmation screen 61.

When the call between the customer and the operator ends, the operator terminal 8 displays a confirmation screen 61 as shown in Figure 7. Similar to the call screen 51 (see Figure 6), the confirmation screen 61 has a voice recognition result display section 62. The voice recognition result display section 62 displays the entire voice recognition result output from the voice recognition section 4. By viewing the confirmation screen 61, the operator can confirm whether the voice recognition result is appropriate.

In addition, the speech recognition result display unit 62 allows the operator to specify the target range for item extraction processing for the speech recognition results by performing a predetermined operation (e.g., a drag operation). In the example shown in Figure 7, the target range is specified to include the portion of the conversation related to the customer representative, but since the customer representative is manually entered by the operator, the target range may be specified to exclude the portion of the conversation related to the customer representative.

The confirmation screen 61 also has an item extraction button 63. When the operator operates the item extraction button 63, the text analysis unit 5 executes item extraction processing for the speech recognition results displayed on the confirmation screen 61 that fall within the target range specified by the operator.

When responding to a customer inquiry over the phone, the conversation may change to another topic midway through the call, and a single inquiry call may include multiple topics. In this case, separate customer interaction records are created for each of the multiple topics. First, the range of speech recognition results related to each of the multiple topics is selected separately, and item extraction processing for each of the multiple topics is executed separately in the text analysis unit 5, resulting in item extraction results for each of the multiple topics.

In the examples shown in Figures 7 and 8, two topics are included in one call, and ranges related to each topic are selected twice. In the example shown in Figure 7, the range related to the first of the two topics is selected. In the example shown in Figure 8, the range related to the second of the two topics is selected.

In this way, in this embodiment, the operator can specify the required range from the speech recognition results, so unnecessary parts of the speech recognition results, such as casual conversation, can be excluded from the target range of the item extraction process, and the accuracy of the item extraction results can be improved with a simple operation by the operator. Furthermore, when a single inquiry call contains multiple topics, the different topics contained in the speech recognition results can be separated, so that separate customer interaction records for each of the multiple topics can be created with a simple operation by the operator. Furthermore, because the number of characters in the speech recognition results is reduced, the load on the text analysis unit 5 can be reduced and the wait time until the item extraction results are output can be shortened.

In the examples shown in Figures 7 and 8, the operator specifies the target range for the item extraction process on the confirmation screen 61 that is displayed after the call ends, but it is also possible for the operator to specify the target range for the item extraction process on the call screen 51 (see Figure 6) that is displayed during the call.

Furthermore, in this embodiment, when a single inquiry call includes multiple topics, the operator selects the range of speech recognition results for each of the multiple topics separately, and item extraction processing for each of the multiple topics is performed separately. However, the processing for extracting the range for each of the multiple topics from the speech recognition results for a single inquiry call may also be performed using generation AI in the text analysis unit 5.

In addition, in the examples shown in Figures 7 and 8, a continuous string of characters is specified as the target range for the item extraction process in the voice recognition result display section 62 of the confirmation screen 61, but a discontinuous string of characters may also be specified.

Next, we will explain the registration screen 71 displayed on the operator terminal 8 according to this embodiment. Figures 9 and 10 are explanatory diagrams showing the registration screen 71.

When the operator operates the item extraction button 63 on the confirmation screen 61 (see Figures 7 and 8), the text analysis unit 5 is instructed to execute the item extraction process. Once the item extraction results are output, the registration screen 71 shown in Figures 9 and 10 is displayed on the operator terminal 8 based on the item extraction results.

The registration screen 71 has input fields 72, 73, and 74 for each item in the customer interaction record, specifically, the operator ID, customer number, customer name, customer representative, reception category, causing device, inquiry details, fault details, power indicator, recovery possibility, recovery method, and result summary.

The operator ID input field 72 pre-displays the operator ID entered by the operator when logging in. Note that the operator can manually edit the displayed character string in the input field 72 using the input device 12.

The customer number and customer name input fields 73 are pre-displayed with the customer number and customer name calculated from the customer's telephone number. Note that the operator can manually edit the displayed text in the input fields 73 using the input device 12.

The item extraction results output from the text analysis unit 5 are reflected in the input fields 74 for the following items: customer representative, reception category, causing device, inquiry content, failure content, power lamp, recovery possibility, recovery method, and result summary. The item extraction results serve as a draft (original plan) of the registration information for each item in the customer interaction record, and if the item extraction results are inappropriate, the operator can make corrections in the input fields 74. Note that text (summary) is entered in the input fields 74 for the inquiry content, failure content, recovery method, and result summary.

In addition, the input field 74 allows the operator to correct any typos that occur during voice recognition. In this case, the correction of the typos is automatically reflected in the voice recognition results registered in the database by the management unit 6.

The registration screen 71 has a registration button 75. When the operator operates the registration button 75, the character strings displayed in the input fields 72, 73, and 74 for each item on the registration screen 71 are registered as registration information for each item in the customer interaction record.

If multiple topics are included in a single inquiry call, one customer interaction record (ticket) is created for each topic. In this case, the operator specifies the range corresponding to each of the multiple topics separately as the target range for item extraction, and a separate registration screen 71 is displayed for each topic, and a separate customer interaction record is created for each topic.

The example shown in Figure 9 is a case where the range related to the first of two topics is selected on the confirmation screen 61 as shown in Figure 7 as the target range for the item extraction process. The example shown in Figure 10 is a case where the range related to the second of two topics is selected on the confirmation screen 61 as shown in Figure 8 as the target range for the item extraction process.

Furthermore, when the target range of the item extraction process is specified for each topic from the speech recognition results for a single inquiry call and a registration screen 71 is displayed for each topic, a tab 76 for each topic is displayed on the confirmation screen 61. The operator can switch between registration screens 71 for each topic by operating the tab 76. Note that the characters displayed on the tab 76 (tab title) may also be extracted from the speech recognition results using generation AI in the text analysis unit 5.

The main control unit 3 may identify the location (e.g., a store) related to the inquiry based on the customer's telephone number, identify the equipment previously delivered to that location, i.e., the equipment that is the subject of the inquiry, and select a dictionary to be used for voice recognition depending on the type of target equipment. This allows a dictionary specialized for the target equipment to be used, thereby improving the accuracy of the voice recognition results. Furthermore, the prompt (instruction sentence) used to instruct the text analysis unit 5 to execute item extraction processing using generation AI may be changed depending on the type of target equipment. This improves the accuracy of the item extraction results.

(First Modification)
Next, a first modified example will be described. It should be noted that the points not specifically mentioned here are the same as those in the above embodiment. Fig. 11 is an explanatory diagram showing a call screen 51 displayed on the operator terminal 8 according to the first modified example.

In this variant, the operator can specify the scope of the item extraction process for the audio of the call with the customer while they are in the middle of a call.

In this modified example, when login is successful on the operator terminal 8, the call screen 81 shown in FIG. 11 is displayed. The call screen 81 is provided with a call start button 82, a start time button 83, an end time button 84, and an item extraction button 85.

When the operator operates the call start button 82, a call between the customer and the operator begins.

When the operator operates the start button 83, the time of operation is set as the start point of the target range of the item extraction process. When the operator operates the end button 84, the time of operation is set as the end point of the target range of the item extraction process.

When the operator operates the item extraction button 85, speech recognition of the call voice included in the target range is performed by the speech recognition unit 4, and item extraction processing based on the speech recognition results is performed by the text analysis unit 5. Once the inquiry conversation ends and casual conversation begins, the operator can operate the item extraction button 85 to instruct the execution of item extraction processing even if the call has not ended.

Furthermore, unlike the call screen 51 (see Figure 6) in the above embodiment, the call screen 81 in this modified example does not display the voice recognition results, but the voice recognition results may also be displayed on the call screen 81. Furthermore, although a screen equivalent to the confirmation screen 61 (see Figures 7 and 8) in the above embodiment is not displayed in this modified example, a confirmation screen may be displayed so that the operator can confirm the voice recognition results.

(Second Modification)
Next, a second modified example will be described. It should be noted that the points not specifically mentioned here are the same as those in the above-described embodiment. Fig. 12 is a sequence diagram showing the procedure of processing performed in the operator support system according to the second modified example. Fig. 13 is an explanatory diagram showing a call screen 51 displayed on the operator terminal 8 according to the second modified example.

In this variant, the main control unit 3 selects an appropriate operator depending on the customer making the inquiry and the content of the inquiry, and instructs that operator to handle the call. First, when an incoming call from a customer is detected, before the call is assigned to an operator, an automated voice guide prompts the customer to explain the outline of their inquiry, and acquires the voice of the inquiry spoken by the customer. Next, speech recognition is performed on the voice of the inquiry content to acquire the text of the inquiry. Next, using generation AI, the information necessary to select an operator is acquired from the text of the inquiry content, and an operator who can appropriately handle the customer's inquiry is determined.

Specifically, as shown in FIG. 12, as in the previous embodiment, when the call unit 1 detects an incoming call from a customer making an inquiry, the contact ID, customer telephone number, and operator ID are sent to the management unit 6 via the main control unit 3 and registered in the database.

Next, the call unit 1 provides automated voice guidance. Specifically, a voice guidance prompting the customer to explain the outline of their inquiry is output. For example, a synthesized voice guidance saying, "Please tell us your problem." The customer responds to the voice guidance by uttering the outline of their inquiry. At this time, the voice intercepting unit 2 collects the voice of the inquiry spoken by the customer and transmits the voice of the inquiry to the main control unit 3. Next, the main control unit 3 transmits the voice of the inquiry to the voice recognition unit 4. Next, the voice recognition unit 4 converts the voice of the inquiry into text using voice recognition, and transmits the voice recognition result (text of the inquiry) to the main control unit 3.

Next, the main control unit 3 sends a prompt to the text analysis unit 5 instructing tagging and summary generation. The prompt is accompanied by the speech recognition results for the voice of the query (text of the query).

Next, the text analysis unit 5 uses the generation AI to obtain tag information (inquiry attribute information) that identifies the inquiry content from the speech recognition result (inquiry text) (tagging process). The tag information is information necessary to determine an operator who can appropriately respond to the customer's inquiry, and may be, for example, a keyword that succinctly describes the customer who made the inquiry or the nature of the problem. The text analysis unit 5 also uses the generation AI to generate a summary of the inquiry content from the speech recognition result (inquiry text) (summary generation process). The text analysis unit 5 then sends the tag information and summary to the main control unit 3.

The main control unit 3 then selects an operator that matches the inquiry content based on the tag information (responder selection process) and sends the operator selection information to the call unit 1. At this time, the main control unit 3 obtains association information that associates the tag information with the operator from the management unit 6, and uses that association information to select an operator from the tag information. The association information is set based on the operator's areas of expertise, the customers the operator is responsible for, etc.

Next, the call unit 1 performs processing to connect the customer's call to the selected operator. The main control unit 3 also sends a summary of the inquiry to the GUI control unit 7. The GUI control unit 7 then displays the summary of the inquiry on the call screen 51. As shown in FIG. 13, the call screen 51 has an inquiry content display field 59. The inquiry content display field 59 displays the summary of the inquiry.

Next, the call between the customer and the operator begins in call unit 1. The rest of the process is the same as in the previous embodiment.

In this way, this modified example allows the operator who can respond appropriately to the customer's inquiry to answer the phone, which allows the phone call to proceed smoothly. Furthermore, because a summary of the inquiry is presented to the operator, there is no need for the operator to explain the outline of the inquiry to the customer again when they begin answering the phone. This increases customer satisfaction.

(Third Modification)
Next, a third modified example will be described. It should be noted that the points not specifically mentioned here are the same as those in the above-described embodiment. Fig. 14 is an explanatory diagram showing a call screen 51 displayed on the operator terminal 8 according to the third modified example.

In this modified example, the operator can edit the speech recognition results displayed on the operator terminal 8. The text analysis unit 5 performs item extraction processing based on the edited speech recognition results.

In this modified example, the call screen 51 shown in Figure 14 is displayed on the operator terminal 8. The call screen 51 has an edit button 54. When the operator operates the edit button 54, the system switches to edit mode.

In edit mode, the operator can manually modify the voice recognition results displayed on the voice recognition result display unit 53 using the input device 12. Also, in edit mode, the reflection of new voice recognition results on the call screen 51 is temporarily stopped. That is, only the voice recognition results reflected on the call screen 51 before switching to edit mode are displayed on the call screen 51, and voice recognition results output from the voice recognition unit 4 after switching to edit mode are not reflected on the call screen 51.

When the operator has finished editing, they can cancel the edit mode by operating the edit button 54 again. When the edit mode is canceled, the voice recognition results that had been stopped from being reflected on the call screen 51 up until that point are reflected on the call screen 51, and the voice recognition results output from the voice recognition unit 4 after the edit mode is canceled are also reflected sequentially on the call screen 51.

In this way, in this modified example, when the operator switches to edit mode, the reflection of the voice recognition results on the call screen 51 is temporarily stopped, and when the editing operation is completed and the edit mode is released, the reflection of the voice recognition results on the call screen 51 resumes. This prevents any corrections made manually by the operator during an editing operation on the voice recognition results from being overwritten by the next voice recognition result.

(Fourth Modification)
Next, a fourth modified example will be described. It should be noted that the points not specifically mentioned here are the same as those in the above-described embodiment. Fig. 15 is an explanatory diagram showing a call screen 51 displayed on the operator terminal 8 according to the fourth modified example.

In this modified example, the call screen 51 shown in Figure 15 is displayed on the operator terminal 8. A separator mark 55 is displayed in the voice recognition result display area 53 of the call screen 51.

In this modified example, as in the previous embodiment, speech recognition is performed on the voice of a call in predetermined time intervals (e.g., 30 seconds), and the speech recognition results for each predetermined time interval are displayed sequentially in the speech recognition result display area 53 on the call screen 51. At this time, a separator mark 55 is inserted each time a speech recognition result for a predetermined time interval is displayed. That is, the next speech recognition result output for the predetermined time interval is added after the separator mark 55. The speech recognition results are confirmed within the range separated by the separator mark 55. On the other hand, the speech recognition results added after the separator mark 55 are unconfirmed, and speech recognition results are added one after another as the speech recognition process progresses.

In the range delimited by the delimiter marks 55, i.e., the range where the speech recognition results have been finalized, the operator can freely edit the speech recognition results. The delimiter marks 55 also allow the operator to identify the range where the speech recognition results have been finalized. This allows the operator to manually correct the speech recognition results using the input device 12 for the finalized range before the delimiter marks 55, preventing the operator's manual corrections from being overwritten by the speech recognition results.

In this modified example, delimiter marks 55 are inserted to distinguish between the range of voice recognition results for a predetermined time unit that are displayed sequentially in the voice recognition result display area 53 of the call screen 51. However, since it is only necessary to distinguish between the range of confirmed voice recognition results and the range of unconfirmed voice recognition results, it is also possible to leave only the most recent delimiter mark 55 and erase the previous delimiter marks 55. Furthermore, it is also possible to distinguish between the range of confirmed voice recognition results and the range of unconfirmed voice recognition results by changing the display format of the voice recognition result characters (such as character color, italics, etc.) or by changing the display format of the background of the voice recognition result characters (such as color, pattern, etc.).

(Fifth Modification)
Next, a fifth modified example will be described. It should be noted that the points not specifically mentioned here are the same as those in the above-described embodiment. Fig. 16 is an explanatory diagram showing a call screen 51 displayed on the operator terminal 8 according to the fifth modified example.

In this modified example, the call screen 51 shown in Figure 16 is displayed on the operator terminal 8. The call screen 51 has a memo input field 56. During a call, the operator can enter phrases or sentences that express the content of the call that the operator finds interesting in the call with the customer as notes in the memo input field 56.

When the main control unit 3 sends a prompt to the text analysis unit 5 instructing it to perform item extraction processing on the speech recognition results, the main control unit 3 adds the notes entered in the note input field 56 to the prompt along with the text of the speech recognition results and sends them to the text analysis unit 5. The text analysis unit 5 uses generation AI to perform item extraction processing based on the speech recognition results and notes.

In this way, this modified example allows the operator to take notes of things that they notice during a call before they forget them. Furthermore, because the contents of the notes are reflected in the item extraction results, the accuracy of the item extraction process can be improved, thereby improving the quality of the customer interaction records registered based on the item extraction results.

(Sixth Modification)
Next, a sixth modified example will be described. It should be noted that the points not specifically mentioned here are the same as those of the above-described embodiment. Fig. 17 is an explanatory diagram showing a call screen 51 displayed on an operator terminal 8 according to the sixth modified example. Fig. 18 is a sequence diagram showing the procedure of processing performed in the operator support system according to the sixth modified example.

In this variant, past cases similar to the current inquiry are searched for during the call and presented to the operator. This allows for smoother phone responses to customer inquiries and increases customer satisfaction.

In this modified example, the call screen 51 shown in Figure 17 is displayed on the operator terminal 8. The call screen 51 has a similar case display field 57. The similar case display field 57 displays information about past cases similar to the current customer's inquiry (for example, a summary of the inquiry).

The similar case search is also performed by the text analysis unit 5 using generation AI. The text analysis unit 5 obtains information about past cases from the management unit 6 and searches for cases similar to the current inquiry. At this time, the relevance of the past cases to the current inquiry is obtained, and cases with high relevance are presented to the operator. In the example shown in Figure 17, the three most highly relevant cases are selected, and information about those three cases is displayed in the similar case display field 57.

The call screen 51 also has a hold button 58. The operator operates the hold button 58 when they cannot immediately respond to a customer's inquiry. When the operator operates the hold button 58, the call between the customer and the operator is put on hold, and at the same time, the text analysis unit 5 performs a similar case search based on the speech recognition results up to that point, and the similar cases are displayed in the similar case display field 57. A dedicated operation button for instructing the execution of a similar case search may also be provided. Furthermore, during the call, the text analysis unit 5 may periodically (for example, every 30 seconds) perform a similar case search based on the speech recognition results up to that point, and the similar cases displayed in the similar case display field 57 may be periodically updated.

Specifically, as shown in FIG. 18, when the operator operates the hold button 58 on the call screen 51, the GUI control unit 7 sends a hold instruction and a similar case search instruction to the main control unit 3. Next, the main control unit 3 sends the hold instruction to the call unit 1, which then enters a hold state. The main control unit 3 also sends a similar case search instruction (prompt) to the text analysis unit 5. Next, the text analysis unit 5 executes a similar case search process using generation AI and sends information on similar cases as the search results to the main control unit 3. At this time, the text analysis unit 5 requests information on past cases from the management unit 6, obtains information on past cases from the management unit 6, and searches for cases similar to the current inquiry from among those cases. Next, the main control unit 3 sends information on similar cases as the search results to the GUI control unit 7. Next, the GUI control unit 7 displays the information on similar cases in the similar case display field 57 on the call screen 51. Once the operator has confirmed the similar case, they can press the hold button 58 again to release the hold and resume the call with the customer, allowing them to respond to the inquiry, for example, by explaining measures to resolve the problem.

(Seventh Modification)
Next, a seventh modified example will be described. It should be noted that the points not specifically mentioned here are the same as those in the above-described embodiment. Fig. 19 is an explanatory diagram showing a confirmation screen 61 displayed on the operator terminal 8 according to the seventh modified example. Fig. 20 is a sequence diagram showing the procedure of processing performed in the operator support system according to the seventh modified example.

In this variant, the text analysis unit 5 uses generation AI to proofread the speech recognition results output from the speech recognition unit 4.

In this modified example, a confirmation screen 61 shown in FIG. 19 is displayed on the operator terminal 8. A proofreading button 64 is provided on the confirmation screen 61. By operating the proofreading button 64, the operator can have the text analysis unit 5 proofread the speech recognition results.

Specifically, as shown in FIG. 20, when the operator operates the proofreading button 64 on the confirmation screen 61, the GUI control unit 7 sends a proofreading instruction to the main control unit 3. Next, the main control unit 3 sends a proofreading instruction (prompt) to the text analysis unit 5. Next, the text analysis unit 5 performs proofreading processing using AI and sends the proofreading results to the main control unit 3. Next, the main control unit 3 sends the proofreading results to the GUI control unit 7. Next, the GUI control unit 7 updates the display content of the speech recognition result display unit 62 on the confirmation screen 61 to the proofreading results (proofreaded speech recognition results). When the operator checks the proofreading results and operates the item extraction button 63, the text analysis unit 5 executes item extraction processing based on the proofreaded speech recognition results.

In this way, in this variant, the speech recognition results are proofread by the generation AI before the item extraction process is performed, thereby improving the accuracy of the item extraction process and the quality of the customer interaction records registered based on the item extraction results.

In addition, in this modified example, the operator can specify the target range of the proofreading performed by the text analysis unit 5. Specifically, the operator can specify the target range of the proofreading process for the speech recognition results displayed in the speech recognition result display area 62 of the confirmation screen 61 shown in FIG. 19 by performing an operation similar to the operation for specifying the target range for the item extraction process (see FIGS. 7 and 8). The text analysis unit 5 performs proofreading on the speech recognition results included in the specified target range.

In this modified example, when the operator performs an operation on the operator terminal 8 to instruct proofreading, proofreading is performed in the text analysis unit 5, and the proofreading result (proofreaded speech recognition result) is temporarily displayed on the operator terminal 8. When the operator then performs an operation to instruct item extraction, item extraction is performed in the text analysis unit 5. On the other hand, when the operator performs an operation on the operator terminal 8 to instruct proofreading and item extraction, proofreading and item extraction may be performed consecutively in the text analysis unit 5.

As mentioned above, the embodiments have been described as examples of the technology disclosed in this application. However, the technology in this disclosure is not limited to these, and can also be applied to embodiments in which modifications, substitutions, additions, omissions, etc. are made. Furthermore, it is possible to combine the components described in the above embodiments to create new embodiments.

The operator support system, operator support device, and operator support method according to the present invention assist operators who create customer interaction records, thereby reducing the operator's workload and improving the quality of customer interaction records. They are useful as operator support systems, operator support devices, and operator support methods that support the work of operators who create customer interaction records based on inquiries made by customers over the phone.

1: Communication unit 2: Voice interception unit 3: Main control unit (operator support device)
4: Speech recognition unit 5: Text analysis unit 6: Management unit 7: GUI control unit 8: Operator terminal 9: Telephone 11: Display device 12: Input device 13: Microphone 14: Speaker 23: Processor 33: Processor 51: Call screen 53: Speech recognition result display unit 54: Edit button 55: Delimiter mark 56: Memo input field 57: Similar case display field 58: Hold button 59: Inquiry content display field 61: Confirmation screen 62: Speech recognition result display unit 63: Item extraction button 64: Proof button 71: Registration screen 72, 73, 74: Input field 75: Register button 76: Tab 81: Call screen 83: Start time button 84: End time button 85: Item extraction button

Claims (11)

An operator support system in which one or more processors execute a process for supporting an operator's work of creating a customer interaction record,
The processor:
instructing a speech recognition unit to execute a speech recognition process for converting speech of the customer and the operator into text by speech recognition in a predetermined time unit , and acquiring a speech recognition result for the predetermined time unit ;
instructing a text analysis unit to execute an item extraction process to extract a draft of registration information for each item of the customer interaction record based on the speech recognition results included in the target range in response to an operation by an operator to specify a target range for the speech recognition results, thereby obtaining the item extraction results;
displaying a screen on an operator terminal to present the item extraction results to an operator;
An operator support system characterized in that, in response to an operation by an operator who has confirmed the item extraction results displayed on the screen, the system instructs a management section to register the item extraction results as the customer interaction record. The processor:
The operator support system of claim 1, characterized in that when a single customer interaction includes multiple topics, the text analysis unit is instructed to perform the item extraction process for each of the multiple topics in response to an operator's operation to separately select a range related to each of the multiple topics, and the item extraction results for each of the multiple topics are obtained. The processor:
displaying a screen on the operator terminal to present the speech recognition result to an operator;
2. The operator support system according to claim 1, wherein, in response to an operator's operation of specifying the target range for the speech recognition result displayed on the screen, the operator instructs the text analysis unit to execute the item extraction process based on the speech recognition result included in the target range, thereby obtaining the item extraction result. The processor:
The operator support system according to claim 1, characterized in that, in response to an operator's operation to specify the start and end points of the target range while serving a customer, the text analysis unit is instructed to execute the item extraction process based on the speech recognition result for speech included in the target range, and the item extraction result is obtained. The processor:
Instructs the call section to provide a guide to have the customer explain the outline of the inquiry before the operator responds, and acquires the voice of the inquiry spoken by the customer,
instructing the speech recognition unit to perform the speech recognition process on the speech of the inquiry content, and acquiring text of the inquiry content as the speech recognition result;
instructing the text analysis unit to analyze the text of the query to obtain tag information and a summary of the query;
Selecting an operator who matches the inquiry content based on the tag information;
instructing the call unit to connect the customer's call to the selected operator;
2. The operator support system according to claim 1, wherein a screen for presenting the summary to the operator is displayed on the operator terminal. An operator support system in which one or more processors execute a process for supporting an operator's work of creating a customer interaction record,
The processor:
Instruct the speech recognition unit to execute a speech recognition process that converts the speech of the customer and the operator into text by speech recognition, and obtain the speech recognition results.
instructing a text analysis unit to execute an item extraction process to extract a draft of registration information for each item of the customer interaction record based on the speech recognition results included in the target range in response to an operation by an operator to specify a target range for the speech recognition results, thereby obtaining the item extraction results;
displaying a screen on an operator terminal to present the item extraction results to an operator;
instructing a management unit to register the item extraction result as the customer interaction record in response to an operation by an operator who has confirmed the item extraction result displayed on the screen;
Furthermore,
During the call, a screen for presenting the speech recognition result to an operator is displayed on the operator terminal;
acquiring a corrected speech recognition result in response to a user's operation to correct the speech recognition result displayed on the screen;
instructing the text analysis unit to execute the item extraction process based on the corrected speech recognition result, and obtaining the item extraction result;
In addition, when a screen that presents the voice recognition result to an operator is switched to the edit mode in response to a user operation to instruct the operator to switch to the edit mode, the operator support system accepts a user operation to correct the voice recognition result and temporarily stops reflecting the voice recognition result on the screen. An operator support system in which one or more processors execute a process for supporting an operator's work of creating a customer interaction record,
The processor:
Instruct the speech recognition unit to execute a speech recognition process that converts the speech of the customer and the operator into text by speech recognition, and obtain the speech recognition results.
instructing a text analysis unit to execute an item extraction process to extract a draft of registration information for each item of the customer interaction record based on the speech recognition results included in the target range in response to an operation by an operator to specify a target range for the speech recognition results, thereby obtaining the item extraction results;
displaying a screen on an operator terminal to present the item extraction results to an operator;
instructing a management unit to register the item extraction result as the customer interaction record in response to an operation by an operator who has confirmed the item extraction result displayed on the screen;
Furthermore,
displaying a screen on the operator terminal for the operator to input notes during the call, and acquiring the notes input by the operator;
An operator support system characterized by instructing the text analysis unit to execute the item extraction process based on the speech recognition result and the memo, and obtaining the item extraction result that reflects the content of the memo. An operator support system in which one or more processors execute a process for supporting an operator's work of creating a customer interaction record,
The processor:
Instruct the speech recognition unit to execute a speech recognition process that converts the speech of the customer and the operator into text by speech recognition, and obtain the speech recognition results.
instructing a text analysis unit to execute an item extraction process to extract a draft of registration information for each item of the customer interaction record based on the speech recognition results included in the target range in response to an operation by an operator to specify a target range for the speech recognition results, thereby obtaining the item extraction results;
displaying a screen on an operator terminal to present the item extraction results to an operator;
instructing a management unit to register the item extraction result as the customer interaction record in response to an operation by an operator who has confirmed the item extraction result displayed on the screen;
Furthermore,
Instructing the text analysis unit to execute a similar case search process to search for past cases similar to the content of the customer's current inquiry, and obtaining information on similar cases as a search result;
An operator support system characterized in that the information on the similar cases is displayed on the operator terminal. An operator support system in which one or more processors execute a process for supporting an operator's work of creating a customer interaction record,
The processor:
Instruct the speech recognition unit to execute a speech recognition process that converts the speech of the customer and the operator into text by speech recognition, and obtain the speech recognition results.
instructing a text analysis unit to execute an item extraction process to extract a draft of registration information for each item of the customer interaction record based on the speech recognition results included in the target range in response to an operation by an operator to specify a target range for the speech recognition results, thereby obtaining the item extraction results;
displaying a screen on an operator terminal to present the item extraction results to an operator;
instructing a management unit to register the item extraction result as the customer interaction record in response to an operation by an operator who has confirmed the item extraction result displayed on the screen;
Furthermore,
instructing the text analysis unit to perform a process of proofreading the speech recognition result to obtain a proofread speech recognition result;
an operator support system that instructs the text analysis unit to execute the item extraction process on the calibrated speech recognition result, and obtains the item extraction result; An operator support device in which a processor executes a process for supporting an operator's work of creating a customer interaction record,
The processor:
instructing a speech recognition unit to execute a speech recognition process for converting speech of the customer and the operator into text by speech recognition in a predetermined time unit , and acquiring a speech recognition result for the predetermined time unit ;
instructing a text analysis unit to execute an item extraction process to extract text corresponding to each item of the customer interaction record based on the speech recognition results included in the target range in response to an operator's operation to specify a target range for the speech recognition results, thereby obtaining the item extraction results;
instructing a GUI control unit to display a screen on an operator terminal that presents the item extraction results to an operator;
An operator support device characterized in that, in response to an operation by an operator who has confirmed the item extraction result displayed on the screen, the device instructs a management section to register the item extraction result as the customer interaction record. 1. An operator support method for causing one or more processors to execute a process for supporting an operator's work of creating a customer interaction record, comprising:
instructing a speech recognition unit to execute a speech recognition process for converting speech of the customer and the operator into text by speech recognition in a predetermined time unit , and acquiring a speech recognition result for the predetermined time unit ;
instructing a text analysis unit to execute an item extraction process to extract text corresponding to each item of the customer interaction record based on the speech recognition results included in the target range in response to an operator's operation to specify a target range for the speech recognition results, thereby obtaining the item extraction results;
displaying a screen on an operator terminal to present the item extraction results to an operator;
An operator support method comprising the steps of: instructing a management section to register the item extraction result as the customer interaction record in response to an operation by an operator who has confirmed the item extraction result displayed on the screen.