Detection device, detection method, and program

Abstract

A detection device according to one embodiment of the present invention includes: a detection unit that, on the basis of detection conditions for an attention word representing a word, a term, a phrase, or a sentence to be paid attention to or to be aware of, detects the aforementioned attention word from a voice recognition result from an interaction between a plurality of people; and a display information provision unit that, when the attention word is detected, transmits display information, which is for displaying the detected attention word or a display component representing the attention word in a pre-registered display mode, to a predetermined terminal.

Description

Technical Field

[0001] The present invention relates to a detection device, a detection method, and a program.

Background Art

[0002] For the purpose of improving the response quality of operators in a call center (or also called a contact center) and enhancing the efficiency of the work of supervisors who monitor and support the calls of operators, etc., a technique for detecting keywords during a call and notifying an alert based on the number of occurrences thereof is known (for example, Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, in the prior art, it was not possible to flexibly set the detection conditions of keywords, etc. For this reason, for example, there were cases where business support such as response support for operators and improvement of monitoring efficiency of supervisors could not be fully realized.

[0005] One embodiment of the present invention has been made in view of the above points, and an object thereof is to realize support for specific operations.

Means for Solving the Problems

[0006] To achieve the above object, a detection device according to an embodiment includes a detection unit that detects the target word from the speech recognition results in the conversation among a plurality of persons based on the detection conditions of the target word representing a word, a phrase, a clause, a phrase, or a sentence to be noted or paid attention to, and a display information providing unit that, when the target word is detected, transmits display information for causing a predetermined terminal to display the detected target word or a display component representing the target word in a pre-registered display mode.

Effect of the Invention

[0007] Support for specific operations can be realized.

Brief Description of the Drawings

[0008]

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7

Figure 8

Figure 9

Figure 10

Figure 11

Figure 12

Figure 13

Embodiments for Carrying Out the Invention

[0009] Hereinafter, an embodiment of the present invention will be described. In this embodiment, a contact center system 1 that can assist the operations of operators and supervisors will be described with respect to a contact center. Here, examples of the operations of an operator include, for example, telephone response operations with customers. On the other hand, examples of the operations of a supervisor include, for example, operations of monitoring the calls of operators and assisting in the telephone responses of operators, operations of analyzing the calls of each operator, and the like. However, it goes without saying that these operations are just examples, and operators and supervisors can perform various other operations. Also, for example, the operation of analyzing the calls of each operator may be performed by other persons such as analysts instead of supervisors.

[0010] Note that targeting a contact center is just an example, and it can be similarly applied to cases where, for example, an office or the like is targeted to assist in the telephone response operations of the persons working there, the monitoring and support operations thereof, the call analysis operations, and the like.

[0011] Hereinafter, cases will be described where when a keyword registered in advance is spoken during a call between an operator and a customer, the fact that the keyword has been spoken is notified to the operator or supervisor in real time, or where the call containing the keyword is searched for analysis such as response quality evaluation after the call ends. Thereby, it is possible to assist the telephone response operations of operators and the monitoring and support operations of supervisors, and it is also possible to assist call analysis operations. Note that the keyword is a word, term, phrase, sentence, etc. that an operator or supervisor should pay attention to or be aware of, and will be referred to as a "target word" hereinafter.

[0012] <Overall Configuration of Contact Center System 1> An example of the overall configuration of the contact center system 1 according to this embodiment is shown in FIG. 1. As shown in FIG. 1, the contact center system 1 according to this embodiment includes a keyword detection device 10, one or more operator terminals 20, one or more supervisor terminals 30, a PBX (Private Branch Exchange) 40, and a customer terminal 50. Here, the keyword detection device 10, the operator terminal 20, the supervisor terminal 30, and the PBX 40 are installed in a contact center environment E, which is a system environment of the contact center. Note that the contact center environment E is not limited to a system environment within the same building, and may be, for example, a system environment within a plurality of geographically separated buildings.

[0013] The keyword detection device 10 converts an audio call between a customer and an operator into text by voice recognition, then detects a keyword from this text, and causes the detected keyword to be displayed on the operator terminal 20 or notifies the supervisor terminal 30 that the keyword has been detected. Further, in response to a search request from the supervisor terminal 30, the keyword detection device 10 searches for calls containing the keyword from past call histories and displays the search results on the supervisor terminal 30.

[0014] The operator terminal 20 is various terminals such as a PC (Personal Computer) used by an operator and functions as an IP (Internet Protocol) telephone. A response support screen is displayed on the operator terminal 20 during a call with a customer. The response support screen is a screen that visualizes the content of an audio call with a customer in real time, and visualizes the text obtained by converting the audio call with the customer by voice recognition and the keyword detected from it.

[0015] The supervisor terminal 30 is various terminals such as a PC (Personal Computer) used by the supervisor. Note that the supervisor is a person who monitors the operator's calls and supports the operator's telephone response operations when some problem is likely to occur or in response to a request from the operator. Usually, it is common for the calls of several to a dozen or so operators to be monitored by one supervisor.

[0016] On the supervisor terminal 30, an operator support screen, a call list screen, etc. are displayed. The operator support screen is a screen for monitoring and supporting the operator's calls, and when a keyword is detected during the operator's call, that fact is notified. Also, the call list screen is a screen for searching calls for analysis such as response quality evaluation. For example, by specifying a keyword as a search condition, a list of calls containing that keyword is displayed.

[0017] The PBX 40 is a telephone exchange (IP-PBX) and is connected to a communication network 60 including a VoIP (Voice over Internet Protocol) network and a PSTN (Public Switched Telephone Network). When there is an incoming call from the customer terminal 50, the PBX 40 calls one or more predetermined operator terminals 20 and connects the customer terminal 50 to any operator terminal 20 that responds to the call.

[0018] The customer terminal 50 is various terminals such as a smartphone, mobile phone, fixed phone, etc. used by the customer.

[0019] Note that the overall configuration of the contact center system 1 shown in FIG. 1 is an example, and other configurations may be used. For example, although the keyword detection device 10 is included in the contact center environment E (that is, the keyword detection device 10 is on-premises), all or some of the functions of the keyword detection device 10 may be realized by a cloud service or the like. Similarly, in the example shown in FIG. 1, the PBX 40 is an on-premises telephone exchange, but it may be realized by a cloud service.

[0020] <Functional Configuration of the Attention Word Detection Device 10> The functional configuration of the attention word detection device 10 according to the present embodiment is shown in FIG. 2. As shown in FIG. 2, the attention word detection device 10 according to the present embodiment includes an attention word registration unit 101, a speech recognition text conversion unit 102, an attention word detection unit 103, a display information providing unit 104, and a search unit 105. Each of these units is realized, for example, by processing in which one or more programs installed in the attention word detection device 10 cause a processor such as a CPU (Central Processing Unit) to execute.

[0021] Further, the attention word detection device 10 according to the present embodiment includes a registration information DB 106 and a call history information DB 107. Each of these DBs (databases) is realized, for example, by an auxiliary storage device such as an HDD (Hard Disk Drive) or an SSD (Solid State Drive). Note that at least one of these DBs may be realized, for example, by a database server or the like connected to the attention word detection device 10 via a communication network.

[0022] The attention word registration unit 101 stores registration information including detection conditions for attention words and the like in the registration information DB 106 in response to a registration request from the supervisor terminal 30. As a result, the detection conditions for attention words and the like are registered in the attention word detection device 10.

[0023] The speech recognition text conversion unit 102 converts a voice call between the operator terminal 20 and the customer terminal 50 into text by speech recognition. At this time, the speech recognition text conversion unit 102 performs speech recognition for each speaker and converts it into text. As a result, the voice of the operator and the voice of the customer are each converted into text.

[0024] The attention word detection unit 103 detects attention words in the text converted by the speech recognition text conversion unit 102 based on the detection conditions of the registration information stored in the registration information DB 106.

[0025] The display information providing unit 104 transmits information (hereinafter also referred to as display information) for causing various screens (for example, a response support screen, an operator support screen, a call list screen, etc.) to be displayed on the operator terminal 20 or the supervisor terminal 30.

[0026] In response to a search request from the supervisor terminal 30, the search unit 105 searches the call history information DB 107 for call history information that matches the search conditions.

[0027] The registration information DB 106 stores registration information. The registration information is information that at least includes detection conditions for detecting a keyword. As will be described later, the registration information includes various information (for example, the organization to be the detection target of the keyword, the type and name of the keyword, the icon and color representing the keyword, the screen to be the display destination when the keyword is detected, etc.) set by the supervisor on the keyword registration screen.

[0028] The call history information DB 107 stores call history information. The call history information is information representing past call contents, and includes, for example, call ID, call date and time (call start date and time and call end date and time), information about the operator corresponding to the call (operator ID, name, affiliated organization, etc.), the extension number of the operator and the customer's phone number, the text converted by the voice recognition text conversion unit 102, information about the keyword in the text, etc. Here, the information about the keyword is, for example, the keyword and information necessary for displaying the keyword on various screens (for example, information indicating the display mode such as the color and icon type when the keyword is displayed).

[0029] Note that an operator belongs to at least one organization, and it is common for the response content to vary depending on the organization. To give a typical example, there are organizations such as "Contract", "Maintenance", and "Service Guide". An operator belonging to the "Contract" organization is responsible for handling inquiries related to contracts, an operator belonging to the "Maintenance" organization is responsible for handling inquiries related to maintenance such as troubleshooting, and an operator belonging to the "Service Guide" organization is responsible for handling inquiries related to guides such as providing information on service content.

[0030] <Registration Process of Keywords of Interest> The following describes the registration process of keywords of interest with reference to FIG. 3.

[0031] In response to a display request from the supervisor terminal 30, the display information providing unit 104 of the keyword detection device 10 transmits the display information of the keyword registration screen to the supervisor terminal 30 (step S101). As a result, the keyword registration screen is displayed on the display of the supervisor terminal 30.

[0032] Here, an example of the keyword registration screen is shown in FIG. 4. The keyword registration screen 1000 shown in FIG. 4 includes an organization setting column 1001, a keyword type setting column 1002, a name setting column 1003, a display order setting column 1004, an icon setting column 1005, a color setting column 1006, a setting column 1007 for whether to display the response support screen, a setting column 1008 for whether to display the operator support screen, a setting column 1009 for whether to display the call list screen, a setting column 1010 for whether to use response quality evaluation, a detection condition setting column 1011, a preview display button 1012, a cancel button 1013, and a registration button 1014.

[0033] The organization setting column 1001 is a setting column in which the organization to be the detection target of the keyword of interest is set. Calls of operators belonging to the organization set in the organization setting column 1001 become the detection target of the keyword of interest. Note that when "Root" is set, it means that all organizations are detection targets.

[0034] The keyword type setting column 1002 is a setting column in which a character string (keyword type) that uniquely identifies a keyword is set under the organization set in the organization setting column 1001. The keyword type can be set arbitrarily. For example, "NG" which indicates an inappropriate keyword under the organization, "〇〇 service" which indicates a keyword related to a certain specific service under the organization, etc. can be considered. Thus, a keyword can be uniquely identified as "organization / character string" such as "Root / NG", "Root / 〇〇 service". This is because there are keywords common to all organizations and also keywords specific to an organization. Incidentally, hereinafter, "Root" will be simply denoted as "R", and for example, "Root / NG" may be denoted as "R / NG".

[0035] The name setting column 1003 is a setting column in which an arbitrary name for the keyword is set. The display order setting column 1004 is a setting column in which the order of arrangement when arranging and displaying the icons of the keywords on the operator support screen described later is set. The icon setting column 1005 is a setting column in which the icon to be displayed on the operator support screen described later is set. Incidentally, the icon is not only selected and set from among those determined in advance, but for example, arbitrary image data etc. may be uploaded and the image may be set as the icon.

[0036] The color setting column 1006 is a setting column in which the color for highlighting the keyword when the keyword is detected on the response support screen etc. described later is set. Incidentally, the color is not only selected and set from among those determined in advance, but for example, an arbitrary color may be set by a color code etc.

[0037] The response support screen display presence / absence setting field 1007 is a setting field for setting whether or not to highlight the attention word on the response support screen when the attention word is detected. The operator support screen display presence / absence setting field 1008 is a setting field for setting whether or not to display the presence / absence of detection of the attention word on the operator support screen. The call list screen display presence / absence setting field 1009 is a setting field for setting whether or not to display the attention word on the call list screen when the attention word is detected.

[0038] The response quality evaluation usage presence / absence setting field 1010 is a setting field for setting whether or not to use the detection result of the attention word for the analysis of response quality evaluation, and whether the detection of the attention word is to be a point addition target or a point deduction target when using it.

[0039] The detection condition setting field 1011 is a setting field for setting the attention word and its detection conditions, and includes a match count setting field 1101, a speaker setting field 1102 of one or more, and a detection pattern setting field 1103. Here, the speaker setting field 1102 and the detection pattern setting field 1103 can be increased or decreased. A new speaker setting field 1102 and detection pattern setting field 1103 are added by the add button 1104, and are deleted by the delete button 1105 corresponding to the speaker setting field 1102 and the detection pattern setting field 1103. In the example shown in FIG. 4, two rows of the speaker setting field 1102 and the detection pattern setting field 1103, namely the speaker setting field 1102-1 and the detection pattern setting field 1103-1, and the speaker setting field 1102-2 and the detection pattern setting field 1103-2, are displayed.

[0040] The match count setting field 1101 is a setting field where the match count is set. The speaker setting field 1102 is a setting field where the speaker of the keyword of interest is set. Note that as the speaker, any one of "Operator" representing the operator, "Customer" representing the customer, and "Both" representing both the operator and the customer can be set. The detection pattern setting field 1103 is a setting field where the keyword of interest is set. Here, in the detection pattern setting field 1103, a character string representing the keyword of interest can be set as it is, or the keyword of interest can be set using a regular expression. By setting the keyword of interest using a regular expression, for example, a set of keywords of interest such as keywords that are similar to each other or keywords that differ only in part can be set.

[0041] For example, when "1" is set in the match count setting field 1101, "Customer" is set in the speaker setting field 1102, and "Application" is set in the detection pattern setting field 1103, if the customer makes a statement of "Application" once during the call, the keyword of interest "Application" is detected.

[0042] Also, for example, when "3" is set in the match count setting field 1101, "Operator" is set in the speaker setting field 1102, and "AAA Service" is set in the detection pattern setting field 1103, if the operator makes a statement of "AAA Service" three times during the call, the keyword of interest "AAA Service" is detected. Note that in this case, if the operator makes a statement of "AAA Service" only twice during the call, it is not detected as a keyword of interest.

[0043] The preview display button 1012 is a button for previewing and displaying the response support screen, operator support screen, and call list screen when a keyword of interest is detected. The cancel button 1013 is a button for canceling the registration of the keyword of interest. The registration button 1014 is a button for registering the information set on the keyword of interest registration screen 1000. When the registration button 1014 is pressed by the supervisor, a registration request including the information set in the organization setting field 1001 to the detection condition setting field 1011 is sent from the supervisor terminal 30 to the keyword of interest detection device 10.

[0044] Return to FIG. 3. The keyword registration unit 101 of the keyword detection device 10 receives a registration request from the supervisor terminal 30 (step S102).

[0045] Then, the keyword registration unit 101 of the keyword detection device 10 creates registration information from the registration request and stores the created registration information in the registration information DB 106 (step S103). Note that the keyword registration unit 101 may create, for example, registration information including a registration information ID and the information included in the registration request.

[0046] <Display processing of response support screen and operator support screen> Hereinafter, the display processing of the response support screen and the operator support screen will be described with reference to FIG. 5. The response support screen is displayed on the display of the operator terminal 20 during a call with a customer, and the operator support screen is displayed on the display of the supervisor terminal 30 during business. Hereinafter, it is assumed that a certain operator and a certain customer are having a call, and the display processing will be described for the response support screen displayed on the operator terminal 20 used by this operator and the operator support screen displayed on the supervisor terminal 30 used by the supervisor who monitors and supports this operator. Also, assuming that the voice for each certain segment (for example, the voice for each segment such as a sentence or a phrase) is converted into text by the voice recognition text conversion unit 102, the following steps S201 to S206 are repeatedly executed every time the voice recognition result for the segment unit is obtained.

[0047] The keyword detection unit 103 of the keyword detection device 10 detects a keyword from the text obtained by converting the call by voice recognition, using the detection conditions included in the registration information of the organization to which the operator belongs among the registration information stored in the registration information DB 106 (step S201). As described above, the detection conditions include the number of matches, the speaker, and the detection pattern. When the speaker utters the keyword represented by the detection pattern more than the number of matches during the call, the keyword is detected.

[0048] The keyword detection unit 103 of the keyword detection device 10 determines whether a keyword has been detected in step S201 above (step S202).

[0049] If it is determined in step S202 above that a keyword has been detected, the display information providing unit 104 of the keyword detection device 10 transmits display information including the text representing the voice recognition result, the keyword detected in step S201 above, and highlighting information for highlighting the keyword in the color set in the color setting column 1006 when registering this keyword to the operator terminal 20 (step S203). As a result, on the response support screen of the operator terminal 20, the text is displayed and the keyword in the text is highlighted.

[0050] Here, an example of the response support screen is shown in FIG. 6. The response support screen 2000 shown in FIG. 6 includes a speech history column 2100 in which the text representing the voice recognition result is displayed in real time in chronological order. In the example shown in FIG. 6, speech texts 2101 to 2106 are displayed in the speech history column 2100. These speech texts 2101 to 2106 are those in which the text representing the voice recognition result is displayed in chronological order. In the speech history column 2100, fillers such as "um" are displayed in parentheses.

[0051] Also, in the example shown in FIG. 6, "maintenance service" in the speech text 2102 is highlighted. This indicates that the keyword "maintenance service" has been detected in this speech text 2102.

[0052] Similarly, "Ah" and "maintenance service" in the speech text 2103 are highlighted. This indicates that the attention words "Ah" and "maintenance service" have been detected in this speech text 2103. Note that "Ah" and "maintenance service" are different attention words and are set in different colors, so they are highlighted in different colors.

[0053] Similarly, "Ah" in the speech text 2106 is highlighted. This also indicates that the attention word "Ah" has been detected in this speech text 2106.

[0054] As a result, when an attention word (for example, a word, term, phrase, sentence, etc. that should be noted or paid attention to) is detected, the operator can easily confirm it visually and in real time. Moreover, since it is highlighted in different colors according to the attention word, it is also possible to easily know what kind of attention word it is by the color (for example, the attention word highlighted in red is an "NG" phrase, etc.). Therefore, the operator can, for example, narrow down the confirmation to only the attention words that he / she should truly pay attention to, and can efficiently confirm the attention words.

[0055] In addition, when, for example, a pointer or the like is superimposed on the highlighted attention word, the attention word type set in the attention word type setting column 1002 or the name set in the name setting column 1003 when registering the attention word may be displayed. For example, as shown in FIG. 7, when a pointer or the like is superimposed on "Ah" in the speech text 2103, a balloon 2110 including "NG" representing the type or name of the attention word "Ah" is displayed.

[0056] Return to FIG. 5. Subsequently to step S203, the display information providing unit 104 of the keyword detection device 10 transmits display information including the notification information of the keyword to the supervisor terminal 30 (step S204). Note that the notification information of the keyword is information for notifying that the keyword has been detected. As a result, on the operator support screen of the supervisor terminal 30, it is notified that the keyword has been detected.

[0057] Here, an example of the operator support screen is shown in FIG. 8. The operator support screen 3000 shown in FIG. 8 includes a notification display column 3100 for notifying that a keyword has been detected in the call of an operator to be monitored / supported by the supervisor. In the example shown in FIG. 8, the operators to be monitored / supported are "Hanako Suzuki", "Taro Tanaka", and "Jiro Tanaka", and icons 3101 to 3103 corresponding to the keyword are displayed for each of them. That is, icons 3101-1 to 3103-1 are displayed for "Hanako Suzuki", icons 3101-2 to 3103-2 are displayed for "Taro Tanaka", and icons 3101-3 to 3103-3 are displayed for "Jiro Tanaka". These icons 3101 to 3103 are set in the icon setting column 1005 when registering the keyword.

[0058] Then, on the operator support screen 3000, when a keyword is detected, the icon corresponding to this keyword is highlighted. For example, in the example shown in FIG. 8, icons 3101-1 and 3103-1 are highlighted. This means that in the call of the operator "Hanako Suzuki", the keyword corresponding to icon 3101 and the keyword corresponding to icon 3103 have been detected. Similarly, in the example shown in FIG. 8, icon 3102-2 is highlighted. This means that in the call of the operator "Jiro Tanaka", the keyword corresponding to icon 3102 has been detected.

[0059] As a result, the supervisor can visually and easily confirm in real time when the target word is detected. Moreover, since the icon corresponding to the target word is highlighted, it is also possible to easily know what target word has been detected by that icon. Therefore, the supervisor can, for example, narrow down the confirmation to only the target words that the supervisor should truly pay attention to, and can efficiently confirm the target words.

[0060] In addition, the supervisor can display, on the supervisor terminal 30, a screen similar to the operator's response support screen by selecting the operator name on the operator support screen. As a result, when the target word is detected, it is possible to check the content of the call in which the target word was detected and provide support such as preventing troubles in advance or giving advice to the operator.

[0061] Return to FIG. 5. When it is not determined in step S202 above that the target word has been detected, the display information providing unit 104 of the target word detection device 10 transmits display information including text representing the speech recognition result to the operator terminal 20 (step S205). As a result, the text is displayed on the response support screen of the operator terminal 20.

[0062] When the call ends, the keyword detection unit 103 of the keyword detection device 10 creates call history information that includes at least the text obtained by converting the voice during this call by the voice recognition text conversion unit 102 and information related to the keyword during the call, and stores the created call history information in the call history information DB 107 (step S206). Note that in addition to the text and information related to the keyword, the call history information includes, for example, a call ID, call date and time (call start date and time and call end date and time), information related to the operator corresponding to the call (operator ID, name, affiliated organization, etc.), the extension number of the operator and the customer's phone number, etc. Further, the information related to the keyword includes not only the keyword but also information such as the color for highlighting the keyword and the type of icon corresponding to the keyword. Thereby, for example, after the call ends, it becomes possible to use the call history information for analysis such as response quality evaluation.

[0063] <Search Processing on Call List Screen> The search process on the call list screen will be described below with reference to FIG. 9.

[0064] The search unit 105 of the keyword detection device 10 receives a search request from the supervisor terminal 30 (step S301). Here, the supervisor terminal 30 can send a search request including the specified search conditions to the keyword detection device 10 by specifying the search conditions on the call list screen. For example, when the call list screen 4000 shown in FIG. 10 is displayed on the supervisor terminal 30, the supervisor can specify the desired search conditions from the search condition specification column 4100.

[0065] In the search condition specification field 4100 of the call list screen 4000 shown in FIG. 10, various search conditions such as "call content", "favorite", "call start date and time", "operator", "organization", "phone number", "emotion estimation", "keyword of interest", "important matter", and "memo" can be specified. For example, for "operator", the operator name, operator ID, etc. can be specified as search conditions. Similarly, for "organization", the organization name, organization ID, etc. can be specified as search conditions, and for "phone number", the phone number can be specified as a search condition. Also, for example, for "keyword of interest", the keyword of interest or the regular expression representing the keyword of interest can be specified as a search condition. Furthermore, in the search condition specification field 4100, when the keyword of interest or the regular expression representing the keyword of interest is specified as a search condition, the number of occurrences of that keyword of interest during the call can be further specified as a search condition. Thereby, for example, it becomes possible to search for "calls in which a certain keyword of interest was detected 1 or more times" or "calls in which a certain keyword of interest was detected 2 or more times".

[0066] Note that not only one keyword of interest but also a plurality of keywords of interest can be specified as search conditions. Also, it is possible to combine them with other search conditions (for example, specifying a search condition that combines "keyword of interest" and "organization", etc.).

[0067] Return to FIG. 9. The search unit 105 of the keyword of interest detection device 10 searches the call history information DB 107 for call history information that matches the search conditions included in the search request from the supervisor terminal 30 (step S302). For example, when the keyword of interest "〇〇 service" and the number of occurrences "2 or more times" are specified as search conditions, the search unit 105 searches the call history information DB 107 for the call history information of calls in which the keyword of interest "〇〇 service" was searched 2 or more times.

[0068] Then, the display information providing unit 104 of the keyword detection device 10 transmits the display information including the call history information retrieved in the above step S302 to the supervisor terminal 30 (step S303). As a result, the retrieved calls are listed and displayed on the call list screen on the supervisor terminal 30. For example, in the call information display column 4200 of the call list screen 4000 shown in FIG. 10, the call history information 4210 and the call history information 4220 retrieved in the above step S302 are displayed.

[0069] Also, at this time, in the call history information 4210, the type of the keyword detected in the call corresponding to this call history information 4210 is displayed in the keyword display column 4211. Similarly, in the call history information 4220, the type of the keyword detected in the call corresponding to this call history information 4220 is displayed in the keyword display column 4221. As a result, the supervisor can easily know which call detected which keyword. Therefore, for example, when performing an analysis such as coping quality evaluation from the viewpoint of whether a keyword has been detected or not, it is possible to easily find the call history information including the target keyword.

[0070] Note that in the keyword display column, for example, display components of "organization / keyword type" such as "R / NG" and "R / OP talk" are displayed, but further, the number of times the keyword has been searched may be displayed. That is, for example, display components of "organization / keyword type / number of searches" may be displayed. As a result, the supervisor can know how many times the keyword has been searched so far.

[0071] <Summary> As described above, in the keyword detection device 10 according to the present embodiment, at the organizational unit level, it is possible to set keywords using regular expressions, set the speaker thereof, and set the number of occurrences thereof during a single call as a detection condition. Further, the color, icon, etc. of the highlighting display when a keyword is detected can also be arbitrarily set. In this way, the detection conditions for keywords and the display mode when the keyword is detected can be flexibly set. Therefore, even when telephone response operations, monitoring / support operations, analysis operations, etc. may differ from organization to organization, it is possible to set appropriate detection conditions and display modes according to the organization.

[0072] Furthermore, since keywords can be set using regular expressions and similar or related keywords can be set simultaneously, it is also possible to reduce the work cost required for registering keywords.

[0073] <Modification Example> Hereinafter, a modification example of the present embodiment will be described. Note that the following modification examples can also be combined as appropriate.

[0074] ≪Modification Example 1≫ On the keyword registration screen, it may be possible to set whether to sound an alarm on the operator support screen when a keyword is detected. In this case, for example, as in the keyword registration screen 5000 shown in FIG. 11, an alarm notification setting column 5001 is included.

[0075] Then, on the operator support screen, when a keyword for which an alarm is to be sounded is detected, a predetermined alarm sound is output. Thereby, the supervisor can know by sound that a keyword has been detected.

[0076] Note that the alarm sound output when a keyword is detected may be selected from pre-determined sounds, or uploaded music data may be used as the alarm sound. Further, the alarm may be output only when a keyword with particularly high importance is detected.

[0077] <<Modification Example 2>> In the detection condition setting field of the keyword registration screen, it may be possible to set the scene to be the detection target of the keyword. For example, as shown in FIG. 12, the detection condition setting field 5010 may include a match count setting field 5101, a speaker setting field 5102 for one or more speakers, a detection pattern setting field 5103, and a scene setting field 5104. Here, the scene refers to the scene of the conversation taking place between the operator and the customer. Typical scenes include, for example, an "opening scene" representing a scene such as an initial greeting, an "inquiry grasping scene" representing a scene for grasping the content of an inquiry from the customer, a "response scene" representing a case of making a response to the inquiry content, an "identity confirmation scene" representing a case of performing identity confirmation, etc., a "service recommendation scene" representing a scene for recommending some service, a "closing scene" representing a scene such as a final greeting, and the like. Note that the scene can be specified by known techniques (for example, the techniques disclosed in International Publication No. 2020 / 036189).

[0078] In the example shown in FIG. 12, "1" is set in the match count setting field 5101, "operator" is set in the speaker setting field 5102-1, "thank you" is set in the detection pattern setting field 5103-1, and "service recommendation scene" is set in the scene setting field 5104-1. This means that when the operator makes a statement of "thank you" once in the service recommendation scene during the call, the keyword "thank you" is detected.

[0079] Similarly, in the example shown in FIG. 12, "1" is set in the match count setting field 5101, "both" is set in the speaker setting field 5102-2, "contract status" is set in the detection pattern setting field 5103-2, and "identity confirmation scene" is set in the scene setting field 5104-2. This means that when both the operator and the customer make a statement of "contract status" once in the identity confirmation scene during the call, the keyword "contract status" is detected.

[0080] This prevents the attention words set in the detection pattern setting field from being detected outside the scenes where the attention words are to be detected. For example, in the example shown in FIG. 12, "Thank you" spoken by the operator in a scene other than the service recommendation scene (for example, the opening scene) can be prevented from being detected as an attention word. Therefore, for example, it becomes possible to set such that words that are frequently spoken throughout the call but rarely spoken in a specific scene are detected as attention words.

[0081] <<Modification Example 3>> In the detection condition setting field of the attention word registration screen, it may be possible to set the type of utterance (question utterance and non-question utterance) for which the attention word is to be detected. For example, as shown in FIG. 13, the detection condition setting field 5020 may include a match count setting field 5201, a speaker setting field 5202 for one or more speakers, a detection pattern setting field 5203, and an utterance type setting field 5204. Here, the utterance type refers to the type of whether it is a question utterance representing a question to the conversation partner or a non-question utterance representing other utterances. Note that whether a certain utterance is a question utterance or a non-question utterance can be specified by known techniques.

[0082] In the example shown in FIG. 13, "1" is set in the match count setting field 5201, "Operator" is set in the speaker setting field 5202-1, "〇〇 Service" is set in the detection pattern setting field 5203-1, and "Question utterance" is set in the utterance type setting field 5204-1. This means that when the operator makes a question utterance including "〇〇 Service" once during the call, the attention word "〇〇 Service" is detected.

[0083] Similarly, in the example shown in FIG. 13, "1" is set in the match count setting field 5201, "Customer" is set in the speaker setting field 5202-2, "×× Service" is set in the detection pattern setting field 5203-2, and "Non-question utterance" is set in the utterance type setting field 5204-2. This means that when the customer makes a non-question utterance including "×× Service" once during the call, the attention word "×× Service" is detected.

[0084] This prevents the detection of the attention words set in the detection pattern setting field except for the utterance type to be the detection target of the attention words. For example, in the example shown in FIG. 13, when the operator makes a non-question utterance including "〇〇 service", it can be prevented from being detected as an attention word. Therefore, for example, it is possible to perform a setting such that a word that is frequently uttered in non-question utterances but rarely uttered in question utterances is detected as an attention word.

[0085] ≪Modification Example 4≫ In the above embodiment, the number of matches is set in the number of matches setting field. However, for example, for each attention word, the number of matches may be calculated from the same word in the past by a known test method or the like. Hereinafter, for the sake of simplicity, the call history information stored in the call history information DB 107 is referred to as "past call".

[0086] Specifically, taking a certain attention word as w, the average number of appearances μ of the attention word w in the past call w and the standard deviation as σ w are used. At this time, the number of matches is set to μ w + 2σ w . However, this is just an example. For example, μ w + 3σ w etc. may be set as the number of matches.

[0087] ≪Modification Example 5≫ In the above embodiment, the attention word is detected based on whether it matches the detection pattern set in the detection pattern setting field. However, for example, the attention word may be detected in consideration of the words that appeared before the word that matches the detection pattern. Thereby, it becomes possible to detect the attention word with higher accuracy in consideration of not only whether it simply matches the detection pattern but also the flow and context of the utterance up to that point.

[0088] Specifically, after performing the following preprocessing, the following detection process is performed in step S201 of FIG. 5, so that words of interest including the surrounding vocabulary of the detection pattern can be detected. Hereinafter, the word represented by the detection pattern is w x , and the past call is represented by i

[0089] ·Preprocessing (1-1) In the past call i, the utterance S x including w i,k , and the N (N is an integer of 1 or more determined in advance) utterances S i,k uttered before that utterance S i,k-1 , S i,k-2 , ···, S i,k-N are acquired. Then, the collection of these utterances is D i = {S i,k , S i,k-1 , S i,k-2 , ···, S i,k-N}}. This is done for each past call to obtain {D i |i = i 1 , i 2 , ···}

[0090] (1-2) For each D i (i = i 1 , i 2 , ···), for example, an existing technique such as Doc2Vec is applied to calculate the vector V i (i = i 1 , i 2 , ···).

[0091] (1-3) Calculate the average vector V i (i = i 1 , i 2 , ···) of the vector V ave .

[0092] ·Detection process (2-1) Represent the current call by r. When the utterance S x including w r,k is detected, the N utterances S r,k uttered before that utterance S r,k-1 , S r,k-2 , ···, Sr,k-N Obtain these. Then, the summary of these utterances is D r ={S r,k , S r,k-1 , S r,k-2 , ···, S r,k-N}.

[0093] (2-2) For D r , apply an existing technique such as Doc2Vec to calculate the vector V r .

[0094] (2-3) Calculate the similarity (e.g., cosine similarity, etc.) sim between the average vector V ave and the vector V r .

[0095] (2-4) When the similarity sim is greater than a predetermined threshold (e.g., 0.7, etc.), detect w x as the keyword of interest; otherwise, do not detect it as the keyword of interest.

[0096] Note that in the above (2-4), the number of matches is not considered (or the number of matches is set to 1), but the number of matches may be considered. In this case, when the similarity sim is greater than the predetermined threshold, w x may be set as the matched one.

[0097] ≪Modification Example 6≫ In the above embodiment, the detection of the keyword of interest was performed in real time in step S201 of FIG. 5. However, when the setting indicating that it is not displayed in the response support screen display presence / absence setting column 1007 and the operator support screen display presence / absence setting column 1008 is made, it is not necessarily required to be performed in real time. In this case, for example, the detection of the keyword of interest may be performed at an arbitrary timing (e.g., the timing after the call ends, the timing when the call history information is searched, etc.).

[0098] The present invention is not limited to the specifically disclosed above embodiments, and various modifications, changes, combinations with known techniques, etc. are possible without departing from the description of the claims.

Description of Symbols

[0099] 1 Contact center system 10 Keyword detection device 20 Operator terminal 30 Supervisor terminal 40 PBX 50 Customer terminal 60 Communication network 101 Keyword registration section 102 Speech recognition text conversion section 103 Keyword detection section 104 Display information providing section 105 Search section 106 Registration information DB 107 Call history information DB E Contact center environment

Claims

1. Based on either a regular expression of attention words representing words, phrases, clauses, phrases, or sentences that should be noted or paid attention to, or a detection pattern representing a character string indicating the attention words, when the detection pattern appears in the speech recognition result in the conversation among multiple persons and the scene or speech type is a predetermined scene or speech type, a detection unit that detects the attention words; A display information providing unit that, when the attention words are detected, transmits display information for displaying the detected attention words or display components representing the attention words in a pre-registered display mode to a predetermined terminal; A detection device having the above.

2. The detection unit is The detection device according to claim 1, wherein when the detection pattern appears a predetermined number of times or more in the speech recognition result, the attention words are detected.

3. The detection unit is The detection device according to claim 2, wherein the number of appearances is determined based on the appearance distribution of the detection pattern in past conversations.

4. The detection unit is The detection device according to any one of claims 1 to 3, wherein the attention words are detected based on the context until the detection pattern appears in the speech recognition result.

5. The display information providing unit is The detection device according to any one of claims 1 to 4, wherein display information for highlighting the detected attention words in a pre-registered color is transmitted to a first terminal used by a predetermined one of the speakers in the conversation.

6. The display information providing unit is The detection device according to claim 5, wherein display information for highlighting the display components in a pre-registered color is transmitted to a second terminal used by a person who monitors and supports the one speaker.

7. The display information providing unit is The detection device according to claim 6, wherein when the attention words are detected, information for notifying, in a pre-registered sound, that the attention words have been detected is transmitted to the second terminal.

8. It has a search unit that searches the conversation history information stored in a database representing the speech recognition results in past conversations using the attention words as search conditions, The display information providing unit is The detection device according to any one of claims 1 to 7, wherein display information for displaying the searched conversation history information and the attention words included in the speech recognition result represented by the conversation history information is transmitted to a third terminal that specified the search conditions.

9. A detection procedure for detecting a target word, based on a detection pattern representing either a regular expression of the target word representing a word, word, phrase, or sentence to be noted or a string representing the target word, when the detection pattern appears in the speech recognition result in the conversation among a plurality of persons and the scene or speech type is a predetermined scene or speech type; A display information providing procedure for transmitting, to a predetermined terminal, display information for causing the detected target word or a display component representing the target word to be displayed in a pre-registered display mode when the target word is detected; A detection method executed by a computer.

10. A program for causing a computer to function as the detection device according to any one of Claims 1 to 8.

Images