JP2018106551A - Conversation recording device, conversation recording method, and conversation recording program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a conversation recording device, a conversation recording method, and a conversation recording program that suppress a decrease in a service level caused by a system failure.SOLUTION: A conversation recording device 14 includes: an acquisition unit 32 that acquires voice data of a conversation transmitted and received via communication; and a data processing unit 34 that extracts extraction data from voice data on the basis of conditions prescribed for each item included in a system failure reporting scenario, and records the extraction data in associated with each item.SELECTED DRAWING: Figure 2

Description

  The present invention relates to a conversation recording apparatus, a conversation recording method, and a conversation recording program.

  In a data center, a plurality of virtual machines (Virtual Machines) cooperate to operate as a business system. In such a data center, dozens to hundreds of virtual machines may operate on one virtual infrastructure.

  The data center provides a service that maintains a state in which a system such as a physical machine or a virtual machine can continue to operate normally. Services provided by the data center include operations such as monitoring whether a system abnormality or sign has occurred, maintenance work such as updating the operation system, reporting when a failure occurs, and recovery. The report when a failure occurs includes, for example, a task of reporting a failure to a person in charge of a customer through a telephone call or the like when a system failure such as a failure occurs in a physical machine. In addition, the operation staff further includes a task of reporting the contents of the failure report to the customer staff to the manager in charge of the operation, for example, an advanced SE.

JP 2004-133527 A JP 2010-282545 A

  However, as described below, when a system failure occurs, there is a possibility that the operations in charge of operations cannot be maintained.

  In other words, with the development of the cloud business, the number of customers who are users of virtual machines has increased, and the multiplexing and scale-up of systems has also progressed. Along with this, if there is a case in which the number of users of virtual machines that have increased as a result of the system increase due to a system failure that occurred on a single physical machine, a failure report to all customers There is a risk that the period until will end will be prolonged. In this way, if the failure handling operation is delayed, there is a risk that the service provided by the data center will not function due to a problem caused by the system failure.

  In one aspect, an object of the present invention is to provide a conversation recording apparatus, a conversation recording method, and a conversation recording program that can suppress a decrease in service level due to a system failure.

  In one aspect, the conversation recording device is based on a condition defined for each item included in a system failure report scenario from an acquisition unit that acquires voice data of a conversation transmitted and received via communication, and the voice data A data processing unit that extracts extracted data and records the extracted data in association with each item.

  Service level degradation due to system failures can be suppressed.

FIG. 1 is a diagram illustrating the overall configuration of a conversation recording system according to an embodiment. FIG. 2 is a block diagram showing a functional configuration of the conversation recording apparatus. FIG. 3 is a diagram showing a specific example of the conversation recording process. FIG. 4 is a diagram for explaining extracted data patterns based on different conversations. FIG. 5 is a diagram for explaining extraction data editing processing by the data processing unit. FIG. 6 is a diagram illustrating an example of a report scenario after editing. FIG. 7 is a flowchart showing the procedure of the voice acquisition process. FIG. 8 is a flowchart showing a data processing procedure. FIG. 9 is a flowchart showing the procedure of the editing process. FIG. 10 is a diagram illustrating a hardware configuration of a computer that executes the conversation recording program according to the embodiment.

  A conversation recording device, a conversation recording method, and a conversation recording program according to the present application will be described below with reference to the accompanying drawings. Note that this embodiment does not limit the disclosed technology. Each embodiment can be appropriately combined within a range in which processing contents are not contradictory.

<Example>
FIG. 1 is a diagram illustrating the overall configuration of a conversation recording system 10 according to the embodiment. The conversation recording system 10 shown in FIG. 1 extracts extracted data from text data obtained by text-converting voice data of a call between a responsible SE and a customer SE according to conditions determined for each item included in a system failure report scenario. Provide a conversation recording service that records extracted data for each item.

  Here, “SE in charge” refers to a person in charge of a business related to a user (hereinafter referred to as a customer) who uses the cloud service among the parties operating the data center providing the cloud service. Further, “advanced SE” refers to an administrator who manages one or a plurality of responsible SEs among the parties operating the data center. Further, “customer SE” refers to a system engineer serving as a customer's contact.

  As shown in FIG. 1, the conversation recording system 10 includes a host terminal 12, one or more conversation recording devices 14, and one or more customer communication terminals 16.

  Among these, the upper terminal 12 and the conversation recording device 14 are communicably connected via a predetermined network 18. The network 18 may be any type of communication network such as the Internet (Internet), LAN (Local Area Network), and VPN (Virtual Private Network), regardless of wired or wireless. The conversation recording device 14 and the customer communication terminal 16 can be connected to various telephone lines such as a public telephone network, a mobile communication network, and a dedicated line (not shown).

  The upper terminal 12 is used by the advanced SE. For example, various computers such as a personal computer can be adopted as the upper terminal 12. For example, the upper terminal 12 can detect an abnormality such as a system failure, for example, a cloud service used by a customer, for example, a failure of a virtual machine such as a Web server or a mail server. For example, the upper terminal 12 receives an input for instructing execution of a task related to a failure report via a user interface (not shown) or the like. In accordance with this instruction input, as an example, a message or the like that prompts the execution of a business related to a failure report to the customer SE in charge of each responsible SE is displayed on the display unit 26 or the like of the conversation recording device 14. Note that the advanced SE may orally communicate a contact instruction to the SE in charge who uses the conversation recording device 14 without using the upper terminal 12. The host terminal 12 acquires extracted data obtained by extracting a part of the conversation between the responsible SE and the customer SE from the conversation recording device 14.

  The conversation recording device 14 is used by the responsible SE. The conversation recording device 14 includes an information processing device 20 and a communication unit 22.

  The information processing apparatus 20 and the communication unit 22 are connected so as to communicate with each other regardless of wired or wireless. Among these, the information processing apparatus 20 can employ various computers such as a personal computer. In addition to the fixed telephone that can be connected to the public telephone network, the communication unit 22 can be a mobile communication terminal that can be connected to the mobile communication network, such as various telephones such as a smartphone or a mobile phone. For example, the communication unit 22 accepts designation of a telephone number assigned to the customer communication terminal 16 from a responsible SE who performs a failure report in the event of a system failure. When a call is started between the communication unit 22 and the customer communication terminal 16 after the customer communication terminal 16 is called by specifying the telephone number, the information processing apparatus 20 records the call between the responsible SE and the customer SE. Start together. After that, when the call ends between the communication unit 22 and the customer communication terminal 16, the information processing apparatus 20 converts voice data obtained by recording the call into text data. Then, the information processing apparatus 20 extracts the extracted data according to the conditions defined for each item included in the system failure report scenario, and records the extracted data for each item.

  The customer communication terminal 16 is used by the customer SE. For example, the customer communication terminal 16 may be a fixed telephone that can be connected to a public telephone network, or a mobile communication terminal that can be connected to a mobile communication network, such as a smartphone or a mobile phone. The customer communication terminal 16 has a performance capable of being connected to the communication unit 22 of the conversation recording device 14 through a telephone line or the like. Thereby, the customer communication terminal 16 transmits / receives voice data to / from the communication unit 22 and realizes a call between the responsible SE and the customer SE.

  FIG. 2 is a functional block diagram showing a functional configuration of the conversation recording device 14. As illustrated in FIG. 2, the information processing apparatus 20 of the conversation recording apparatus 14 includes an input unit 24, a display unit 26, a storage unit 28, and a control unit 30.

  The input unit 24 is a device that receives input of a responsible SE such as a keyboard, a touch panel, and a mouse. For example, the input unit 24 receives data for editing the extracted data. The input unit 24 outputs the input received from the responsible SE to the control unit 30.

  The display unit 26 is a display device that can display an image, such as a liquid crystal display device. The display unit 26 displays an image based on the image data acquired from the control unit 30. For example, the display unit 26 displays the extracted data received from the control unit 30.

  The storage unit 28 can be implemented by an auxiliary storage device such as an HDD (Hard Disk Drive) and an SSD (Solid State Drive). The storage unit 28 does not necessarily have to be provided in the information processing apparatus 20, and can also be implemented as a storage device on the network, for example, a file server. In addition, the storage unit 28 may be mounted as a removable medium and connected via a port such as a connector or a terminal.

  The storage unit 28 is generated by a program executed by the control unit 30, for example, an OS (Operating System) and a conversation recording program for realizing the conversation recording service, data used for executing the program, and execution of the program. Store the data. For example, the storage unit 28 stores word dictionary data 36, spoken word dictionary data 38, and a report scenario 42 including a plurality of extracted data 40. In addition to the word dictionary data 36, the spoken word dictionary data 38, and the report scenario 42, the storage unit 28 can store other electronic data, for example, voice data acquired by the acquisition unit 32.

  The word dictionary data 36 is dictionary data in which keywords are registered in advance. The word dictionary data 36 is, for example, a word associated with an item, and includes text data words such as the name of the customer SE and the name of the system handled by the responsible SE. Word dictionary data 36 associates words with items. The word dictionary data 36 may include the name of the customer SE associated with the phone number of the customer SE.

  The spoken word dictionary data 38 is dictionary data in which one or more spoken words that are human spoken words are registered in advance. The spoken word dictionary data 38 includes, for example, a spoken word as text data. Here, as the spoken language, for example, a fixed sentence described in the conversation manual of the responsible SE can be adopted. Furthermore, the spoken word dictionary data 38 associates spoken words with items and processes. An example of the processing is setting a flag such as “present” or “absent”, or extracting a predetermined time or a predetermined number of words before or after a spoken word from text data.

  The report scenario 42 is data relating to a scenario in which the responsible SE reports to the customer SE when a system failure occurs. This report scenario 42 includes items such as “conversation start date / time”, “system name”, and “customer name” as items used to identify a failure report, and system failures are used for business in the system. This includes items such as “business impact” given to the virtual machine to be processed and “additional questions” that require a post response to the customer after the call ends. For each of these “Conversation start date / time”, “System name”, “Customer name”, “Business impact”, and “Additional questions”, characters corresponding to each item from the text data obtained by converting the voice data of the call into text A column is extracted as extracted data 40.

  The control unit 30 can be implemented by a hardware processor such as a CPU (Central Processing Unit) or an MPU (Micro Processing Unit). Here, the CPU and the MPU are illustrated as an example of the processor. However, the processor and the MPU can be mounted by any processor regardless of a general-purpose type or a specialized type. In addition, the control unit 15 may be realized by a hard wired logic such as an application specific integrated circuit (ASIC) or a field programmable gate array (FPGA).

  The control unit 30 develops the conversation recording program as a process on a work area of a RAM such as a DRAM (Dynamic Random Access Memory) or SRAM (Static Random Access Memory) mounted as a main storage (not shown). The functions of the acquisition unit 32 and the data processing unit 34 shown in FIG. 1 are virtually realized.

  The acquisition unit 32 acquires voice data of a conversation (for example, a call) transmitted / received via communication from the communication unit 22. The acquisition unit 32 records (that is, records) the acquired audio data in the storage unit 28. The acquisition unit 32 may store the audio data in a work area of an internal memory (not shown) that is referred to by the data processing unit 34 without saving the audio data in the storage unit 28.

  The data processing unit 34 acquires the audio data acquired by the acquisition unit 32 from the storage unit 28. Note that the data processing unit 34 may directly acquire audio data from the acquisition unit 32. The data processing unit 34 converts the voice data acquired from the acquisition unit 32 into text and generates text data. The data processing unit 34 extracts a character string relating to a conversation such as failure or maintenance as extracted data 40 for each item in accordance with the conditions defined for each item included in the system failure report scenario from the text data. For example, the data processing unit 34 may extract a character string that matches a keyword registered in the word dictionary data 36 from the text data as the extracted data 40. In addition, the data processing unit 34 may extract a character string corresponding to a predetermined number of words or a predetermined time before or after a spoken word registered in the spoken word dictionary data 38 as extracted data 40 from the text data. After that, the data processing unit 34 records the extracted data 40 extracted from the text data in the report scenario 42 in the storage unit 28 in association with the item. The data processing unit 34 may store the report scenario 42 in an external storage device, for example, a storage device on the network or the higher order terminal 12 of the advanced SE.

  The data processing unit 34 can display a report scenario 42 including the extracted data 40 extracted for each item as an editing screen, and can accept editing of a character string described in association with each item. For example, the data processing unit 34 may accept editing content from the responsible SE via the input unit 24. As a result of this editing, a report of a failure report to the customer submitted by the SE in charge to the senior SE is generated.

  FIG. 3 is a diagram for explaining a specific example of the conversation recording process.

  As shown in the upper left of FIG. 3, when the responsible SE starts a call with the customer SE by connecting to the customer communication terminal 16 via the communication unit 22, the acquisition unit 32 of the information processing apparatus 20 Recording of voice data exchanged between the terminal 22 and the customer communication terminal 16 is started. Thereby, the voice data of the call between the responsible SE and the customer SE is recorded. Thereafter, when the call between the communication unit 22 and the customer communication terminal 16 ends, the acquisition unit 32 stores the voice data of the call in the storage unit 28 or outputs it to the data processing unit 34. Note that the acquisition unit 32 indicates the date and time at the time when the communication unit 22 makes a call to the customer communication terminal 16 or when a call is started between the communication unit 22 and the customer communication terminal 16. Output to.

  After the call between the responsible SE and the customer SE shown in the lower left of FIG. 3 is completed, the data processing unit 34 converts the voice data of the conversation CV1 acquired by the acquisition unit 32 into text. Then, the data processing unit 34 extracts the extracted data 40 for each item in accordance with the conditions defined for each item included in the system failure report scenario from the text data obtained by converting the voice data of the conversation CV1 into text. For example, the above items include “conversation start date / time”, “system name”, “customer name”, “conversation content”, “conversation content (individual)”, and the like. Here, the case where the voice data is converted into text after the call ends is illustrated, but the timing of the text conversion is not limited to after the call ends. For example, text conversion can be performed every time voice data is acquired.

  The data processing unit 34 displays the date and time “yyyy-mm-dd hh: mm: ss” when the communication unit 22 made a call to the customer communication terminal 16 or when the call started between the communication unit 22 and the customer communication terminal 16. Is extracted as extracted data 40A. Then, the data processing unit 34 registers the extracted data 40 </ b> A in association with the item “conversation start date” among the items included in the report scenario 42.

  The data processing unit 34 searches the text data of the conversation CV1 for the name “system name” of the system in which the system failure has occurred among the items included in the report scenario 42. In this example, since the keyword “yy system” registered in the word dictionary data 36 in association with the item “system name” matches the character string “yy system” included in the text data of the conversation CV1, the character string “Yy system” is extracted as the extracted data 40B. As a result, the data processing unit 34 registers the extracted data 40B “yy system” in association with the item “system name” among the items included in the report scenario 42.

  The data processing unit 34 searches the text data of the conversation CV1 for the name “customer charge name” of the customer SE who is the partner of the conversation CV1 among the items included in the report scenario 42. In this example, since the keyword “zz” registered in the word dictionary data 36 in association with the item “customer name” matches the character string “zz” included in the text data of the conversation CV1, the character string “zz” is matched. zz ″ is extracted as the extracted data 40C. As a result, the data processing unit 34 registers the extracted data 40C “zz” in association with the item “customer charge name” among the items included in the report scenario 42.

  Further, the data processing unit 34, among the items included in the report scenario 42, has a fixed phrase spoken word (for example, “business impact shown in the conversation CV1”) associated with the item “conversation content” and registered in the spoken word dictionary data 38. If there is a conversation that matches "", execute processing associated with the spoken word, for example, a predetermined time after the conversation or a predetermined number of words, for example, immediately after the conversation data The content spoken by the customer SE is extracted and registered in the extracted data 40D of the item “conversation content”. For example, in the example of FIG. 3, an utterance made by the customer SE immediately after the responsible SE uttered the spoken word “Tell me about the business impact”, for example, one sentence “I am in trouble with the occurrence of xx”. After being extracted as extracted data 40D, it is stored in association with the conversation content item.

  In addition, the data processing unit 34 includes, among items included in the report scenario 42, a specific spoken word (for example, “shown in the conversation CV 2” associated with the item “conversation content (individual)” and registered in the spoken word dictionary data 38. If there is a conversation that matches ")", the process associated with the spoken word is executed. For example, before and after the conversation (for example, the number of words corresponding to one minute or one minute before and after the conversation) The contents are extracted and registered in the extracted data 40E of the item “conversation contents (individual)”. For example, in the example of FIG. 3, the utterance made by the customer SE immediately before the utterance “I will give you another report” by the responsible SE, for example, one sentence “Is there a relationship with the failure that occurred last time? "Is extracted as the extracted data 40E and stored in association with the item of conversation content (individual).

  After the extracted data 40A to 40E of each item included in the report scenario 42 is recorded in this way, the data processing unit 34, as shown in the lower right of FIG. An editing screen for editing the report of the trouble report to the customer to be submitted is displayed on the display unit 26. Subsequently, the data processing unit 34 receives, via the input unit 24, editing of the extracted data 40 </ b> A to 40 </ b> E associated with the items included in the editing screen into the character string. Thereafter, when the editing on the editing screen is completed, the data processing unit 34 saves the edited report scenario in the storage unit 28 as a failure report and outputs it to the host terminal 12. Thereby, the senior SE can grasp the contents of the conversation between the responsible SE and the customer SE even if the result of the failure report is not reported verbally from the responsible SE.

  FIG. 4 is a diagram for explaining extracted data patterns based on different conversations.

  As shown in FIG. 4, after the responsible SE and the customer SE have spoken the conversation CV3, the spoken words registered in the spoken word dictionary data 38 have “business influence”, “no business influence”, and , “If there is a conversation that includes a separate confirmation and report,” the data processing unit 34 determines whether the item “work impact” or “additional question” ”Is further registered, and the report scenario 42 is classified into a plurality of failure report patterns.

  Specifically, when the responsible SE and the customer SE speak the conversation CV3 and then speak the conversation CV3a, the data processing unit 34 registers the spoken word “there was no business influence” registered in the spoken word dictionary data 38. Search for and extract. In this case, the data processing unit 34 registers the extracted data 40 of “no business effect” and “no additional question” in the item “business effect presence / absence” and the item “additional question presence / absence”.

  When the responsible SE and the customer SE speak the conversation CV3 and then speak the conversation CV3b, the data processing unit 34 searches and extracts the spoken word “there is business influence” registered in the spoken word dictionary data 38. . In this case, the data processing unit 34 registers the extracted data 40 of “with business impact” and “without additional question” in the item “presence / absence of business impact” and the item “presence / absence of additional question”.

  When the responsible SE and the customer SE speak the conversation CV3 and then speak the conversation CV3c, the data processing unit 34 reports the spoken word registered in the spoken word dictionary data 38 after confirming separately. I will search and extract "". In this case, the data processing unit 34 registers the extracted data 40 of “with additional questions” in the item “additional question presence / absence”. In the conversation CV3c, the data processing unit 34 registers “with business impact” or “without business impact” according to the content of “business impact is ...”. As described above, the data processing unit 34 classifies each report scenario 42 into a plurality of patterns (for example, four patterns) depending on the presence / absence of business influence and the presence / absence of additional questions.

  FIG. 5 is a diagram for explaining editing processing of the extracted data 40 by the data processing unit 34.

  The items of the extracted data 40 shown in FIG. 5 further include a “business impact presence / absence” flag and an “additional question presence / absence” flag for classifying patterns. As described with reference to FIG. 4, the data processing unit 34 registers “present” (for example, “1”) in the extracted data 40 of the item “business influence presence / absence” in the case of the pattern “with business influence”. In the case of the “no business impact” pattern, “none” (for example, “0”) is registered in the extracted data 40 of the item “business impact presence / absence”.

  In the case of the “with additional question” pattern, the data processing unit 34 registers “present” (for example, “1”) in the extracted data 40 of the item “additional question presence / absence” and sets the pattern of “without additional question”. In this case, “none” (for example, “0”) is registered in the extracted data 40 of the item “existing questions”.

  A case will be described in which the responsible SE and the customer SE have spoken the content indicated by the conversation CV4 in the lower column of FIG. In this case, the data processing unit 34 acquires the conversation start date and time acquired by the acquisition unit 32 and the voice data of the conversation CV4. The data processing unit 34 registers the acquired conversation start date and time in the extracted data 40 of the item “conversation start date and time”. Further, the data processing unit 34 registers “in-house public Web” extracted from the conversation CV4 in accordance with the above-described processing in the extracted data 40 of the item “system name”, and sets “xxx Taro” as the item “customer contact name”. Registered in the extracted data 40, and registered "xx server could not be displayed" in the extracted data 40 of the item "business impact details" and "I am sorry ... I will report". Registered in the extracted data 40 of the item “additional question details”. The item “business impact details” is an example of the item “conversation content”. The item “additional question details” is an example of the item “conversation content (individual)”.

  Further, since the current conversation CV4 includes the business impact details and the additional question details, the data processing unit 34 registers “Yes” in the item “Additional question presence / absence” and “Yes” in the item “Business impact presence / absence”. Register.

  The data processing unit 34 causes the display unit 26 to display an editing screen 44 including the extracted data 40 indicated by “Before editing” in the upper part of FIG. Here, the data processing unit 34 generates an edit screen 44 including an edit start button 46 and an edit completion button 48 together with a login name indicating the name of the responsible SE.

  When the responsible SE operates the input unit 24 and selects (for example, clicks) the edit start button 46, the data processing unit 34 causes the display unit 26 to display the edit screen 44 that has been changed for acceptance, and displays each item. Start accepting edits for editing. On the editing screen 44 for acceptance, the data processing unit 34 displays, for example, “Yes” and “None” in the items “presence / absence of business impact” and “presence / absence of additional question”.

  In a state where the accepting edit screen 44 is displayed, the data processing unit 34 accepts, for example, selection of a cell of the extracted data 40 selected (for example, clicked) by the SE in charge by operating the input unit 24. As a result, the data processing unit 34 accepts editing related to the extracted data of the cell. For example, the data processing unit 34 accepts deletion of a part of the text data indicating the editing content input by the responsible SE by operating the input unit 24 or a part of the extracted data 40. In the example indicated by “Editing” in the second row from the top of FIG. 5, the data processing unit 34 accepts deletion of a part of the extracted data 40 of the items “work impact details” and “additional question details”. The extracted data 40 is edited.

  When the data processing unit 34 receives the selection of the edit completion button 48 input by the responsible SE by operating the input unit 24, the data processing unit 34 displays the edit screen 44 shown in “after editing” on the third row from the top of FIG. 5. To display.

  FIG. 6 is a diagram showing an example of the report scenario 42 after editing. The data processing unit 34 stores the report scenario 42 of the set of the extracted data 40 after editing shown in FIG.

  FIG. 7 is a flowchart showing the procedure of the voice acquisition process. In the voice acquisition process shown in FIG. 7, as an example, the communication unit 22 control unit 30 executes the voice acquisition process by reading a conversation recording program.

  In the voice acquisition process shown in FIG. 7, when the input of the telephone number of the customer communication terminal 16 used by the customer SE is received by the communication unit 22, the acquisition unit 32 transmits the call to the customer communication terminal 16 of the customer SE and stores it. Recording of the voice data of the call to the unit 28 is started, and the start date and time of the call among the items included in the report scenario 42 is registered in the storage unit 28 (S102). Thereafter, the acquisition unit 32 causes the display unit 26 to display a screen notifying that the recording of the call has started (S104).

  Subsequently, the acquisition unit 32 arrives at the customer communication terminal 16 and starts a call between the communication unit 22 and the customer communication terminal 16 (S106). At this stage, the conversation between the responsible SE and the customer SE becomes possible. And the acquisition part 32 will complete | finish the recording of the audio | voice data with respect to the memory | storage part 28, if the telephone call of the communication part 22 and the customer communication terminal 16 is complete | finished (S108).

  Thereafter, the data processing unit 34 converts the text data obtained by text-converting the voice data stored in the storage unit 28 from the last extraction time from the end of the conversation into the first variable (for example, the variable name “1 minute”). Is stored in the storage unit 28 (S110). For example, the extraction time is a preset time and is, for example, an approximate number of words corresponding to one minute or one minute.

  Further, the data processing unit 34 stores the text data obtained by text-converting all the voice data stored in the storage unit 28 as the second variable (for example, variable name “full text”) in the storage unit 28 (S112).

  Thereby, the voice acquisition process is terminated.

  FIG. 8 is a flowchart showing a data processing procedure. The data processing shown in FIG. 8 starts as an example when text data converted from voice data of a call is obtained by the voice acquisition processing shown in FIG.

  As shown in FIG. 8, the data processing unit 34 extracts the date and time when a call is made from the communication unit 22 to the customer communication terminal 16 as the extracted data 40, and the extracted data 40 is the item included in the report scenario 42. Of these, registration is made in association with the item “conversation start date” (S202).

  Subsequently, the data processing unit 34 searches for the keyword registered in the word dictionary data 36 in association with the item “system name” from the text data obtained by converting the voice data of the call in step S112 into text (step S204).

  At this time, if there is a hit keyword in the keyword of the system name (Yes in step S206), it is determined which of the systems operating in the data center is related to the failure report. In this case, the data processing unit 34 extracts a character string corresponding to the keyword of the system name hit in the search as the extracted data 40, and records the extracted data 40 in association with the item “system name” of the report scenario 42 ( Step S208). If there is no hit keyword in the system name keyword (No in step S206), the process of step S208 is skipped, and the process proceeds to the following step S210.

  Subsequently, the data processing unit 34 searches for the keyword registered in the word dictionary data 36 in association with the item “customer name” from the text data obtained by converting the voice data of the call into text (step S210).

  Then, if there is a hit keyword in the keyword of the system name (Yes in step S212), it is determined which customer SE is the target of the customer SE for which the failure report is handled by the responsible SE. In this case, the data processing unit 34 extracts a character string corresponding to the keyword of the name of the customer SE hit in the search as the extracted data 40, and associates the extracted data 40 with the item “customer charge name” of the report scenario 42. Recording is performed (step S214). On the other hand, if there is no hit keyword in the customer name keyword (No in step S212), the process of step S214 is skipped, and the process proceeds to the following step S216. The data processing unit 34 extracts the customer contact name when the affirmative spoken word registered in the spoken word dictionary data 38 such as “Yes, yes” follows the extracted customer contact name. Also good.

  Thereafter, the data processing unit 34, from the text data obtained by converting the voice data of the call into the text data, is associated with the item “work presence / absence” and is registered in the spoken word dictionary data 38 and corresponds to the business word having the business influence, for example, a trouble report. A character string that matches or resembles the phrase “There is business impact” in the fixed phrase described in the manual is searched (step S218). In the search here, the fixed phrase is not necessarily uttered as it is in the conversation, so it is not always necessary to make a perfect match as a condition. For example, if the phrase of a fixed phrase is “work-affected”, the words “has a business influence” or “has a business influence” with the same meaning are included in the search results as words similar to the phrase of the fixed phrase be able to.

  At this time, if a character string that matches or is similar to the spoken word corresponding to the presence of business impact does not hit (No in step S218), the data processing unit 34 associates the flag “none” with the item “business impact presence / absence” of the report scenario 42. Recording is performed (step S220).

  On the other hand, if a character string that matches or resembles a spoken word corresponding to the presence of business influence is found (Yes in step S218), the boilerplate provides a description of the business impact from the customer SE to the responsible SE and the boilerplate specified in the trouble report manual You can see that you are repeating. In this case, the data processing unit 34 records the flag “present” in association with the item “business operation presence / absence” of the report scenario 42 (step S222). Furthermore, the data processing unit 34 adds a predetermined number of words following the spoken word “has a job influence” corresponding to “has a job influence” to the item “job influence details” of the report scenario 42, in the example of FIG. "It means that you could not access the public web site for a minute" and record it in association (step S224).

  In the above description of step S216, as an example, the case where “there is business impact” is used as the spoken word corresponding to the presence of business impact. However, in order to more accurately extract the business impact from text data, other spoken words are used. Can also be used. For example, the content of the business influence uttered by the customer SE or the spoken phrase of the fixed phrase that the summary of which is replayed by the responsible SE can be registered in the spoken word dictionary data 38 as a speech corresponding to the presence of business influence. For example, a text string that matches or is similar to both of the two spoken words “I understand” and “That means” may be searched from text data. By using these two spoken words, the place where the utterance was made in the section starting from the first spoken word "I knew" and ending with the second spoken word "Is that?" Could not be accessed "can be extracted as business impact details.

  Further, in the description of step S218, the case where No is determined in the branch of step S218 when a character string that matches or is similar to the spoken word corresponding to the presence of business influence is not exemplified, but the present invention is not limited to this example. That is, when a character string that matches or is similar to the corresponding spoken word without any business influence is hit, it may be determined as No in the branch of step S218. For example, in addition to the spoken word “no business influence” corresponding to no business influence, similar words such as “no business influence” and “no business influence” may be searched from text data.

  Thereafter, the data processing unit 34, from the text data in which the voice data of the call is converted into text, the spoken word corresponding to the presence of an additional question associated with the item “additional question presence / absence” and registered in the spoken word dictionary data 38, for example, a trouble report Search for a character string that matches or is similar to the phrase of the fixed phrase described in the manual of “(I will report to you separately)” (step S226). In this search as well, the fixed phrase is not necessarily uttered as it is in the conversation, so it is not always necessary to make a perfect match as a condition.

  At this time, if the character string that matches or is similar to the spoken word corresponding to the presence of the additional question does not hit (No in step S226), the data processing unit 34 associates the flag “none” with the item “additional question presence / absence” of the report scenario 42. Recording is performed (step S228).

  On the other hand, when a character string that matches or resembles a spoken word corresponding to the presence of an additional question is hit (Yes in step S226), a standard that is defined in the trouble report manual is a promise to make a subsequent response to the additional question from the customer SE to the responsible SE. You can see that the SE in charge confirms the phrase. In this case, the data processing unit 34 records the flag “present” in association with the item “presence / absence of additional question” of the report scenario 42 (step S230). In addition, the data processing unit 34 sets the item “additional question details” of the report scenario 42 to a predetermined number of words before the spoken word “I will report after confirming separately” corresponding to the presence of the additional question, the example of FIG. In other words, one sentence of the utterance made by the customer SE immediately before “Is there any relevance to the failure that occurred last time?” Is recorded in association (step S232).

  Thereafter, the data processing unit 34 edits the report of the failure report to the customer submitted by the responsible SE to the senior SE according to the report scenario 42 stored in the storage unit 28, for example, the editing shown in the uppermost part of FIG. The screen 44 is displayed on the display unit 26 (step S234).

  After displaying the editing screen in this way, the data processing is terminated.

  FIG. 9 is a flowchart showing the procedure of the editing process. For example, the editing process illustrated in FIG. 9 is started when the editing screen is displayed in step S234.

  As shown in FIG. 9, the data processing unit 34 determines whether or not the responsible SE has operated the edit start button 46 (S302). For example, the data processing unit 34 receives an input of the responsible SE via the input unit 24 and determines whether or not the edit start button 46 has been operated. The data processing unit 34 is in a standby state until the operation of the edit start button 46 is accepted (S302: No).

  When the data processing unit 34 receives the operation of the editing start button 46 (S302: Yes), the data processing unit 34 displays the editing screen 44 for receiving shown in the second row from the top of FIG. Is started (S304).

  If the data processing unit 34 determines that the edited content by the responsible SE has not been received (S304: No), the data processing unit 34 determines whether the responsible SE has operated the editing completion button 48 (S308). For example, the data processing unit 34 receives the input of the responsible SE via the input unit 24 and determines whether or not the edit completion button 48 has been operated. If the data processing unit 34 determines that the edit completion button 48 has not been operated (S308: No), it repeats step S304 and subsequent steps. In other words, the data processing unit 34 continues to accept editing contents until an operation of the editing completion button 48 is accepted.

  For example, when the responsible SE operates the input unit 24 to select the extracted data 40 of any item and inputs the text data or the presence / absence of the extracted data 40, the data processing unit 34 receives the editing content. Determination is made (S304: Yes), and the extracted data 40 is edited and newly registered with the received editing content (S306).

  When receiving the operation of the edit completion button 48 by the responsible SE (S308: Yes), the data processing unit 34 notifies the upper terminal 12 of the advanced SE of the update (S310).

[One aspect of effect]
As described above, the conversation recording device 14 is based on conditions set in advance from text data obtained by text conversion of conversation between the responsible SE and the customer SE, for example, data registered in the word dictionary data 36 and the spoken word dictionary data 38. Thus, the extraction data 40 is extracted. Thereby, the conversation recording apparatus 14 can reduce the work load of the responsible SE for extracting necessary data from the conversation in which everything is recorded. For example, the conversation recording device 14 can reduce the response time per customer SE of the responsible SE (for example, from 4 minutes to 2 minutes) and reduce the time required for reporting to the advanced SE (for example, 3 minutes to 1). Half the minute). As a result, the conversation recording device 14 can significantly reduce the time required for the SE in charge of handling a system failure (for example, 0.02 man-days for 100 systems). Further, the conversation recording device 14 can realize an immediate response to the customer SE by the responsible SE, and can realize an appropriate response according to the recovery situation by making the failure response instruction issued by the advanced SE closer to real time. In this way, it is possible to suppress the failure handling operation from being delayed, and it is also possible to suppress the situation where the service provided by the data center does not function due to a problem caused by the system failure. Therefore, according to the conversation recording device 14, it is possible to suppress a decrease in service level due to a system failure.

  Since the conversation recording device 14 extracts a part of all conversations as the extracted data 40, the necessary storage capacity can be reduced.

  The conversation recording apparatus 14 registers the extracted data 40 of the text data obtained by converting the conversation of the voice data into text, so that it can be output in a display format that can be easily understood by the responsible SE and advanced SE.

  Since the conversation recording device 14 can cause the responsible SE to edit the extracted data 40 while viewing the text data of the voice data on the editing screen 44, it is higher for the responsible SE than when inputting the text data while talking. Does not require typing skills.

  The functions, connection relationships, and numbers of the configurations of the above-described embodiments may be changed as appropriate.

  For example, in the above-described embodiment, the case where various kinds of processing, for example, extraction of extracted data, is performed on text data obtained by converting voice data of a call into text is illustrated. You don't have to convert. That is, it is possible to recognize a keyword registered in the word dictionary data and a spoken word registered in the spoken word dictionary data through voice recognition such as word spotting from the voice data of the call. For example, when word dictionary data is referred to, a keyword recognized by voice recognition may be extracted as extracted data as it is, and when word dictionary data is referred to, it is recognized by voice recognition from voice data of a call. Partial voice data corresponding to a predetermined number of words or a predetermined time before or after a spoken word or a predetermined time can be extracted as extracted data, or a character string obtained by converting the partial voice data into text can be extracted as extracted data. The partial speech data or character string extracted in this way can be associated with an item. Also, when partial audio data is associated with an item, the advanced SE can extract and listen to partial audio data out of audio data, so when reading a character string associated with an item or a character string edited from it As with, can understand the contents of business impact and additional questions in a short time.

  For example, in the above-described embodiment, the description has been made on the premise of the conversation corresponding to the failure. However, the above-described processing may be applied to other correspondence, for example, a conversation related to maintenance.

  In the above-described embodiment, the information processing device 20 of the responsible SE is provided with the functions of the acquisition unit 32 and the data processing unit 34. However, the present invention is not limited to this. For example, the functions of the acquisition unit 32 and the data processing unit 34 may be provided in the upper terminal 12 or another device.

[Conversation recording program]
The various processes described in the above embodiments can be realized by executing a prepared program on a computer such as a personal computer or a workstation. In the following, an example of a computer that executes a conversation recording program having the same function as that of the above embodiment will be described with reference to FIG.

  FIG. 10 is a diagram illustrating a hardware configuration example of a computer that executes the conversation recording program according to the embodiment. As shown in FIG. 10, the computer 50 includes a CPU 52, a ROM 54, an HDD 56, and a RAM 58. The CPU 52, ROM 54, HDD 56, and RAM 58 are connected via the bus 60.

  As shown in FIG. 10, the HDD 56 stores a conversation recording program 62 that exhibits the same functions as those of the acquisition unit 32 and the data processing unit 34 described in the above embodiment. The conversation recording program 62 may be integrated or separated as with each component of the acquisition unit 32 and the data processing unit 34 shown in FIG. That is, the HDD 56 does not necessarily have to store all the data shown in the above-described embodiment, and data used for processing may be stored in the HDD 56.

  Under such an environment, the CPU 52 reads the conversation recording program 62 from the HDD 56 and develops it in the RAM 58. As a result, the conversation recording program 62 functions as a conversation recording process 64 as shown in FIG. The conversation recording process 64 expands various data read from the HDD 56 in an area allocated to the conversation recording process 64 in the storage area of the RAM 58, and executes various processes using the expanded various data. For example, as an example of processing executed by the conversation recording process 64, processing shown in FIGS. In the CPU 52, not all the processing units shown in the above embodiments need to operate, and a processing unit corresponding to a process to be executed may be virtually realized.

  Note that the conversation recording program 62 is not necessarily stored in the HDD 56 or the ROM 54 from the beginning. For example, the conversation recording program 62 is stored in a “portable physical medium” such as a flexible disk inserted into the computer 50, so-called FD, CD-ROM, DVD disk, magneto-optical disk, IC card or the like. Then, the computer 50 may acquire and execute the conversation recording program 62 from these portable physical media. Further, the conversation recording program 62 is stored in another computer or a server device connected to the computer 50 via a public line, the Internet, a LAN, a WAN, etc., and the computer 50 acquires the conversation recording program 62 from these. May be executed.

14 ... Conversation recording device 22 ... Communication unit 32 ... Acquisition unit 34 ... Data processing unit 36 ... Word dictionary data 38 ... Spoken word dictionary data 40 ... Extracted data 42 ... Report scenario 44 ... Editing screen 50 ... Computer

Claims (9)

An acquisition unit for acquiring voice data of conversations transmitted and received via communication;
A data processing unit that extracts extracted data based on conditions defined for each item included in a system failure report scenario from the voice data, and records the extracted data in association with each item;
A conversation recording apparatus characterized by comprising:   2. The item according to claim 1, wherein the item includes at least one of a system name in which the system failure occurs, a customer contact name to which the system failure is reported, or a combination thereof. Conversation recording device.   The data processing unit refers to word dictionary data in which a keyword corresponding to the system name or the customer contact name is registered, and extracts the keyword from the voice data as the extracted data. Item 3. The conversation recording device according to Item 2.   The items include the presence / absence of the impact of the system failure on the virtual machine used in the system, the details of the impact, and the subsequent to the customer who is the report destination of the system failure after the call made via the communication is terminated The conversation recording apparatus according to claim 1, 2 or 3, characterized in that it includes the presence / absence of a question requiring an answer, the content of the question, or a combination thereof.   The data processing unit refers to spoken word dictionary data in which a spoken word representing a fixed phrase for prompting an answer to the influence or a fixed phrase indicating the presence of the question is registered, and the spoken word is present in the speech data 5. The conversation recording according to claim 4, wherein partial speech data or a character string corresponding to the partial speech data corresponding to a predetermined number of words or a predetermined time before or after the spoken word is extracted as the extracted data. apparatus.   The data processing unit refers to spoken word dictionary data in which a spoken word indicating the presence or absence of the influence or the presence or absence of the question is registered, and the spoken word corresponding to the presence of the influence or the presence of the question is included in the voice data. If present, records the presence / absence of the influence or the presence / absence of the question in association with information indicating the presence of the influence or the presence of the question, and corresponds to a predetermined number of words or a predetermined time before or after the spoken word 5. The conversation recording apparatus according to claim 4, wherein partial voice data or a character string corresponding to the partial voice data is extracted and recorded in association with the content of the influence or the content of the question.   When there is no spoken word corresponding to the presence of the influence or the presence of the question in the voice data, the data processing unit includes the absence of the influence or the question in the presence / absence of the influence or the presence / absence of the question. The conversation recording apparatus according to claim 6, which records information indicating the absence or absence in association with each other. Obtain voice data of conversations sent and received via communication,
Extracting extracted data from the voice data based on conditions defined for each item included in the system failure report scenario, and recording the extracted data in association with each item,
A conversation recording method, wherein a computer executes a process. Obtain voice data of conversations sent and received via communication,
Extracting extracted data from the voice data based on conditions defined for each item included in the system failure report scenario, and recording the extracted data in association with each item,
A conversation recording program characterized by causing a computer to execute processing.

Images