JP2019185191A - Faq support device, faq support system, faq support method and faq support program - Google Patents

Abstract

To support providing an appropriate answer to a question from a user.SOLUTION: A FAQ support device 100 for answering a question from a client 105 operated by a user U preliminarily holds a plurality of words of questions as assumed questions from the client 105 and answers relating to the questions in a FAQ data file 121 as FAQ information and comprises a bot 123 which performs search processing of, in response to a question from the client 105, presenting question candidates resulting from narrowing-down of pertinent FAQ information to the client 105 on the basis of an operation of selecting a category or a tag relating to the FAQ information.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a FAQ support apparatus, a FAQ support system, a FAQ support method, and a FAQ support program that support a work of FAQ (Frequency Asked Question).

  For example, in a FAQ operation that accepts a user's question (inquiry) and automatically answers it, a server in the FAQ operation center analyzes a text message input by the user and returns a corresponding answer to the user. Conventionally, there is a technique that improves search efficiency by performing word segmentation on a user input keyword and dividing it into a plurality of keywords, and then performing a search based on the keyword selected by the user (for example, the following patent document) 1).

  In addition, a chat service that enables conversations between friends and groups using a mobile device such as a smartphone such as LINE (registered trademark) has been provided, and simpler message transmission has become possible (for example, see Patent Document 2 below). .)

Special table 2018-503917 gazette JP 2014-146380 A

  However, in text search, it is necessary for the user himself / herself to input a keyword corresponding to the content of the inquiry, and if the inquiry keyword input by the user is not appropriate, the user cannot obtain a desired answer. there were. In order for a user to obtain an appropriate answer, the user needs to ponder a plurality of keywords, and an appropriate answer cannot be obtained as the user's search skill is lower. Further, the conventional chat service exchanges messages between users, and it has been difficult to use for business that requires immediacy.

  For this reason, the chat service cannot simply be applied to the FAQ business center, and neither the user side nor the answering side (service side) can perform an appropriate exchange in the conventional FAQ business. On the user side, an appropriate answer to the originally desired inquiry cannot be quickly obtained. Also, the service side cannot provide an appropriate answer to the inquiry requested by the user, and the accuracy and speed in the FAQ work cannot be improved.

  In view of the above problems, an object of the present invention is to support the provision of an appropriate answer to a user's inquiry.

  In order to achieve the above object, the FAQ support apparatus according to the present invention is a FAQ support apparatus that answers a client's inquiry, and includes a plurality of inquiries in advance as FAQ information regarding an inquiry that assumes an inquiry from the client and an answer related to the inquiry. And a control unit for performing search processing for presenting inquiry candidates narrowed down to the corresponding FAQ information to the client based on a selection operation of a category or tag related to the FAQ information at the time of inquiry of the client It is characterized by that.

  Further, the control unit holds a FAQ data file in which the category and the tag are associated with the FAQ information as a FAQ database, presents the category or the tag candidate to the client, and the client performs the The search process is performed by selecting a category or the tag.

  The FAQ database holds the categories in a plurality of hierarchies, presents the categories in a plurality of hierarchies to the client, and performs the search process by a selection operation of the categories in the hierarchies by the client. To do.

  The control unit may generate a tag from a character string included in the FAQ information and add the tag to the FAQ data file.

  The control unit may add a character string selected by the client to the FAQ data file as a new tag for the character string included in the FAQ information presented to the client.

  In addition, the control unit further transmits / receives information about an inquiry from the client and information on the search process to the client in an interactive form by chat.

  In addition, the control unit transmits a plurality of stamps corresponding to the number of candidates, and performs the search process based on the selection of the stamps by the client. It is characterized by that.

  The FAQ support system of the present invention is a FAQ support system including a chat server that answers a client inquiry and a FAQ server that performs a search process based on the client inquiry. A plurality of inquiries and inquiries that are related to the inquiry are stored in advance as FAQ information, and the relevant FAQ information is selected based on a selection operation of a category or tag related to the FAQ information at the time of inquiry of the client. The chat server performs a search process for presenting inquiry candidates narrowed down to the client, and the chat server sends information related to the inquiry from the client and information on the search process to the client in an interactive manner by chat. To signal, characterized in that.

  Further, the chat server and the FAQ server functions are arranged on a cloud.

  In the FAQ support method of the present invention, in the FAQ support method for answering a client inquiry, the computer holds in advance a plurality of inquiry information that is assumed to be an inquiry from the client and a reply word related to the inquiry as FAQ information. Performing a search process for presenting inquiry candidates narrowed down to the corresponding FAQ information to the client based on a selection operation of a category or tag related to the FAQ information at the time of inquiry of the client. Features.

  The FAQ support program according to the present invention is a FAQ support program for answering a client inquiry, and causes a computer to retain in advance a plurality of inquiry information that is assumed to be an inquiry from the client and a reply word related to the inquiry as FAQ information. When the client makes an inquiry, based on a category or tag selection operation related to the FAQ information, a search process for presenting inquiry candidates narrowed down to the corresponding FAQ information to the client is executed. It is characterized by.

  According to the present invention, it is possible to support provision of an appropriate answer to a user's inquiry.

FIG. 1 is a schematic diagram of the FAQ support system according to the embodiment. FIG. 2 is a diagram illustrating a hardware configuration example of the chat server of the FAQ support system according to the embodiment. FIG. 3 is a diagram for explaining the data hierarchy of the FAQ data file of the FAQ support apparatus according to the embodiment. FIG. 4 is a flowchart illustrating a search process example of the FAQ support apparatus according to the embodiment. FIG. 5 is a flowchart illustrating an example of an automatic tag generation process of the FAQ support apparatus according to the embodiment. FIG. 6 is a diagram illustrating a data structure example of a FAQ data file of the FAQ support apparatus according to the embodiment. FIG. 7 is a diagram illustrating a data format example of a FAQ data file of the FAQ support apparatus according to the embodiment. FIG. 8 is a diagram illustrating an environment setting example of the FAQ support apparatus according to the embodiment. FIG. 9 is an example of a display screen showing an operation processing example of the FAQ support apparatus according to the embodiment. (Part 1) FIG. 10 is an example of a display screen showing an operation processing example of the FAQ support apparatus according to the embodiment. (Part 2) FIG. 11 is an example of a display screen showing an operation processing example of the FAQ support apparatus according to the embodiment. (Part 3) FIG. 12 is an example of a display screen showing an operation processing example of the FAQ support apparatus according to the embodiment. (Part 4) FIG. 13 is an example of a display screen showing an operation processing example of the FAQ support apparatus according to the embodiment. (Part 5) FIG. 14 is an example of a display screen showing an operation processing example of the FAQ support apparatus according to the embodiment. (Part 6) FIG. 15 is a diagram for explaining an operation process of narrowing down by the FAQ support apparatus according to the embodiment. FIG. 16 is a diagram for explaining an example of narrowing down by combining each search method by the FAQ support apparatus according to the embodiment. FIG. 17 is a diagram illustrating an example of user text input by the FAQ support apparatus according to the embodiment. FIG. 18 is a chart showing the times when the user can use each search method during the search process by the FAQ support apparatus according to the embodiment. FIG. 19 is a diagram for explaining an example of processing for confirming an answer to a user for a search result by the FAQ support apparatus according to the embodiment. FIG. 20 is a diagram for explaining a tagging process example by the FAQ support apparatus according to the embodiment.

(Embodiment)
Exemplary embodiments of a FAQ support apparatus, a FAQ support system, a FAQ support method, and a FAQ support program according to the present invention will be described below in detail with reference to the accompanying drawings.

(System outline configuration)
FIG. 1 is a schematic diagram of the FAQ support system according to the embodiment. The FAQ support system includes an FAQ support apparatus 100 and a client (terminal) 105 for each user (user) U that is connected to the FAQ support apparatus 100 via a communication network such as the Internet.

  The FAQ support apparatus 100 includes a chat server 101 and an FAQ server 102. Chat server 101 and FAQ server 102 may be provided in a business center such as a company and may be configured by the same server. The chat server 101 and the FAQ server 102 can also be provided on the cloud. In the example of FIG. 1, for the sake of convenience, the chat server 101 and the FAQ server 102 are configured to be connected via a communication network. However, the chat server 101 and the FAQ server 102 each have a function on the cloud (for example, an AI engine). Can be configured.

  As the client 105, a personal computer (PC), a smartphone, a tablet, or the like can be used.

  In the following description, the FAQ support system describes an example in which the user U makes an inquiry to the FAQ business center. Note that “inquiry” and “question” described in the following description have the same meaning.

  For example, the chat server 101 generates a display screen for FAQ (inquiry and answer) by the UI execution unit 111 such as a Web page or a Web widget, and chats with the client 105 of the user U. The user U's client 105 displays the display screen generated by the chat server 101 and answers the inquiry to the chat server 101 by a user operation. The chat server 101 manages FAQ chats exchanged with the client 105.

  The chat server 101 transfers an inquiry from the client 105 of the user U to the FAQ server 102 and transfers an answer from the FAQ server 102 to the client 105.

  The FAQ server 102 includes a FAQ database (DB) 122 that holds a FAQ data file 121 created in advance, and a bot 123.

  The bot 123 refers to the FAQ database 122 based on the query (question) text (character string such as text), and searches the chat server 101 for a response text corresponding to the query text. “bot” means an automation program.

  The FAQ data file 121 held by the FAQ database 122 is, for example, a tabular data file, and has FAQ information that associates the keyword of the user U's query that has been assumed in advance and an answer (natural language information) corresponding to the keyword. . In addition, the FAQ data file 121 may include category list information for a plurality of categories related to the query keyword. Further, the FAQ data file 121 may include tag list information of a plurality of tags automatically generated by a keyword learning function. The tag list is associated with FAQ information.

  The bot 123 presents the keyword candidates at the time of inquiry to the user U, refers to the FAQ information in the FAQ data file 121 by selecting a keyword from the keyword candidates by the user U, and answers corresponding to the inquiry (keyword). The search results (restricted answer candidates) are transferred to the chat server 101.

  At this time, the bot 123 refers to the category list and searches for answers to FAQ information related to the category when the category can be narrowed down. By narrowing down the category, it becomes possible to search for an answer corresponding to the inquiry appropriately.

  Further, the bot 123 has a tag function for presenting a list of tags related to the inquiry if the keyword at the time of the inquiry is automatically generated as a tag. That is, the bot 123 has a function of automatically generating a tag by learning a specific keyword included in the FAQ information at the time of inquiry. The automatically generated tag list allows the user U to save keyword input by selecting a tag corresponding to the keyword from the tag list displayed on the client 105 when the keyword (character string) is input. Become.

  Although details will be described later, the FAQ data file 121 has a data structure in which a plurality of categories are arranged in the highest hierarchy, and FAQ information (inquiry + answer information) is arranged under each category hierarchy. In addition, the data structure is associated with a plurality of tag information corresponding to the wording of the inquiry.

  When there are a plurality of answers corresponding to the query word, the bot 123 transmits the plurality of options to the client 105 of the user U as, for example, a stamp obtained by imaging the word (text). By repeatedly selecting the stamp by the user U, it is possible to narrow down the answers corresponding to the inquiry.

  The FAQ database (DB) 122 stores a plurality of combinations of inquiries and answers to the user U in advance for frequently used FAQs. The bot 123 refers to the FAQ database 122 and sequentially executes inquiry processing.

  The above-described user U operates the client 105 in a distributed manner, and performs, for example, an inquiry communication connection to one chat server 101. The chat server 101 and the FAQ server 102 can be arranged at arbitrary positions by mutual communication connection. The chat server 101 and the FAQ server 102 can also be configured to use a cloud system.

  FIG. 2 is a diagram illustrating a hardware configuration example of the chat server of the FAQ support system according to the embodiment. For example, the chat server 101 shown in FIG. 1 has the configuration shown in FIG. The chat server 101 includes a CPU 201, ROM 202, RAM 203, magnetic disk drive 204, magnetic disk 205, optical disk drive 206, optical disk 207, input device 208, video I / F 209, display 210, communication I / F 211, and the like. Each component 201-211 is connected by a bus 220.

  The CPU 201 has a function of a control unit that controls the entire chat server 101. The ROM 202 records a server boot program. The RAM 203 is used as a work area for the CPU 201. That is, the CPU 201 controls the entire chat server 101 by executing various programs recorded in the ROM 202 while using the RAM 203 as a work area.

  The magnetic disk drive 204 controls the reading / writing of the data with respect to the magnetic disk 205 according to control of CPU201. The magnetic disk 205 records data written under the control of the magnetic disk drive 204. For example, an HD (hard disk) can be used as the magnetic disk 205.

  The optical disk drive 206 controls reading / writing of data with respect to the optical disk 207 according to the control of the CPU 201. The optical disk 207 is a detachable recording medium from which data is read according to the control of the optical disk drive 206. As the optical disk 207, a writable recording medium can be used. In addition, an SSD, a memory card, or the like can be used as a recording medium.

  Examples of the input device 208 include a remote controller including a plurality of keys for inputting characters, numerical values, various instructions, a keyboard, a touch panel, and the like. The input device 208 may be realized by any one form of a remote controller, a keyboard, and a touch panel, but can also be realized by a plurality of forms.

  The video I / F 209 is connected to the display 210. Specifically, the video I / F 209 is output from, for example, a graphic controller that controls the entire display 210, a buffer memory such as a VRAM (Video RAM) that temporarily records image information that can be displayed immediately, and a graphic controller. And a control IC for controlling the display 210 based on the image data to be processed.

  The display 210 displays icons, cursors, menus, windows, or various data such as characters and images. As the display 210, for example, a TFT liquid crystal display, an organic EL display, or the like can be used.

  The communication I / F 211 is connected to the network and functions as an interface for communication connection with the chat server 101 and the client 105. Communication networks that function as networks include the Internet connected by wire or wirelessly, public line networks, mobile phone networks, LANs, WANs, and the like.

  The chat server 101 (UI execution unit 111) shown in FIG. 1 uses the programs and data recorded in the ROM 202, RAM 203, magnetic disk 205, optical disk 207, etc. shown in FIG. By doing so, the control unit realizes the function of the chat server 101.

  Also, the FAQ server 102 described in FIG. 1 can use the same hardware configuration as in FIG. 2, and the function of the bot 123 of the FAQ server 102 is controlled by the CPU 201 shown in FIG. Can be realized. Further, the functions of the UI execution unit 111 of the chat server 101 and the bot 123 of the FAQ server 102 shown in FIG. 1 may be executed by different CPUs 201 or may be executed by one CPU 201 in the same server. Good. Further, the FAQ database 122 described in FIG. 1 can be realized by using the RAM 203 or the like described in FIG.

  Also, the client 105 shown in FIG. 1 has the same configuration as that shown in FIG. 2, and the CPU 201 functions as a control unit to control the entire process. In a portable device such as a smartphone or a tablet used as the client 105, for example, a flash memory may be used instead of the magnetic disk 205 or the optical disk 207.

(Data structure of FAQ data file)
FIG. 3 is a diagram for explaining the data hierarchy of the FAQ data file of the FAQ support apparatus according to the embodiment. The FAQ data file 121 handled by the bot 123 includes information on an FAQ category 301, FAQ information 302, and a tag 303.

  The category 301 is obtained by classifying the FAQ into a plurality of categories, and each classification has a hierarchical structure. The example of FIG. 3 is an example of FAQ in a company or the like that provides weather information and the like, and the highest level (first level) of the category 301 includes “meteorology”, “earthquake / tsunami / volcano”, “global environment” ”,“ Others ”are set, and“ Weather Forecast ”,“ Term ”,“ Typhoon ”,“ Temperature ”, etc. are set as the second layer of“ Meteorology ”.

  The FAQ information 302 is information in which a plurality of inquiries (Q) and answers (A) are associated with each other. In the example of FIG. 3, as one of a plurality of FAQ information related to the “weather forecast” in the second layer of the category 301, an inquiry Q “How is the weather predicted?” Corresponding response A “natural language information (text information)” is set as “A forecast using meteorological satellite observations and numerical forecasts based on ground and sky observation data”.

  As for the tag 303, specific text is automatically generated as a tag by learning text information included in the FAQ information 302. In the example of FIG. 3, the bot 123 learns the text of the reply A of the FAQ information 302 “A forecast using a weather satellite observation and a numerical forecast based on observation data of the ground and the sky”. , Two words of “satellite” and “observation” are automatically generated as tags. The automatic tag generation will be described later. Note that the tag 303 may be automatically generated by a learning unit (not shown) other than the bot 123.

(About search types)
As described above, in the embodiment, the FAQ data file 121 held by the FAQ database 122 includes the category 301, the FAQ information 302, and the tag 303. By using these, the FAQ server 102 (bot 123) Category search Tag search, 3. Each keyword search can be executed.

1. The category search is a search selected from a category immediately below the current category (hereinafter, the current category). In the case of the category search shown in FIG. 3, in the category search, if “weather” is the current category, a list of “weather forecast”, “term”, “typhoon”, “temperature”, etc. immediately below is displayed. When U selects one, the FAQ information 302 is narrowed down. In addition to the selection using the category list, a search can also be performed by the user U inputting a character string of “category: category name” or “> category name”.

2. The tag search is a search selected by the user U by displaying a tag list of the tags 303 related to the narrowed FAQ information (question + answer) 302. The search is not limited to the selection using the tag list, but can also be searched by inputting a character string of “tag: tag name” or “#tag name”. The search target at this time is a category 301 and a tag 303.

And in tag search,
a. FAQ information (question + answer) 302 belonging to the category 301 having a character string of the search tag 303 as a part is targeted for narrowing down.
b. FAQ information (question + answer) 302 having the same tag as the character string of the search tag 303 is targeted for narrowing down.
c. A. b. Items other than are excluded from narrowing down.

3. In the keyword search, a character string designated by the inquiry of the user U is used as a keyword. Search targets are a category 301, a tag 303, and FAQ information (question + answer) 302. In this keyword search,
a. FAQ information (question + answer) 302 belonging to a category 301 having a keyword as a part is targeted for narrowing down.
b. FAQ information (question + answer) 302 having a tag 303 having a keyword as a part is targeted for narrowing down.
c. FAQ information (question + answer) 302 including a question having a keyword as a part is targeted for narrowing down.
d. FAQ information (question + answer) 302 including an answer having a keyword as a part is targeted for narrowing down.
e. A. ~ D. Items other than are excluded from narrowing down.

(About navigation type search by linking different types of search)
The FAQ server 102 (bot 123) according to the embodiment is configured as described in 1. above. Category search Tag search, 3. By connecting the keyword search, the search that the user U has appropriately narrowed down is supported.

  FIG. 4 is a flowchart illustrating a search process example of the FAQ support apparatus according to the embodiment. An example of navigation type search processing by connecting different types of searches performed by the bot 123 of the FAQ server 102 will be described.

  The FAQ server 102 (bot 123) includes information (character string) of an inquiry made by the user U inputted via the chat server 101, and the above-described 1. Category search Tag search, 3. The specification of three search methods for keyword search is accepted (step S401).

  Thereby, the FAQ server 102 makes an inquiry by referring to the FAQ database 122 by one designated search method, that is, category search (step S402), tag search (step S403), and keyword search (step S404). Search for the corresponding answer. For example, the FAQ server 102 can return to the previous search result by holding the result of each search having a different search method during the period until the search is completed.

  Thereafter, the FAQ server 102 determines whether answer candidates have been narrowed down (step S405). If it is determined that the search has been narrowed down (step S405: Yes), the FAQ server 102 transmits a plurality of narrowed-down answers as search results to the client 105 of the user U via the chat server 101 (step S406). .

  On the other hand, if it is determined that the narrowing is not sufficient (step S405: No), the FAQ server 102 transmits information on another search method appropriate at that time to the client 105 of the user U via the chat server 101. Then, the process returns to step S401. Then, the search by the search method selected by the user U is executed again.

  For example, first, Refine by category search, then 2. 2. From the search results of category search, narrow down by keyword search, and then It is possible to further narrow down by category search from the categories in the search results of the keyword search.

  If the search result is narrowed down too much, the search process is returned to one. You can also resume the search from the category search. As described above, in the embodiment, since flexible narrowing can be performed, an appropriate answer to the inquiry of the user U can be obtained (the target FAQ information 302 can be easily reached).

(About automatic tag generation)
With regard to the above-described narrowing at the time of search, the narrowing-down conditions are relaxed by the following procedure.
1. 1. List of categories 301 whose classification is clearly determined 2. List of tags 303 that can be freely added regardless of the category 301 If there is no absorption of fluctuations by the tag 303, an arbitrary character string can be input as a keyword.

  FIG. 5 is a flowchart illustrating an example of an automatic tag generation process of the FAQ support apparatus according to the embodiment. For example, a processing example of automatic tag generation performed by the bot 123 of the FAQ server 102 or the learning unit will be described. First, when the answer requested by the user U is obtained (step S501), the FAQ server 102 determines whether a keyword search has been performed (step S502).

  If a keyword search has been performed (step S502: Yes), the FAQ server 102 determines whether the keyword at that time is a tag (step S503). If the keyword is not the tag 303 (step S503: No), the keyword is associated with the obtained FAQ information (question + answer) 302 as the tag 303 (step S504), and the above processing is terminated. In addition, when the keyword search is not performed in step S502 (step S502: No), and if the keyword is the tag 303 in step S503 (step S503: Yes), the FAQ server 102 The above process ends.

  According to the above processing, the FAQ server 102 adds the keyword to the FAQ information (question + answer) 302 of the search result if the keyword search is performed when the answer requested by the user U is obtained. In this way, tag search learning is performed. After the tag 303 is added, the FAQ server 102 determines that the tag 303 of a specific number or more is commonly added to the same category 301 to which the FAQ information (question + answer) 302 belongs. The sub-category is automatically generated with the tag 303 to learn category search.

  By automatically generating and learning the tag 303, the user U does not need to input a character string because the keyword is displayed in the tag list on the client 105 at the next inquiry. In addition, since the characteristic tag 303 is displayed as the category 301, it is possible to save the user U from searching for the optimum tag 303 among the many tags 303.

(Specific example of FAQ)
FIG. 6 is a diagram illustrating a data structure example of a FAQ data file of the FAQ support apparatus according to the embodiment. In the following, a specific configuration and processing example will be described using a FAQ of a company or the like providing weather information as an example. In the following description, the FAQ server 102 holds the FAQ data file 121 shown in FIG. 6 in the FAQ database 122, and the bot 123 performs a search with reference to the FAQ data file 121.

  The FAQ data file 121 includes FAQ category 301, FAQ information 302, and tag 303 data. The category 301 includes an upper category 301a and a subcategory 301b below the category 301a.

  In the example of FIG. 6, data (values) such as “disaster”, “weather”, “environment”, “facility”, etc. are set in the upper category 301a, and “seasonal information” is set in the subcategory 301b of “weather”. "," Weather forecast "," term "," observation equipment ", etc. are set. Further, data such as “forecast”, “probability”, “date / time”, “region”, etc. are set as the subcategory 301b of “weather forecast”. The subcategory 301b has a lower layer and can set a plurality of hierarchies.

  In the example of FIG. 6, data “probability” 612 is set as the subcategory 301 b in the data “weather forecast” 611 in the subcategory 301 b in the lower layer of the data “weather” in the category 301 a. In the FAQ information 302 related to the probability 612, data of a plurality of questions (Q1 to Q4) 621 and answers (A1 to A4) 622 is set.

  For example, the data for question Q1 (621) is “How is the precipitation probability determined?”, Whereas the data for answer A1 (622) is “the probability that there is more than 1 mm of precipitation in the specified time zone. "..." is set. In the FAQ information 302 (Q1, A1), “precipitation” 631 and “1 mm” 632 are set as data of the tag 303.

  FIG. 7 is a diagram illustrating a data format example of a FAQ data file of the FAQ support apparatus according to the embodiment. 7 shows an exemplary data format of the FAQ data file 121 shown in FIG. The FAQ data file 121 can be set in a tabular data format as shown in FIG.

  As shown in FIG. 7, the data of the FAQ data file 121 shown in FIG. 6 is set in a sheet 701 named “FAQ”. The sheet 701 in FIG. 7 has a matrix in the form of a vertical and horizontal matrix, and the upper half of the row (1st to 3rd rows) includes “# weather” 601 of category 301a in column A, “## weather forecast” 611, Data of “#### probability” 612 is set. The data starts with # and the number of # indicates the hierarchy.

  In the lower half (lines 4 to 7), FAQ information (question + answer) 302 and tag 303 data related to the FAQ information are set. Data 621 of questions (Q1 to Q4) is set in the A column, and data 622 of answers (A1 to A4) is set in the B column. After column C, data 623 and 624 of the tag 303 related to the FAQ information 302 are set.

  In this way, the row in which the data in the column A does not start with “#” indicates that the FAQ information 302 and the tag 303 are included. As a result, the bot 123 can identify the category 301a, the subcategory 301b, the FAQ information 302, and the tag 303 on the single sheet 701 in the hierarchical structure shown in FIG. 6, and can retrieve and retrieve necessary data.

  FIG. 8 is a diagram illustrating an environment setting example of the FAQ support apparatus according to the embodiment. An example of setting the operating environment of the bot 123 is shown. The operating environment of the bot 123 can also be set in the tabular data format shown in FIG.

  The environment setting table 801 illustrated in FIG. 8 is set in the sheet 802 having the name “_config_”. Environment variables are set in the A column, and data (values) can be set in the B column. In the first line, a message is displayed when the bot 123 cannot answer (the answer to the inquiry could not be retrieved). In the example of FIG. 8, “An inquiry answer was not found” is set in the B column. The initial value (default value) is “No corresponding candidate was found”.

  The second row is a message when the answer cannot be made for each category 301. In this example, the message when the answer to the “weather” set in the column A cannot be made is shown in the column B as “inquiry answer”. Could not be found. Please re-search or visit the weather page. "

  In the third row, the number of candidates “1 to 8” for ending the narrowing display in bot 123 is set. The default value is “5”. In the fourth line, whether or not tagging using the tag 303 by the user U is set “on / off” is set. The default value is “on”.

(Example of screen display for FAQ inquiry)
9 to 14 are display screen examples showing operation processing examples of the FAQ support apparatus according to the embodiment. Next, operation processing of the FAQ support apparatus 100 at the time of an inquiry by the user U will be described. In these drawings, a display example of information transmitted from the FAQ support apparatus 100 to the client 105 and displayed on the display screen of the client 105 is shown. In each display screen, the UI execution unit 111 displays a chat stamp with the user U (client 105) based on the search process of the bot 123.

  Hereinafter, using the category 301, FAQ information 302, and tag 303 corresponding to the example shown in FIG. 6, the FAQ support apparatus 100 causes the user U to select the category 301 and the tag 303, and then the FAQ information 302 (question 621). A description will be given of an example in which search guidance for the user U is operated and guided in a chat order in which the user U is selected.

  FIG. 9 shows a top screen 901. At the top of the top screen 901, the number of candidate answers 902 “candidate: 1000” for the inquiry of the user U is displayed, and at the bottom thereof, a message 903 “select a category or input a keyword for the user U” Please narrow down. ”Is displayed.

  At the bottom of the message 903, a plurality of (for example, a maximum of 6) “disaster”, “weather”,... If there are previous and next pages, “previous category” and “next category” stamps 904 are displayed at the upper and lower positions, respectively.

  FIG. 10 is a subcategory selection screen 1001 after the stamp 904 of “weather” 611 is selected in the category 301a of FIG. At the top of the sub-category selection screen 1001, a candidate number 1002 “candidate: 253” of “weather> weather forecast” narrowed down by selection of the category 301 a is displayed. If there is a lower sub-category 301b at the bottom, a message 1003 to the user U “Select a category or enter a keyword to narrow down.” Is displayed.

  In the lower part of the message 1003, a plurality of (for example, a maximum of 6) “forecast”, “probability”,... If there are previous and next pages, “previous category” and “next category” stamps 1004 are displayed at the upper and lower positions, respectively. At the bottom, a stamp “reducing narrowing by one” 1004 for returning to the upper category is displayed, and when selected, the top screen 901 of FIG. 9 can be returned.

  FIG. 11 shows a tag selection screen 1101 after the stamp 1004 of “probability” 612 is selected in the subcategory 301b of FIG. The candidate number 1102 “candidate: 61” of “weather> weather forecast> probability” narrowed down by selection of the subcategory 301b is displayed at the top of the tag selection screen 1101. If there is a tag 303 at the bottom, a message 1103 to the user U “Select a tag or enter a keyword to narrow down.” Is displayed.

  In the lower part of the message 1103, “# precipitation”, “# 1 mm”,..., “# Grade” stamps 1104 of a plurality of tags 303 (for example, a maximum of seven) are displayed. In this example, only a predetermined number that can be displayed on one page is displayed in order from the top to the bottom of the screen in the order of the number of tags, and the stamps 1104 of the previous and subsequent pages are not displayed. Under the stamp 1104, those not including the displayed tag 301 are displayed as candidates (in the example of the figure, “no corresponding tag”). At the bottom, a stamp “reducing the narrowing down by 1” 1104 for returning to the upper category is displayed, and upon selection, the subcategory selection screen 1001 in FIG. 10 can be returned.

  FIG. 12 shows a keyword input screen 1201 when the stamp 1104 of “#precipitation” 631 is selected with the tag 303 of FIG. At the top of this keyword input screen 1201, the number of candidates 1202 “candidate: 13” of the above-mentioned narrowed down “weather> weather forecast> probability> #precipitation” is displayed. In the lower part, when there is no category 301 or tag 303 candidate for further narrowing down, a message 1203 to the user U “Please enter keywords and narrow down.” Is displayed. At the bottom of the message 1203, a stamp “reducing the narrowing down by one” 1204 for returning to the higher category is displayed, and when selected, the tag selection screen 1101 of FIG. 11 can be returned.

  FIG. 13 shows a candidate display screen 1301 for the question 621 of the FAQ information 302. The FAQ support apparatus 100 (bot 123) displays candidates for the question 621 of the FAQ information 302 on the candidate display screen 1301 when the number of candidates is narrowed down to a predetermined number (for example, five or less) by the category 301 and the tag 303 described above.

  At the top of this candidate display screen 1301, the number of candidates 1302 “candidate: 5” of the above-mentioned narrowed down “weather> weather forecast> probability> #precipitation>“ time ”” is displayed. Below that, a message 1303 “Candidates are narrowed down.” Is displayed indicating that the candidates for FAQ information 302 (question) are narrowed down. At the bottom of the message 1303, the question 621 of the narrowed-down FAQ information 302 “How will the precipitation probability be determined?”, “What does the precipitation probability mean?”, “Probability of hitting the weather forecast” The stamp 1304 is displayed. At the bottom, a stamp “reducing the narrowing down by one” 1304 for returning to the upper category is displayed, and when selected, the display can be returned to the keyword input screen 1201 of FIG.

  Thereafter, when the user U selects the stamp 1304 of any question 621 in the FAQ information 302, the client 105 of the user U outputs the selected stamp 1304 to the FAQ support apparatus 100 as an inquiry, and provides FAQ support. The apparatus 100 (bot 123) transmits an answer 622 corresponding to the question 621 of the selected stamp 1304 to the user U (client 105).

  FIG. 14 shows a no-candidate display screen 1401 when there is no candidate for the question 621 in the FAQ information 302 as a result of narrowing down. The FAQ support apparatus 100 (bot 123), when there is no candidate by the above-described narrowing by the category 301 and the tag 303, first, as the transmission message 1402 to the user U (client 105), the above narrowing “weather> weather forecast> probability > #Precipitation ”candidate number 1403“ Candidate: 13 ”is displayed, and there is no FAQ information 302 (question) candidate below it, so as a message 1404“ Please enter a keyword to narrow down. ” indicate. In the lower part of the message 1404, a stamp “reducing the narrowing down by 1” 1405 for returning to the upper category is displayed.

  Thereafter, it is assumed that the user U operates the client 105 and inputs a keyword “early morning” 1413 for narrowing down.

  The FAQ support apparatus 100 (bot 123) searches for this keyword “early morning”, but if the corresponding FAQ information 302 (question 621) cannot be searched, the message 1404 “corresponding candidate was not found” and the user U (client 105). In addition, the FAQ support apparatus 100 (bot 123), similarly to the above, as the next transmission message 1405 to the user U (client 105), the number of candidates 1403 “candidate” of the above-mentioned narrowed down “weather> weather forecast> probability> #precipitation” : 13 ”is displayed, and since there is no FAQ information 302 (question) candidate,“ Please enter a keyword and narrow down. ”Is displayed as a message 1404. At the bottom of the message 1403, a stamp “reducing the narrowing down by 1” 1405 for returning to the higher category is displayed.

  According to the above processing, the FAQ support apparatus 100 (bot 123) can narrow down the candidates of the question 621 of the FAQ information 302 using the category 301 and the tag 303 by chatting with the user U (client 105). At the time of narrowing down, the user U can easily reach the question 621 to be inquired by a simple operation of selecting a stamp transmitted by the FAQ support apparatus 100 (bot 123) without inputting a keyword.

  FIG. 15 is a diagram for explaining an operation process of narrowing down by the FAQ support apparatus according to the embodiment. An example of the narrowing operation performed by the FAQ support apparatus 100 (bot 123) using the hierarchical structure of the FAQ data file 121 will be described.

  The FAQ support apparatus 100 (bot 123) performs narrowing down by combining three types of search methods as shown in FIGS. In the category search shown in FIG. 15A, the bot 123 selects the predetermined data (in the above example, “weather”) 601 of the highest category 301a by the user U, and the subcategory 301b below this “weather” 601. Data 611 (in the above example, “seasonal forecast”, “weather forecast”, etc.) 611 are selected. When there are a plurality of subcategories, data 612 under the similarly selected subcategory 301b (in the above example, “forecast”, “probability”, etc. under “weather forecast”) is selected.

  In the category search, the bot 123 narrows down the user U by selecting the stamp displayed on the client 105 or inputting the text “category: category name” or “> category name”. At this time, the highest layer of a category having a plurality of candidates is targeted.

  In the tag search shown in FIG. 15B, the user U selects the data of the tag 303 displayed on the client 105 (in the above example, “Precipitation” 631, “1 mm” 632, etc.) The bot 123 narrows down by inputting the text “tag: tag name” or “#tag name”. At the time of tag search, the bot 123 narrows down tags, but if the characters input by the user U are two or more characters, the category 301 (category 301a and subcategory 301b) is also narrowed down by partial match search.

  In the keyword search shown in FIG. 15C, the user U inputs the keyword text on the client 105 to narrow down the corresponding FAQ information 302 data (“question” 621, “answer” 622). If the keyword is included in the FAQ information (“Question” 621 and “Answer” 622), the bot 123 is a selection target for narrowing down. At the time of keyword search, the bot 123 narrows down the FAQ information (question + answer) 302, but if there are two or more characters input by the user U, the category 301 (category 301a and subcategory 301b) and the tag 303 Are also targeted for narrowing down by partial match search.

  FIG. 16 is a diagram for explaining an example of narrowing down by combining each search method by the FAQ support apparatus according to the embodiment. FIG. 16 shows a state in which search result candidates (see FIG. 13) obtained by combining the three search methods described above are obtained.

  At the top of the candidate display screen 1301, the number of candidates 1302 “candidate: 5” of the above-mentioned narrowed down “weather> weather forecast> probability> #precipitation>“ time ”” is displayed. At this time, the bot 123 is narrowed down by “weather” 601 of the category 301a and narrowed down by “weather forecast” 611 and “probability” 612 of the subcategory 301b. Further, the bot 123 is narrowed down by “#precipitation” 631 of the tag 303 and narrowed down by the keyword “time”.

  As a result, the bot 123 displays that the selection candidates have been narrowed down to five as shown in FIG. The bot 123 is a tag 303 with “#” added to the head, the text surrounded by ““ ”is recognized as a keyword, and the other is recognized as a category 301 (category 301a or subcategory 301b). . Note that the user U only needs to input the keyword, and the bot 123 adds ““ ”and“ ”” before and after the keyword input by the user U.

  FIG. 17 is a diagram illustrating an example of user text input by the FAQ support apparatus according to the embodiment. An input character when the user U inputs a character (text) on the client 105 is shown. As shown in FIG. 17, when the category 301 is designated, the user U inputs “category:” or “?” Added to the head. When the tag 303 is designated, the user U inputs “tag:” or “#” with the head added. The keyword itself is input except for specifying the category 301 or the tag 303 and canceling the narrowing down.

  When specifying multiple filters, separate them with a space. For example, in the illustrated example, a space is input after the keyword “time”, a space is input after inputting a category 301 “category: probability of weather”, and then a tag 303 “#precipitation” is input. At this time, the bot 123 processes the input text (character string) from the top, and if there is no corresponding candidate, the narrowing process is skipped.

  Further, when “<” is input at the head (“<” stamp selection in FIG. 10 etc.), the narrowing down is reduced by one (one narrowing down is canceled). Also, in order to cancel an arbitrary number of refinements, the input of “<” is repeated. In the example shown in the figure, “<<” is input when three are released. Alternatively, the number to be narrowed down is designated after “<”. In the example shown in the figure, “<3” is input when three are released.

  Also, to cancel all the narrowing down, input “<all”. The bot 123 recognizes the user U's text input and performs a corresponding narrowing or cancellation of the narrowing.

  FIG. 18 is a chart showing the times when the user can use each search method during the search process by the FAQ support apparatus according to the embodiment. The horizontal axis represents time, and as shown in FIG. 18A, the FAQ support apparatus 100 (bot 123) guides and displays an operable search method to the user U (client 105).

  At the start of the search process, the bot 123 displays a message indicating the selection of the category 301a to the user U (step S1801), and then displays a message indicating the selection of the subcategory 301b (step S1802). Thereafter, the bot 123 displays a message indicating that the tag 303 has been selected (step S1803), and then displays a message indicating that only keyword input is possible (step S1804), and displays the search result candidates for refinement to the user U (step S1804). S1805).

  At the time of the search process, the bot 123 describes a time when characters can be input by the user U (client 105). The period T1 during which characters can be input for selection of the category 301 is a period until the selection of the category 301a and the subcategory 301b is completed. The period T2 during which characters can be input for selection of the tag 303 is a period until the selection of the category 301a to the tag 303 is completed. The period T3 during which the keyword characters can be input is possible from the category 301a to the entire period in which candidates are displayed. Time T4 when the narrowing can be canceled is a period from subcategory 301b to candidate display.

  FIG. 19 is a diagram for explaining an example of a process for confirming an answer to a user with respect to a search result by the FAQ support apparatus according to the embodiment. A processing example of the user U (the FAQ support apparatus 100 (bot 123) after confirmation by the client 105) after the search result is displayed will be described.

  As shown in FIG. 19, the bot 123 displays search result candidates (see FIG. 13) obtained by combining the three search methods described above on the candidate display screen 1901. Reference numeral 1902 is an answer display with one stamp of the candidate displays.

  Subsequently, the bot 123 displays an inquiry screen 1903 that inquires the user U about “the answer you want?” With respect to the narrowed response display 1902 indicated by the candidate display screen 1901. On the inquiry screen 1903, stamps “Yes” 1911 and “No” 1912 are displayed.

  When the user U selects the stamp “No” 1912, the user U returns to the candidate display screen 1901 and makes it possible to select another answer candidate display. When the user U selects the “Yes” 1911 stamp, tagging processing 1921 by the user U is performed. In the tagging process 1921, as described above, the bot 123 performs automatic learning and tag generation. Then, the bot 123 returns to the category selection and displays the top screen 901 (see FIG. 9).

  FIG. 20 is a diagram for explaining a tagging process example by the FAQ support apparatus according to the embodiment. When “Yes” 1911 which is an answer requested by the user U is selected on the answer display 1902 of the search result shown in FIG. 19, the bot 123 performs a tagging process. At this time, the bot 123 transmits a tagging display screen 2001 to the user U (client 105). On the tagging display screen 2001, “Please tag. You can also tag with text input.” Is displayed as a message 2002.

  Here, when the user U inputs “# time” 2011 as a keyword to be tagged from the character string (text) included in the answer display 1902, the bot 123 newly sets the keyword “# time”. This tag 303 is generated as a tag related to the answer display 1902. For example, it is additionally updated as data 623 and 624 of the tag 303 in the FAQ database 122 (the FAQ data file 121 shown in FIG. 7).

  Here, the bot 123 does not perform the tagging process unless there is a keyword input by the user U. Further, tagging is not performed even when tagging is set to “off” in the environment setting (see FIG. 8).

  According to the embodiment described above, appropriate search can be performed by arbitrarily selecting and combining the category search and the tag search in addition to the keyword search at the time of the search process, and the user seeks Answers can be provided accurately.

  The FAQ support apparatus holds, as a database, an FAQ data file in which categories and tags are associated with FAQ information of assumed user inquiries (questions) and answers. Then, by selecting a category and a tag by the user, it is possible to appropriately narrow down candidates for FAQ information (question) that the user wants to make an inquiry to. As a result, the FAQ support apparatus can provide the user with an inquiry (question) itself without any keyword operation input by the user, and the search accuracy can be reduced while saving labor of the user's search. Can be improved. In addition, appropriate answers related to the inquiry can be provided to the user.

  In addition, by setting a plurality of hierarchies for a category, it is possible to improve the accuracy of narrowing down during search. Furthermore, by generating a tag from a character string (text) included in the FAQ information (question + answer), a new tag can be presented as a candidate at the time of search and search can be performed. In addition, each time a FAQ search is performed, the search accuracy can be improved, and in addition, the user's labor for searching can be further saved.

  When a tag is added, if a specific number of tags or more are commonly added to the same category to which the corresponding FAQ information (question + answer) belongs, the lower-level category (subcategory) is added. ) May be automatically generated. Thereby, the target search result can be obtained without the keyword input by the category selection by the user. In addition, since characteristic tags are displayed as categories, it is possible to perform a search without the need to search for a desired tag from among a large number of tags.

  Also, the FAQ support apparatus transmits the chat stamp to the user, so that the user can simplify the text input operation as much as possible and can proceed with the search only by selecting the stamp. As a result, the search can be performed efficiently.

  Further, according to the embodiment, it is possible to perform an automatic response without requiring manual operations such as an operator, and to perform a 24-hour FAQ operation.

  In the above-described embodiment, the description has been given by taking as an example the FAQ work that receives an inquiry from a user of a certain company. Quality control by inquiries and responses to periodic quality control at multiple restaurants in the restaurant industry, restaurant reservations, shift work management by periodic shift requests and responses to part-time job managers, taxi and other driving diaries and vehicles It can be applied to various tasks such as daily report management of information and advance payments.

  Note that the program related to FAQ business support described in the present embodiment can be realized by executing a program prepared in advance by a computer. The program is recorded on a computer-readable recording medium such as a semiconductor memory, a hard disk, a flexible disk, a CD-ROM, or a DVD, and is executed by being read from the recording medium by the computer. Further, this program may be distributed through a network such as the Internet.

  As described above, the present invention is useful for application to an apparatus or a support system that supports a FAQ job using a chat service.

DESCRIPTION OF SYMBOLS 100 FAQ support apparatus 101 Chat server 102 FAQ server 105 Client 111 UI execution part 121 FAQ data file 122 FAQ database 201 CPU
202 ROM
203 RAM
211 Communication I / F
U users

In order to achieve the above object, the FAQ support apparatus according to the present invention is a FAQ support apparatus that answers a client's inquiry, and includes a plurality of inquiries in advance as FAQ information regarding an inquiry that assumes an inquiry from the client and an answer related to the inquiry. A control unit that holds and performs a search process using a narrowed-down dialog UI that arbitrarily combines a keyword of the inquiry, a category related to the FAQ information, or a tag according to a selection operation at the time of inquiry of the client; Generates a character string of an inquiry from the client that is not included in the FAQ information as the new tag .

Also, the FAQ support apparatus of the present invention is a FAQ support apparatus that answers a client inquiry, and stores in advance a plurality of inquiry information that is assumed to be an inquiry from the client and an answer word related to the inquiry as FAQ information. A control unit that performs a search process for presenting inquiry candidates narrowed down to the corresponding FAQ information to the client based on a category or tag selection operation related to the FAQ information at the time of the inquiry, A FAQ data file in which the category and the tag are respectively associated with the FAQ information is held as a FAQ database, the category or a list of the tags is presented to the client, and the category or the tag is selected by the client operation The search process is performed, a tag is generated from the character string included in the FAQ information, the tag is added to the FAQ data file, and the character string included in the FAQ information presented to the client is The character string selected by the client is added to the FAQ data file as a new tag .

Further, the control unit determines a search keyword from a character string at the time of inquiry of the client, narrows down the FAQ information required by the client based on the search keyword, and can be selected by the client in the process of the narrowing down If there is no tag, the client is informed that the character string at the time of the inquiry of the client can be added as a new tag, and the character string at the time of the inquiry is added to the new tag based on an addition instruction from the client. And added to the FAQ data file .

Further, the control unit transmits a plurality of stamps obtained by imaging the tag and the category information to be presented to the client according to the number of categories or the list of tags, and the stamp of the client by the client is transmitted . The search process is performed based on the selection.

In addition, the FAQ support system of the present invention is a FAQ support system including a client, a chat server that answers the client's inquiry, and an FAQ server that performs a search process based on the client's inquiry. A plurality of inquiries regarding inquiries that are assumed to be inquiries from the client and the answers related to the inquiries are stored in advance as FAQ information, and the inquiry keyword, category related to the FAQ information, or A search process using a narrowed-down dialog UI in which tags are arbitrarily combined is performed, and a character string of an inquiry from the client that is not included in the FAQ information is generated as a new tag. And, interactively by chat, send and receive information regarding inquiry, the information of the search process from the client, it is characterized.

In the FAQ support method of the present invention, in the FAQ support method for answering a client inquiry, the computer holds in advance a plurality of inquiry information that is assumed to be an inquiry from the client and a reply word related to the inquiry as FAQ information. The client that is not included in the FAQ information is subjected to a search process using a narrowed-down dialog UI that arbitrarily combines a query keyword, a category related to the FAQ information, or a tag by a selection operation at the time of inquiry of the client. A process of generating a character string of an inquiry from as a new tag is executed .

Further, FAQ support program of the present invention, in the FAQ support program to respond to the client's query, the computer, the wording of answers associated with assumed query and the query queries from the client in advance a plurality held as FAQ Information The client that is not included in the FAQ information is subjected to a search process by a narrowed-down dialog UI that arbitrarily combines a keyword of the inquiry, a category related to the FAQ information, or a tag by a selection operation at the time of inquiry of the client. A process for generating a character string of an inquiry from as a new tag is executed .

Claims (11)

In the FAQ support device that answers client inquiries,
A plurality of inquiries regarding the inquiry from the client and the wording of answers related to the inquiry are stored in advance as FAQ information.
A control unit that performs a search process for presenting inquiry candidates narrowed down to the corresponding FAQ information to the client based on a category or tag selection operation related to the FAQ information when the client makes an inquiry; FAQ support device. The control unit holds a FAQ data file that associates the category and the tag with the FAQ information as a FAQ database,
2. The FAQ support apparatus according to claim 1, wherein the category or the tag candidate is presented to the client, and the search process is performed by selecting the category or the tag by the client. The FAQ database holds the categories in a plurality of layers,
3. The FAQ support apparatus according to claim 2, wherein the category is presented to the client in a plurality of hierarchies, and the search processing is performed by the selection operation of the categories of the multi hierarchies by the client. The controller is
The FAQ support apparatus according to claim 1, wherein a tag is generated from a character string included in the FAQ information, and the tag is added to the FAQ data file. The controller is
5. The FAQ support apparatus according to claim 4, wherein a character string selected by the client is added to the FAQ data file as a new tag for a character string included in the FAQ information presented to the client. The controller is
The FAQ according to any one of claims 1 to 5, further comprising transmitting and receiving information on an inquiry from the client and information on the search process to the client in an interactive manner by chat. Support device. The controller is
Sending a plurality of stamps corresponding to the number of candidates, the image of the tag and the category information to be presented to the client,
The FAQ support apparatus according to claim 6, wherein the search processing is performed based on the selection of the stamp by the client. In a FAQ support system including a client, a chat server that answers an inquiry from the client, and an FAQ server that performs a search process based on the inquiry from the client,
The FAQ server
A plurality of inquiries regarding the inquiry from the client and the wording of answers related to the inquiry are stored in advance as FAQ information.
When the client makes an inquiry, based on a category or tag selection operation related to the FAQ information, a search process for presenting inquiry candidates narrowed down to the corresponding FAQ information to the client is performed.
The chat server
Sending and receiving information on the inquiry from the client and information on the search process in an interactive form by chat to the client.
FAQ support system characterized by this.   9. The FAQ support system according to claim 8, wherein the functions of the chat server and the FAQ server are arranged on a cloud. In the FAQ support method for answering client inquiries,
Computer
A plurality of inquiries regarding the inquiry from the client and the wording of answers related to the inquiry are stored in advance as FAQ information.
A search process for presenting inquiry candidates narrowed down to the corresponding FAQ information to the client based on a selection operation of a category or tag related to the FAQ information at the time of inquiry of the client;
A FAQ support method characterized by executing processing. In the FAQ support program that answers client inquiries,
On the computer,
A plurality of inquiries regarding the inquiry from the client and the answer words related to the inquiry are stored in advance as FAQ information,
When the client makes an inquiry, based on a category or tag selection operation related to the FAQ information, search processing for presenting inquiry candidates narrowed down to the corresponding FAQ information to the client is performed.
FAQ support program characterized by executing processing.

Images