JP2023073412A - Information provision method, information provision system, information provision device, and computer program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an information provision method capable of providing information by using audio and text input and output as appropriate, an information provision system, an information provision device, and a computer program.

SOLUTION: An information provision system has an information provision device that includes: a user interface placed in a hotel; a request acquisition unit that acquires a request based on a voice input to a user interface; a response information acquisition unit that acquires a piece of response information to the request acquired by the request acquisition unit from a database; an audio output processing unit that performs a series of processing to output a piece of response information acquired by the response information acquisition unit to the user interface via voice; and a processing unit that is configured so as to, when the response information cannot be acquired from the database, perform a series of processing to switch to an automatic response by means of text input/output.

SELECTED DRAWING: Figure 2

COPYRIGHT: (C)2023,JPO&INPIT

Description

Application for application of Article 30, Paragraph 2 of the Patent Law has been filed October 1, 2018 at HOTEL EMIT SHIBUYA, November 9, 2018 at MIMARU Tokyo Akasaka, and November 9, 2018 at MIMARU Kyoto Published at Shinmachi Sanjo and 4 other locations November 20, 2018, International Hotel Ryokan November 20, 2018 No. 440, Page 2, March 5, 2019, International Hotel Ryokan Heisei 31 Published on page 2 of No. 447 on March 5, 2018 Published on page 4 of Nikkei MJ on October 29, 2018 on October 29, 2018 Published on Weekly Building Management on October 22, 2018 Published on page 20 of issue No. 1075 on October 22, 2018 Published on the website of Nikkei Crosstrend on March 27, 2019 (https://xtrend.nikkei.com/atcl/contents/18 /00130/00003/?P=1) (https://xtrend.nikkei.com/atcl/contents/18/00130/00003/?P=2) On March 8, 2019, HOTELS (Weekly Hotel Restaurant ) Published on page 53 of Vol. Iw AR)

Applied for the application of Article 30, Paragraph 2 of the Patent Law Published on the website of PR TIMES on September 28, 2018 (https://prtimes.jp/main/html/rd/p/000000003.000033116.html ) (https://prtimes.jp/main/html/rd/p/000000006.000033116.html) (https://prtimes.jp/main/html/rd/p/000000012.000033116.html) (https: //prtimes.jp/main/html/rd/p/000000014.000033116.html) (https://prtimes.jp/main/html/rd/p/000000015.000033116.html) (https://prtimes.jp/main/html/rd/p/000000015.000033116.html) jp/main/html/rd/p/000000018.000033116.html) (https://prtimes.jp/main/html/rd/p/000000020.000033116.html) On March 18, 2019, N- Published on the LIO website (https://n-lio.com/tradfit-1/?fbclid=IwAR1gsYbwKpX8sPJIANhwfQuy_Qc3TgBbDDJWVgIL_akMbIDWfiyecu5ctNM) (https://n-lio.com/tradfit-2/? fbclid=IwAR0Fsd5HDT3VYszgIkBSWP8XQSneVKNWlBWRnBNW5e9guOxPj4zylfgKbe8)

Applied for application of Article 30, Paragraph 2 of the Patent Law Published on the HotelBank website on January 18, 2019 (https://hotelbank.jp/ai-chatbot-speaker-tradfit-hitocommunications/) (https: //hotelbank.jp/tradfit-interview-2019/?fbclid=IwAR3Ub-wQAkGxikCaQxv08oa6UHgFZEYAav7qId2MRBZLMHEP0d2-kOmDo_E) (https://hotelbank.jp/hyatt -regency-tokyo-tradfit/) On January 7, 2019, Tech in Released on the Asia website (https://www.techinasia.com/tech-southeast-asia-surging-tourist-arrivals) Released on the CANVAS8 website on January 10, 2019 (https:/ /library.canvas8.com/signals/2019/01/10/tradfit-tourism.html) Released on the HOTELLIER website on November 12, 2018 (https://www.hotelier.jp) /inboundnews/service/post-9743.html?fbclid=IwAR1-nE273MPUs9QiLaZw5OQdpa_1tsObxtmFXe3CU3sdAkckZWNcorVNtvo) (https://www.hotelier.jp/inbou ndnews/service/20190129-003.html?fbclid=IwAR2hnK90w1jZASyfyLvuRrZ1IXDqwOME-WZHKGrCprSAcyosePLdxBmOJgU) October 10, 2018 (https://mainichi.jp/articles/20181011/k00/00m/020/043000c)

Applied for application of Article 30, Paragraph 2 of the Patent Act Published on AMP website on September 29, 2018 (https://amp.review/2018/09/29/hotel_ai/) October 2018 Released on the website of the Junkan Travel Newspaper on the 2nd (http://www.ryoko-net.co.jp/?p=42827) On January 28, 2019, on the website of the Kanko Keizai Newspaper Public (https://www.kankokeizai.com/%E5%B0%8F%E7%94%B0%E6%80%A5%E3%83%9B%E3%83%86%E3%83%AB%E3 %82%BB%E3%83%B3%E3%83%81%E3%83%A5%E3%83%AA%E3%83%BC%E3%82%B5%E3%82%B6%E3%83 %B3%E3%82%BF%E3%83%AF%E3%83%BC%E3%80%81ai-%E3%82%B9%E3%83%94/) On January 14, 2019, Published on vietnambiz website (https://vietnambiz.vn/loa-thong-minh-chatbot-dich-ngon-ngu-se-len-ngoi-trong-thoi-dai-du-lich-bung-no- 117051.htm) Published on OK SKY website on February 13, 2019 (https://mag.ok-sky.jp/%E3%83%9B%E3%83%86%E3%83 %AB%E3%81%AB%E3%82%82%E5%B0%8E%E5%85%A5%E3%81%95%E3%82%8C%E3%82%8B%E3%82%B9 %E3%83%9E%E3%83%BC%E3%83%88%E3%82%B9%E3%83%94%E3%83%BC%E3%82%AB%E3%83%BC%E3 %81%A8%E3%83%81%E3%83%A3/) Published on the DIGITAL X website on October 30, 2018 (https://dcross.impress.co.jp/docs/ usecase/000735.html)

Applied for Article 30, Paragraph 2 of the Patent Act Published on the SankeiBiz website on September 28, 2018 (https://www.sankeibiz.jp/business/news/180928/prl1809281313120-n1.htm) Published on the Sankei News website on September 29, 2018 (https://www.sankei.com/economy/news/180929/prl1809290029-n1.html) Published on the website of NEWS RELEASE operated by Co., Ltd. (http://file.swcms.net/file/hitocom/ja/tekiji/auto_20190118460795/pdfFile.pdf) On October 11, 2018, CNET Japan website Published on the website (https://japan.cnet.com/article/35126787/) Published on the Nikkan Kogyo Shimbun electronic version website on January 7, 2019 (https://www.nikkan.co) jp/releases/view/90314) Released on the MIMARU Hotel website on November 9, 2018 (https://mimaruhotels.com/ja-jp/news-topics/news20181109.pdf) 2019 Published on the website of Odakyu Hotel Century Southern Tower on January 18 (https://www.southerntower.co.jp/Portals/0/newsrelease/2018/aispeaker_chatbot.pdf) January 18, 2019 , TV Tokyo released a program called "WBS (World Business Satellite)" On October 12, 2018, AbemaTV released a program called "Keyaki Hills"

The present disclosure relates to an information providing method, an information providing system, an information providing device, and a computer program for providing various information to users who use accommodation facilities.

Systems that allow users to make voice inputs to devices are beginning to spread. Devices called smart speakers or AI (Artificial Intelligence) speakers, for example, are beginning to be installed in homes as such systems. The user can obtain various kinds of information by voice output from the smart speaker by asking questions or the like to the smart speaker by voice. In addition, the user can also control the operation of home appliances by giving voice commands to the smart speaker.

Patent Literature 1 proposes a communication system that establishes dialogue in natural language between a user and a robot. The robot has a microphone and a speaker, transmits the user's speech acquired by the microphone to the server, receives a response from the server, and outputs the response as voice from the speaker. The server has a conversation database of questions and answers used for conversation, refers to this conversation database, generates a response to the user's utterance, and transmits it to the robot.

JP 2017-191531 A

Although devices capable of voice input are beginning to spread, current devices cannot accurately respond to all of the various requests made by users. Also, there is a problem that voice input is less accurate than text input.

The present disclosure has been made in view of such circumstances, and the purpose thereof is to The object is to provide an apparatus and a computer program.

An information providing method according to the present disclosure acquires a request based on voice input at an accommodation facility via a user interface, acquires response information corresponding to the acquired request from a database, and transmits the acquired response information to the user interface. If the response information cannot be acquired from the database, the automatic response is switched to text input/output.

In the present disclosure, a user interface installed in an accommodation facility acquires a request based on voice input from a user, and outputs response information to the acquired request by voice from the user interface. If it is not possible to output the response information for the acquired request, it switches to an automatic response by text input/output. This allows the user to easily switch from voice input/output to text input/output.

According to the present disclosure, input/output of voice and text can be appropriately utilized to provide information to the user.

BRIEF DESCRIPTION OF THE DRAWINGS It is a schematic diagram for demonstrating the outline|summary of the information provision system which concerns on this Embodiment. BRIEF DESCRIPTION OF THE DRAWINGS It is a schematic diagram for demonstrating the outline|summary of the information provision system which concerns on this Embodiment. FIG. 4 is a schematic diagram showing an example of access information displayed on a smart speaker; It is a schematic diagram which shows an example of the chat screen displayed on a smart phone. It is a schematic diagram which shows an example of the chat screen displayed on a smart phone. 1 is a block diagram showing the configuration of a smart speaker according to an embodiment; FIG. 3 is a block diagram showing the configuration of a management server according to this embodiment; FIG. FIG. 4 is a schematic diagram showing a configuration example of an intention identification model; It is a schematic diagram which shows one structural example of response information DB. It is a schematic diagram which shows one structural example of operator DB. 1 is a block diagram showing the configuration of a smart phone according to an embodiment; FIG. 4 is a flowchart showing the procedure of processing performed by the smart speaker according to the embodiment; 7 is a flow chart showing a procedure of switching control processing performed by the management server according to the embodiment; 7 is a flowchart showing the procedure of chat start processing performed by the smartphone according to the embodiment; 6 is a flowchart showing the procedure of chat processing performed by the smartphone according to the embodiment; 7 is a flow chart showing a procedure of switching control processing performed by the management server according to the embodiment; 9 is a flowchart showing a procedure of switching processing to manned response performed by the management server according to the present embodiment; 7 is a flow chart showing the procedure of learning processing of an intention identification model performed by the management server according to the present embodiment; It is a schematic diagram for demonstrating the information provision before accommodation by the information provision system which concerns on this Embodiment. FIG. 10 is a schematic diagram showing an example of access information displayed on the smart speaker according to Embodiment 2; 10 is a flow chart showing the procedure of processing performed by the smart speaker according to Embodiment 2. FIG. FIG. 12 is a flow chart showing a procedure of voice communication start processing performed by the smartphone according to Embodiment 2; FIG.

A specific example of an information providing system according to an embodiment of the present disclosure will be described below with reference to the drawings. The present disclosure is not limited to these examples, but is indicated by the scope of the claims, and is intended to include all modifications within the meaning and scope of equivalents of the scope of the claims.

<System Overview>
1 and 2 are schematic diagrams for explaining the outline of the information providing system according to the present embodiment. The information providing system 10 according to the present embodiment provides information to the user using the smart speaker 20 installed as a user interface in accommodation facilities such as hotel rooms, common areas, or private houses used for private lodging. Can do. The smart speaker 20 shown in FIG. 1 has a configuration in which a touch panel 25, a microphone 26, a speaker 27, a camera 28, and the like are appropriately arranged in a substantially hemispherical housing. However, the shape and the like of the smart speaker 20 is an example and is not limited to this, and various shapes such as a cylindrical shape or a rectangular parallelepiped shape can be adopted. Further, the user interface installed in the guest room is not limited to the smart speaker 20, and may be any device as long as it is an information processing device capable of performing voice input/output.

A user (guest) staying in a guest room can activate the smart speaker 20 by uttering a specific keyword (for example, "start service"). The user can give various requests to the smart speaker 20 by voice input by speaking, and can obtain various information such as information on this accommodation facility by voice response information output from the smart speaker 20. . In the illustrated example, the user utters a request to the smart speaker 20 saying, "Tell me the checkout time." Note that in the present embodiment, various requests, questions, or the like input by the user to the smart speaker 20 or the like are called requests.

A voice request input by the user to the smart speaker 20 is input to the voice processing server 50 via a network such as the Internet. The voice processing server 50 converts input voice information into text (character string) information by performing voice recognition processing on voice input. The voice processing server 50 inputs the converted text to the management server 30 .

The management server 30 performs a process of generating response information to the user's request based on the text information input from the voice processing server 50 . The management server 30 uses an intention identification model 81 and a response information DB (database) 82 for this process. The intention identification model 81 is a pre-learned model that accepts text information relating to a request as an input and outputs identification results that identify the user's intention relating to this request. The response information DB 82 is a database in which intentions related to requests and texts of response messages to the intentions are stored in association with each other. In this example, the intention identification model 81 and the response information DB 82 are provided in the management server 30, but the configuration is not limited to this, and the intention identification model 81 or the response information DB 82 may be provided in another server.

The management server 30 acquires the intention of the user's request by inputting the text information obtained from the speech processing server 50 into the intention identification model 81 and acquiring the information output by the intention identification model 81 . In the example shown in FIG. 1, the intention identification model 81 identifies the user's intention of inquiring about the checkout time for the text information "Tell me the checkout time." The management server 30 then obtains the text information of the response message to the intention of the request by referring to the response information DB 82 based on the identified intention of the user. The management server 30 outputs the text information acquired from the response information DB 82 to the voice processing server 50 . In the example shown in FIG. 1, text information such as "Please check out by 10:00 am" is stored in the response information DB 82 as response information to the intention of inquiring about the check-out time.

The speech processing server 50 to which the text output of the response information is given from the management server 30 converts the given text information into speech information. The audio processing server 50 outputs the converted audio information to the smart speaker 20 . The smart speaker 20 acquires the audio information output from the audio processing server 50 and outputs audio based on this audio information through the speaker 27 . In the example shown in FIG. 1, the smart speaker 20 outputs a voice message "Please check out by 10:00 am".

Note that the management server 30 can also respond to a user's request to the smart speaker 20 in a manner other than voice output. For example, the management server 30 can display an image on the touch panel 25 of the smart speaker 20 . In this case, the management server 30 outputs image information for display to the audio processing server 50 , and the audio processing server 50 provides this image information to the smart speaker 20 .

In addition, the provision of information to the user through the above-described voice input/output is possible, for example, when the language spoken by the user is not included in the languages supported by the voice processing of the voice processing server 50, or when the content of the user's utterance is If the text information cannot be properly converted into text information, there is a possibility that appropriate information cannot be provided to the user. In the information providing system according to the present embodiment, when information cannot be appropriately provided by voice input/output, switching to information provision by text input/output is performed. A system that automatically responds by inputting and outputting text is called a chatbot, and the information providing system according to the present embodiment provides information using the smart speaker 20 and information using It is a system that switches between and as appropriate.

For example, when the management server 30 is notified from the voice processing server 50 that the voice could not be converted to text, or when the intention identification model 81 cannot identify the intention of the user's request, for example, The smart speaker 20 is caused to output a message prompting switching to the chatbot. The message can be, for example, "Your voice request was not recognized. Would you like to use a chatbot?" This message may be audibly output by the smart speaker 20 or may be displayed as text on the touch panel 25 . In response to this message, when the user indicates that he or she will approve the use of the chatbot, the management server 30 sends access information for accessing the chatbot using a terminal device such as the smartphone 40, The smart speaker 20 is made to output. The user's intention as to whether or not to use the chatbot may be received, for example, as a voice input, or may be received, for example, by a touch operation on the touch panel 25 of the smart speaker 20 .

FIG. 3 is a schematic diagram showing an example of access information displayed on the smart speaker 20. As shown in FIG. In the information providing system according to the present embodiment, smart speaker 20 displays access information 20 a on touch panel 25 . The illustrated access information 20a is a two-dimensional code (for example, a barcode or a QR code (registered trademark)), which encodes information such as a URL (Uniform Resource Locator) for accessing the chatbot. Further, in the present embodiment, the access information 20a includes identification information for identifying accommodation facilities, guest rooms, and the like. As a result, the chatbot can acquire the identification information included in the access information 20a and make an automatic response suitable for accommodation facilities, guest rooms, and the like. However, the access information 20a may include only information such as a URL, and may not include identification information such as accommodation facilities and guest rooms. In this case, for example, the user may input information about accommodation facilities, guest rooms, and the like.

The user can acquire information such as a URL for accessing the chatbot by photographing the access information 20a displayed on the smart speaker 20 with the camera of the smartphone 40 of the user. The user who has acquired the access information 20a can start using the chatbot by accessing a server that provides the chatbot service using an application such as a browser installed on the smartphone 40. In this embodiment, the management server 30 provides a chatbot service. A chat screen is displayed on the display unit of the smartphone 40 , and the user can use the text input function of the smartphone 40 to input a request as a text on the chat screen.

4 and 5 are schematic diagrams showing examples of chat screens displayed on the smartphone 40. FIG. FIG. 4 shows a display example of a selection box prompting the user to select a language prior to using the chatbot. In this embodiment, English, Japanese, Chinese, Korean, etc. can be used as chat languages, and selectable languages are listed in the language notation in the selection box. The user can select any one language by performing a touch (tap) operation on the language displayed in the selection box.

Note that the language selection box shown in FIG. 4 is not only displayed prior to the start of the chat, but also when the user touches the language selection icon provided on the lower left side of the chat screen shown in FIG. may also be displayed. The language selection icon is labeled to indicate which language is currently selected. Also, on the chat screen in FIG. 5, a switch icon for switching to manned response is displayed next to the language selection icon.

In this embodiment, the user selects a language using the selection box shown in FIG. 4, but the configuration is not limited to this. For example, information such as a user's nationality may be acquired from a server device that stores information about guests of an accommodation facility, and a language suitable for the user may be automatically selected.

FIG. 5 also shows an example of requests and responses by chat. In the illustrated example, English is selected as the chat language. On the illustrated chat screen, for example, title strings such as "hotel information service", "example hotel" and "inquiry" are displayed at the top, and this chat screen is a chat screen for inquiring about the hotel. It is shown. Below the title string, the text of the request input by the user and the text of the response thereto are displayed in chronological order from top to bottom. Also, the text entered by the user is displayed aligned to the right side of the chat screen, and the text of the response by the chatbot is displayed aligned to the left side. In this example, the text "How many waters would you like to order?"

Information of the text input with the smartphone 40 is input to the translation server 60 via a network such as the Internet. The translation server 60 translates the input text information as necessary. When a text in a language other than the language that the management server 30 can handle is input, the translation server 60 translates this text into a language that the management server 30 can handle. For example, if the management server 30 can handle both Japanese and English, and a request in Chinese is input, the translation server 60 translates the request from Chinese into Japanese or English, and provides information on the translated text. Input to the management server 30 . When text information that does not require translation is input, the translation server 60 may input the text information to the management server 30 without performing translation.

The management server 30 performs a process of generating response information to the user's request based on the text information input from the translation server 60 . The processing of the management server 30 performed at this time is the same as the processing of generating response information to the user's request by voice input, and is performed using the above-described intention identification model 81 and response information DB 82 . The management server 30 outputs the response information of the generated text to the translation server 60 . The translation server 60 appropriately translates the response information given from the management server 30 as necessary, and outputs it to the smartphone 40 . The smartphone 40 automatically responds to the request input by the user by displaying the response information given as text from the translation server 60 as text on the chat screen. The management server 30 has an intention identification model 81 and a response information DB 82 for acquiring response information to requests for voice input, and an intention identification model 81 and response information DB 82 for acquiring response information to requests for text input. It may be provided separately.

The provision of information to the user by the chatbot, for example, when the request input by the user in text cannot be identified by the intention identification model 81, or when, for example, the response information corresponding to the response information DB 82 is not stored, Appropriate information may not be provided to users. In the information providing system according to the present embodiment, when information cannot be appropriately provided by text input/output of the chatbot, the operator switches to a manned response. In the information providing system according to the present embodiment, there is an operator who responds to the user's request instead of the management server 30 when the user's request cannot be automatically responded. The operator can use the operator terminal 70 to exchange text information with the smartphone 40 of the user, that is, chat. The user can chat with the operator using the same chat screen as the automatic response of the chatbot.

When the management server 30 determines that automatic response by text input/output is not possible, the management server 30 causes the smartphone 40 to output a message prompting switching to manned response. The message can be, for example, "We are unable to answer your request. Please touch the switch icon below to switch to manned response." When the user touches the switch icon in response to this message and indicates that the user intends to affirm the switch to the manned response, the management server 30 notifies the operator terminal 70 of requesting the manned response. to switch from the chatbot's automatic response to an operator's manned response.

Switching to the manned response by a touch operation on the switching icon provided on the chat screen may be accepted only when the management server 30 outputs a message prompting the switching to the smartphone 40. You can always accept it regardless. If the switching is accepted only when the management server 30 outputs a message prompting the switching to the smartphone 40, the switching icon need not be displayed on the chat screen when unnecessary.

After switching to the manned response by the operator, the management server 30 transmits to the operator terminal 70 the information of the text input by the user on the smartphone 40 and translated by the translation server 60 as necessary. Based on the text information from the management server 30, the operator terminal 70 displays a chat screen similar to that shown in FIG. The operator inputs response information to the user's request displayed on the operator terminal 70 into the operator terminal 70 using an input device such as a keyboard. In the present embodiment, the operator inputs text to the operator terminal 70, but the configuration is not limited to this, and voice input or the like may be performed. The operator terminal 70 accepts a response text input by the operator and transmits the accepted text information to the management server 30 . The management server 30 provides the text information from the operator terminal 70 to the translation server 60, and the text information translated by the translation server 60 is provided to the smartphone 40 of the user as needed. The smartphone 40 displays the given text information on the chat screen. These enable communication between the user and the operator via text, and the operator can respond to the user's request.

Note that in the present embodiment, switching from information provision by voice input/output using smart speaker 20 to information provision by text input/output using smartphone 40 is performed via access information displayed on smart speaker 20. However, it is not limited to this. For example, when a device that can use both voice input/output and text input/output is installed in a guest room, the management server 30 outputs an instruction to switch from voice input/output to text input/output to the device, Based on the output, the device may switch from voice input/output to text input/output.

<Device configuration>
FIG. 6 is a block diagram showing the configuration of smart speaker 20 according to the present embodiment. The smart speaker 20 according to the present embodiment includes a processing unit (processor) 21, a storage unit (storage) 22, a communication unit (transceiver) 23, a touch panel 25, a microphone 26, a speaker 27, a camera 28, and the like. there is The processing unit 21 is configured using an arithmetic processing device such as a CPU (Central Processing Unit) or an MPU (Micro-Processing Unit). Also, the processing unit 21 may employ a plurality of CPUs, a multi-core CPU, or the like. The processing unit 21 reads out and executes the program 22a stored in the storage unit 22 to perform various processes such as reception of voice input, voice output and image display.

The storage unit 22 is configured using, for example, a non-volatile memory device such as a flash memory, or a magnetic storage device such as a hard disk. The storage unit 22 stores various programs executed by the processing unit 21 and various data required for processing by the processing unit 21 . In the present embodiment, storage unit 22 stores program 22a executed by processing unit 21 and identification information 22b for identifying smart speaker 20 .

The program 22a may be written into the storage unit 22 during the manufacturing stage of the smart speaker 20, for example. For example, the program 22a may be distributed by a remote server device or the like and acquired by the smart speaker 20 through communication. For example, the program 22a may be stored in the storage unit 22 after the smart speaker 20 reads the program 22a recorded in the recording medium 99 such as a memory card or an optical disc. For example, the program 22 a may be recorded in the recording medium 99 and read by a writing device and written in the storage unit 22 of the smart speaker 20 . The program 22 a may be provided in the form of distribution via a network, or may be provided in the form of being recorded on the recording medium 99 .

The identification information 22b is information in which characters, numbers, or the like are appropriately combined, and may be any information that can identify the smart speaker 20 . In the present embodiment, the identification information 22b is a combination of information identifying the lodging facility where the smart speaker 20 is installed and information identifying the guest room. For example, the identification information 22b can be information indicating room 301 of the Example Hotel. The identification information 22b is set by a system administrator, an accommodation facility administrator, or the like when the smart speaker 20 is installed in a guest room of the accommodation facility.

The communication unit 23 can communicate with various devices via a network N including the Internet, a wireless LAN (Local Area Network), a mobile phone communication network, and the like. In this embodiment, the communication unit 23 communicates with the voice processing server 50 via the network N. FIG. The communication unit 23 transmits data provided from the processing unit 21 to the voice processing server 50 and provides data received from the voice processing server 50 to the processing unit 21 .

The touch panel 25 is one of the user interfaces included in the smart speaker 20, and has a display section 25a and an input section 25b. The display unit 25a is configured using a liquid crystal display or the like, and displays various images, characters, and the like. The display unit 25 a displays various images based on the processing of the processing unit 21 . The input unit 25b is provided with a sensor for detecting contact by the user on the surface of the display unit 25a. is notified to the processing unit 21 .

The microphone 26 and the speaker 27 are one of the user interfaces included in the smart speaker 20, and realize a user interface by voice input/output. The microphone 26 acquires surrounding sounds, converts them into digital sound information, and supplies the sound information to the processing unit 21 . The speaker 27 outputs sound based on the sound information given from the processing unit 21 . In addition, the camera 28 takes an image of, for example, the user or the guest room, and provides the image information of the image taken to the processing unit 21 .

The smart speaker 20 according to the present embodiment acquires a request uttered by the user through the speaker 27 , and transmits voice information related to the acquired request to the voice processing server 50 through the communication unit 23 . At this time, the smart speaker 20 transmits the audio information with the identification information 22b stored in the storage unit 22 attached. This voice information is converted into text information by the voice processing server 50 and given to the management server 30, and the management server 30 transmits a response to the request to the voice processing server 50 as text information. The voice processing server 50 converts the text information from the management server 30 into voice information and transmits the voice information to the smart speaker 20 . The smart speaker 20 outputs the audio information given from the audio processing server 50 from the speaker 27 .

When the management server 30 cannot respond to the user's request based on the text information provided from the voice processing server 50, the management server 30 notifies the fact and transmits access information for accessing the chatbot. This notification and access information is provided to smart speaker 20 via audio processing server 50 . In response to this notification, the smart speaker 20 inquires of the user whether or not to switch to the chatbot. Information is displayed on the display section 25 a of the touch panel 25 .

FIG. 7 is a block diagram showing the configuration of the management server 30 according to this embodiment. The management server 30 according to this embodiment includes a processing unit (processor) 31, a storage unit (storage) 32, a communication unit (transceiver) 33, and the like. The processing unit 31 is configured using an arithmetic processing device such as a CPU, MPU, or GPU (Graphics Processing Unit). By reading and executing the server program 32a stored in the storage unit 32, the processing unit 31 performs processing for responding to user requests, processing for switching from voice input/output to chatbot, and switching from chatbot to manned response. and various processing such as switching processing.

The storage unit 32 is configured using, for example, a large-capacity storage device such as a hard disk. The storage unit 32 stores various programs executed by the processing unit 31 and various data required for processing by the processing unit 31 . In the present embodiment, the storage unit 32 stores a server program 32a executed by the processing unit 31 and an intention identification model 81 as a learned model (classifier). The storage unit 32 is also provided with two databases: a response information DB 82 that stores responses to user requests, and an operator DB 83 that stores information about operators who make manned responses.

The communication unit 33 can communicate with various devices via a network N including the Internet, wireless LAN, mobile phone communication network, and the like. In this embodiment, the communication unit 33 communicates with the speech processing server 50, the translation server 60, the operator terminal 70, etc. via the network N. FIG. The communication unit 33 transmits data received from the processing unit 31 to other devices, and provides the processing unit 31 with data received from other devices.

Note that the storage unit 32 may be an external storage device connected to the management server 30 . The management server 30 may be a multicomputer including a plurality of computers, or may be a virtual machine virtually constructed by software. The management server 30 is not limited to the above configuration, and may include, for example, a reading unit that reads information stored in a portable storage medium, an input unit that receives operation inputs, or a display unit that displays images. .

The server program 32a may be written in the storage unit 32 during the manufacturing stage of the management server 30, for example. For example, the server program 32a may be distributed by another remote server device or the like and acquired by the management server 30 through communication. For example, the server program 32a may be stored in the storage unit 32 after the management server 30 reads the server program 32a recorded in the recording medium 98 such as a memory card or an optical disk. For example, the server program 32 a may be recorded in the recording medium 98 and read by a writing device and written in the storage unit 32 of the management server 30 . The server program 32 a may be provided in the form of distribution via a network, or may be provided in the form of being recorded on the recording medium 98 .

The intention identification model 81 is a trained model that has undergone machine learning or deep learning using teacher data in advance. The learned model performs a predetermined calculation on the input value and outputs the calculation result, and the storage unit 32 stores data such as coefficients and threshold values of functions that define this calculation as the intention identification model 81. be done. The intention identification model 81 is a trained model that has been trained to identify the user's intention with respect to sentence information input by the user as voice or text. The processing unit 31 that executes the server program 32a reads the data stored as the intention identification model 81, so that the processing unit 31 can execute the calculation for identifying the intention of the user's request.

FIG. 8 is a schematic diagram showing a configuration example of the intention identification model 81. As shown in FIG. In this embodiment, the intention identification model 81 includes an input layer that accepts input of text information related to a user's request, an intermediate layer that performs a predetermined operation on the input information, and an identification result of the intention of the user's request. is configured as a neural network having an output layer for outputting information indicating In the present embodiment, the intention identification model 81 may adopt, for example, a configuration of RNN (Recurrent Neural Network) as a neural network, but it is not limited to this, and a model other than RNN may be adopted. The management server 30 generates an intention identification model 81 by learning the model of the RNN using text information relating to the request and teacher data in which the intention is associated.

In this embodiment, the intention identification model 81 is provided in the management server 30, but the present invention is not limited to this. For example, the intention identification model 81 is provided in another server device, the text information related to the user's request acquired by the management server 30 is transmitted to the another server device, and the identification result by the intention identification model 81 possessed by the another server device is managed. The server 30 may acquire it.

Further, in the present embodiment, the management server 30 performs learning processing for the intention identification model 81 . However, the learning process may be performed by another server device. The intention identification model 81 learned in this case may be transmitted from another server device to the management server 30 and stored in the storage unit 32, or may be held by another server device. The learned intention identification model 81 may be provided in the form of distribution via a network, or may be provided in the form of being recorded on the recording medium 98, similar to the server program 32a.

FIG. 9 is a schematic diagram showing a configuration example of the response information DB 82. As shown in FIG. The response information DB 82 is a database in which the intention of a user's request and response information such as a message in response to the request are stored in association with each other. Each item of the intention of the request in the response information DB 82 corresponds to the identification result of the intention identification model 81 . The response information in the response information DB 82 is text information of a message that responds to the intention of the request as voice output from the smart speaker 20 or text output from the chatbot. Since the response to the request differs depending on the accommodation facility, the response information DB 82 is provided for each accommodation facility. Although the response information DB 82 shown in FIG. 9 is for Japanese, a response information DB 82 corresponding to other languages such as English or Chinese may be provided as necessary.

The example shown in FIG. 9 is an example of the intention of the request and the response information of the response information DB 82 of "Example Hotel" as an accommodation facility. Note that the request given from the smart speaker 20 to the management server 30 via the voice processing server 50 and the request given from the smartphone 40 to the management server 30 via the translation server 60 include identification information is attached as, for example, header information or the like. The management server 30 has a response information DB 82 for a plurality of accommodation facilities, and can use the response information DB 82 suitable for the accommodation facility according to the identification information attached to the request.

In the illustrated example, the intention of the request "What time is check-in?" ing. Also, the intention of the request "Do you need cash when checking in?" is stored in association with the response information "You can pay by credit card." Also, the intention of the request "Where is the front desk?" Note that the illustrated request intent and response information are merely examples, and are not limited to these.

FIG. 10 is a schematic diagram showing a configuration example of the operator DB 83. As shown in FIG. The operator DB 83 is a database in which conditions related to dates and times and information related to operators to be switched to manned responses are stored in association with each other. The operator DB 83 is provided for each accommodation facility. In the illustrated example, it is set that an operator at the operator center will make a manned response from 8:00 am to 8:00 pm on weekdays and from 6:00 am to 10:00 pm on weekends. In addition, for the dates and times from 12:00 am to 8:00 am and 8:00 pm to 12:00 pm on weekdays and from 0:00 am to 6:00 am and from 10:00 pm to 12:00 pm on holidays, It is set that the operator at the front desk will provide a manned response. When switching to a manned response, the management server 30 refers to the operator DB 83 based on the date and time information, and determines the operator who performs the manned response. The management server 30 transmits a manned response request to the operator terminal 70 used by the determined operator. In the illustrated example, an operator center, a hotel front desk, and the like are set as operators, but actually, identification information of the operator terminal 70 provided in these may be set.

The operator DB 83 may also store information about languages that operators can handle. As a result, the management server 30 can have an operator who can speak the language used by the user make a manned response. Regarding the language used by the user, the management server 30 may allow the user to select a language, for example, when switching to a manned response, or determine the language of the user from information such as a guest list of an accommodation facility. good too.

In the management server 30 according to the present embodiment, the processing unit 31 reads out and executes the server program 32a stored in the storage unit 32, so that the request acquisition unit 31a, the response processing unit 31b, the switching processing unit 31c, and the learning The processing unit 31d and the like are implemented as software functional blocks. The request acquisition unit 31a acquires a request that is converted into text information by the voice processing server 50 after the user inputs voice to the smart speaker 20 by communicating with the voice processing server 50 through the communication unit 33. I do. The request acquisition unit 31a acquires a request input by the user to the smartphone 40 and translated into an appropriate language by the translation server 60 by communicating with the translation server 60 through the communication unit 33. I do. All of the request information acquired by the request acquisition unit 31a is text information.

The response processing unit 31b uses the intention identification model 81 and the response information DB 82 stored in the storage unit 32 to respond to the request acquired by the request acquisition unit 31a. The response processing unit 31b inputs the text information of the request acquired by the request acquisition unit 31a to the intention identification model 81, and acquires the intention identification result output by the intention identification model 81. FIG. The response processing unit 31b refers to the response information DB 82 based on the acquired intent of the request, and acquires response information corresponding to the intent. The response processing unit 31 b transmits the acquired response information to the voice processing server 50 or the translation server 60 , and responds by voice output from the smart speaker 20 or text output from the smartphone 40 .

The switching processing unit 31c switches from the voice response using the smart speaker 20 to the chatbot using the smartphone 40 when an appropriate response cannot be made to the user's request acquired by the request acquisition unit 31a. or switch from chatbot to manned response. For example, when the intention identification model 81 cannot identify the intention in the request text information acquired by the request acquisition unit 31a, or when response information corresponding to the identified intention is received from the response information DB 82, the switching processing unit 31c If not obtained, etc., the switching is performed.

When the switching processing unit 31c cannot respond to a request given from the voice processing server 50, the switching processing unit 31c transmits an instruction prompting the smart speaker 20, which is the source of the request, to switch to the chatbot. do. At this time, the instruction given from the management server 30 to the smart speaker 20 via the voice processing server 50 includes an image of the access information 20a required for accessing the chatbot. The access information 20a is information in which, for example, a URL for accessing the chatbot and identification information of an accommodation facility and a guest room in which the smart speaker 20 is installed are two-dimensionally coded. Upon receiving this instruction, the smart speaker 20 inquires of the user whether or not switching to the chatbot is possible, and displays the access information 20a when a response to that effect is obtained.

When the switching processing unit 31c cannot respond to a request given from the translation server 60, the switching processing unit 31c transmits an instruction prompting switching to manned response to the smartphone 40, which is the transmission source of this request. Upon receiving this instruction, the smartphone 40 inquires of the user whether or not it is possible to switch from the chatbot to the manned response, and transmits a switching request to the management server 30 when a response to the effect of switching is obtained. Upon receiving this request, the switching processing unit 31c of the management server 30 acquires information about the date and time at that time, refers to the operator DB 83 stored in the storage unit 32, and determines the operator to switch to. Also, at this time, the management server 30 may determine the operator to switch to, taking into consideration the languages that the operator can handle. The switching processing unit 31c transmits an instruction to make a manned response to the operator terminal 70 used by the determined operator. After that, the management server 30 relays the text information regarding the user's request and the operator's response between the smartphone 40 and the operator terminal 70 .

The learning processing unit 31 d performs a process of learning (relearning) the intention identification model 81 . For example, the learning processing unit 31d requests the operator to input teaching data after the above-described switching is performed by the switching processing unit 31c and the manned response by the operator is finally completed. At this time, the learning processing unit 31d displays, on the operator terminal 70, the text information related to the user's request that the response processing unit 31b could not respond to, and the request actually determined by the operator with respect to this text information. Prompt for intent input. The learning processing unit 31d acquires from the operator terminal 70 and stores in the storage unit 32, as teacher data, information obtained by combining the text information of the user's request and the intention input by the operator. The learning processing unit 31d re-learns the intention identification model 81 stored in the storage unit 32 using newly stored teacher data when a predetermined amount of teacher data is accumulated or each time teacher data is acquired. do. Re-learning of the intention identification model 81 may be performed by another server device, and the management server 30 may simply store teacher data.

FIG. 11 is a block diagram showing the configuration of smartphone 40 according to the present embodiment. A smartphone 40 according to the present embodiment includes a processing unit (processor) 41, a storage unit (storage) 42, a communication unit (transceiver) 43, a touch panel 44, a camera 45, and the like. The processing unit 41 is configured using an arithmetic processing unit such as a CPU or MPU. The processing unit 41 reads out and executes a program 42a stored in the storage unit 42 to perform various processing such as processing for accessing the management server 30 and processing for implementing chat.

The storage unit 32 is configured using a non-volatile memory device such as a flash memory, for example. The storage unit 42 stores various programs executed by the processing unit 41 and various data required for processing by the processing unit 41 . In the present embodiment, the storage unit 42 stores a program 42a executed by the processing unit 41. FIG. The program 42a may be written into the storage unit 42 during the manufacturing stage of the smart phone 40, for example. For example, the program 42a may be distributed by a remote server device or the like and acquired by the smartphone 40 through communication. For example, the program 42 a may be stored in the storage unit 42 by reading the program 42 a recorded on a recording medium such as a memory card or an optical disc by the smartphone 40 . For example, the program 42a may be recorded in a recording medium, read by a writing device, and written in the storage unit 42 of the smartphone 40 . The program 42a may be provided in the form of distribution via a network, or may be provided in the form of being recorded on a recording medium.

The communication unit 43 can communicate with various devices via a network N including the Internet, wireless LAN, mobile phone communication network, and the like. In this embodiment, the communication unit 43 communicates with the translation server 60 via the network N. FIG. The communication unit 43 transmits the data given from the processing unit 41 to the translation server 60 and gives the data received from the translation server 60 to the processing unit 41 .

The touch panel 44 is one of user interfaces included in the smartphone 40 and has a display section and an input section. A display unit of the touch panel 44 is configured using a liquid crystal display or the like, and displays various images, characters, etc. based on the processing of the processing unit 41 . The input unit of the touch panel 44 is provided with a sensor for detecting contact by the user on the surface of the display unit. is notified to the processing unit 41 .

The camera 45 picks up an image of the surroundings and provides information about the picked up image to the processing unit 41 . In the present embodiment, the camera 45 of the smart phone 40 is used to capture the access information 20a displayed on the touch panel 25 of the smart speaker 20. FIG.

Further, in the smartphone 40 according to the present embodiment, the processing unit 41 reads and executes the program 42a stored in the storage unit 42, so that the access information acquisition unit 41a, the chat processing unit 41b, and the like are software functional blocks. is realized as The access information acquisition unit 41 a performs processing for acquiring the two-dimensional code access information 20 a displayed on the touch panel 25 by the smart speaker 20 . The access information acquisition unit 41 a displays, on the touch panel 44 , a message that prompts the user to capture an image of the access information 20 a displayed on the smart speaker 20 , and acquires an image of the access information 20 a captured by the camera 45 . Based on the acquired image, the access information acquisition unit 41a acquires the URL for accessing the chatbot included in the access information 20a, the identification information of the accommodation facility and guest room where the smart speaker 20 is installed, and the like. do.

The chat processing unit 41b accesses the management server 30 that provides the chatbot (via the translation server 60) based on the access information acquired by the access information acquisition unit 41a. The chat processing unit 41b displays the language selection box shown in FIG. 4 and accepts the selection of the language used in the chat, and then displays the chat screen shown in FIG. The chat processing unit 41 b acquires the text information of the request input by the user, and transmits the acquired text information to the management server 30 via the translation server 60 . The chat processing unit 41b also receives the text information of the response transmitted from the management server 30 via the translation server 60, and displays the received text information on the chat screen. It should be noted that the chat processing unit 41b can proceed with the chat by the same process regardless of whether the chat partner is an automatic response by a chatbot or a manned response by an operator.

<Response switching process>
In the information providing system according to the present embodiment, there are three types of information providing methods: a voice response using the smart speaker 20, an automatic response by a chatbot using the smartphone 40, and a manned response by an operator using the smartphone 40. provided to users. Further, in the information providing system according to the present embodiment, the management server 30 determines whether to switch from the voice response of the smart speaker 20 to the automatic response of the chatbot, and whether to switch from the automatic response of the chatbot to the manned response of the operator. By prompting the user to switch the information provision method, smooth switching of the information provision method is realized.

(1) Switching from smart speaker 20 voice response to chatbot automatic response Request by input. The smart speaker 20 transmits voice information related to the input request to the voice processing server 50 , and the voice processing server 50 converts the voice information into text information and transmits the text information to the management server 30 .

The voice processing server 50 transmits Japanese text information to the management server 30 if the input voice information is in Japanese, and transmits English text information to the management server 30 if the input voice information is in English. When voice information in a language not supported by the voice processing server 50 is input, the voice processing server 50 sends a notification to the management server 30 to the effect that the voice information cannot be converted into text information.

Based on the text information received from the voice processing server 50, the management server 30 uses the intention identification model 81 and the response information DB 82 to acquire response information to the user's request. The management server 30 transmits this response information as text information to the voice processing server 50 , and the voice processing server 50 converts the text information into voice information and transmits the voice information to the smart speaker 20 . The smart speaker 20 receives the audio information from the audio processing server 50 and outputs the received audio information from the speaker 27 as a response to the user's request.

Further, based on the text information received from the voice processing server 50, the management server 30 switches from voice response to chatbot automatic response when response information cannot be obtained even by using the intention identification model 81 and the response information DB 82. A switching instruction prompting switching is transmitted to the smart speaker 20 via the audio processing server 50 . At this time, the management server 30 generates two-dimensional code access information 20a including information such as a URL for accessing the chatbot and identification information of the accommodation facility and guest room where the smart speaker 20 is installed. It is sent along with the switching instruction.

There may be multiple factors that prevent the management server 30 from acquiring the response information. For example, the reason for this is that the voice input to the smart speaker 20 is not clear, so that the voice processing server 50 cannot convert voice information into text information. The same applies when the input voice information is in a language not supported by the voice processing server 50 . In these cases, the voice processing server 50 notifies the management server 30 that the voice processing was not possible, and the management server 30 can determine from this notification that the response information cannot be obtained.

Another factor is, for example, that the intention identification model 81 cannot identify the intention of the request. When the intention identification model 81 is configured to output the intention of the identified request and the certainty of the identification result for the input request, if the output certainty does not exceed a predetermined threshold, The management server 30 can determine that the intention identification model 81 cannot identify the intention, and can determine that response information cannot be acquired.

Another factor is that the intention of the request identified by the intention identification model 81 and the corresponding response information are not registered in the response information DB 82, for example. The management server 30 searches the response information DB 82 based on the intention of the request output by the intention identification model 81, and can determine that the response information cannot be obtained when the corresponding item does not exist in the response information DB 82.

Note that the method for determining that the management server 30 cannot acquire the response information corresponding to the request is not limited to the method described above, and various methods may be adopted.

Upon receiving the switching instruction from the management server 30 via the voice processing server 50, the smart speaker 20 outputs a message or the like inquiring whether to switch to the chatbot automatic response from the speaker 27 or the touch panel 25. , and inquires whether or not the user can switch. When a response affirming the switching is obtained from the user by voice input or operation on the touch panel, the smart speaker 20 displays on the touch panel 25 the two-dimensional code of the access information 20a attached to the switching instruction.

When the access information 20a is displayed on the touch panel 25 of the smart speaker 20, the user reads the access information 20a with his/her smartphone 40 and accesses the chatbot implemented by the management server 30 with the smartphone 40. be able to. The smartphone 40 displays the language selection box shown in FIG. 4 to allow the user to select a language used in chat. Smartphone 40 displays a chat screen in the selected language. On this chat screen, the smartphone 40 accepts text input of a request from the user, and outputs a text response from the management server 30 . In the present embodiment, the smartphone 40 reads the access information 20a using an application program or the like that reads a two-dimensional code, and accesses the URL included in the read access information 20a using a program such as a browser. shall be A chat screen is displayed on the browser, and processing for displaying the chat screen is performed by cooperation between the management server 30 and the smartphone 40 .

In this embodiment, the user selects the language used for chatting, but the configuration is not limited to this. The management server 30 may automatically select the language, for example, based on the identification information of the accommodations and guest rooms included in the access information 20a. For example, the management server 30 acquires the identification information of the accommodation facility and the guest room from the smartphone 40, and from another server device that manages the guests of the accommodation facility, stays in the guest room of the accommodation facility identified by the identification information. By acquiring information such as the nationality of the user, the language used for chat can be selected.

FIG. 12 is a flowchart showing the procedure of processing performed by smart speaker 20 according to the present embodiment. The processing unit 21 of the smart speaker 20 according to the present embodiment determines whether or not the user has voice-inputted a request to the microphone 26 (step S1). If no voice input has been made (S1: NO), the processing unit 21 waits until voice input of the request is made. When voice input is made (S1: YES), the processing unit 21 acquires the voice input by the microphone 26 as voice information of digital data (step S2). The processing unit 21 transmits the acquired voice information to the voice processing server 50 through the communication unit 23 (step S3).

After that, the processing unit 21 determines whether or not a response to the transmitted request has been received (step S4). If no response has been received (S4: NO), the processing unit 21 waits until it receives a response. Here, the response received by the smart speaker 20 is voice response information to the request or an instruction to switch to the chatbot. If a response has been received (S4: YES), the processing unit 21 determines whether or not the received response is an instruction to switch to the chatbot (step S5). If it is not a switching instruction (S5: NO), the processing unit 21 outputs the received voice information of the response from the speaker 27 (step S6), and ends the process.

If the received response is a switching instruction (S5: YES), the processing unit 21 inquires of the user whether or not switching is possible by outputting a message or the like inquiring whether or not switching is possible from the speaker 27 or by displaying it on the touch panel 25. (Step S7). The processing unit 21 determines whether or not an affirmative answer to the inquiry has been obtained (step S8). If an answer affirming the switching is obtained (S8: YES), the processing unit 21 displays the two-dimensional code access information 20a given from the management server 30 together with the switching instruction on the touch panel 25 (step S9). exit. If no answer affirming the switching is obtained (S8: NO), the processing unit 21 terminates the process without displaying the access information 20a.

FIG. 13 is a flow chart showing the procedure of switching control processing performed by the management server 30 according to the present embodiment. The request acquisition unit 31a of the processing unit 31 of the management server 30 according to the present embodiment determines whether or not the communication unit 33 has received a request based on text information from the voice processing server 50 (step S21). If no request has been received (S21: NO), the request acquisition unit 31a waits until a request is received.

If a request has been received (S21: YES), the response processing unit 31b of the processing unit 31 inputs text information of the received request to the intention identification model 81 stored in the storage unit 32 (step S22). Next, the response processing unit 31b acquires the identification result of the intention of the request output by the intention identification model 81 (step S23). Based on the acquired identification result, the response processing unit 31b determines whether or not the intention of the request was identifiable (step S24). If the intention of the request can be identified (S24: YES), the response processing unit 31b refers to the response information DB 82 stored in the storage unit 32 based on the identified intention (step S25). The response processing unit 31b refers to the response information DB 82 and determines whether response information has been obtained (step S26). When the response information is obtained (S26: YES), the response processing unit 31b transmits the response information obtained from the response information DB 82 to the smart speaker 20 that sent the request via the voice processing server 50 (step S27). , terminate the process.

If the intention of the request cannot be identified by the intention identification model 81 (S24: NO), or if response information for the intention of the request cannot be obtained from the response information DB 82 (S26: NO), the processing unit 31 The switching processing unit 31c generates the access information 20a for accessing the chatbot (step S28). The switching processing unit 31c transmits a switching instruction including the generated access information 20a to the smart speaker 20 that is the source of the request via the voice processing server 50 (step S29), and ends the process.

FIG. 14 is a flowchart showing the procedure of chat start processing performed by smartphone 40 according to the present embodiment. The processing unit 41 of the smartphone 40 according to the present embodiment starts an application program according to a user's operation, for example, and shifts to a mode of reading a two-dimensional code (step S41). In this mode, the access information acquisition unit 41a of the processing unit 41 acquires the access information 20a by imaging the access information 20a with the camera 45 (step S42). The access information acquisition unit 41a acquires the URL and identification information included in the acquired access information 20a (step S43).

Next, the chat processing unit 41b of the processing unit 41 accesses the management server 30 based on the acquired URL and identification information (step S44). At this time, the smartphone 40 transmits the identification information acquired from the access information 20a to the management server 30, so that the management server 30 can determine which accommodation facility and guest room user is accessing. can. The chat processing unit 41b displays a language selection box and accepts selection of a language used for chat (step S45). The chat processing unit 41b starts chat processing with the chatbot provided by the management server 30 in the selected language (step S46), and ends the processing.

(2) Switching from automatic response by chatbot to manned response by operator A user who has started a chat using the smartphone 40 can input a request by text input. The text information of the input request is transmitted from the smart phone 40 to the translation server 60, and the translation server 60 translates the text information as necessary. The translation server 60 transmits the translated text information or the text information that does not require translation to the management server 30 . Note that if the text information does not need to be translated, the text information may be directly transmitted from the smart phone 40 to the management server 30 .

Based on the text information received from the translation server 60, the management server 30 uses the intention identification model 81 and the response information DB 82 to obtain response information to the user's request. The management server 30 transmits this response information as text information to the translation server 60, and the translation server 60 translates the text information as necessary. The translation server 60 transmits the translated text information or the text information that does not require translation to the smartphone 40 . The smartphone 40 receives the text information from the translation server 60 and displays the received text information on the chat screen as a response to the user's request.

Further, based on the text information received from the translation server 60, the management server 30, if response information cannot be obtained even by using the intention identification model 81 and the response information DB 82, changes the automatic response of the chatbot to the manned response of the operator. A switching instruction prompting switching to is transmitted to the smartphone 40 via the translation server 60 .

There may be multiple factors that prevent the management server 30 from acquiring the response information. For example, the reason is that the intention identification model 81 cannot identify the intention of the request. Another factor is that the intention of the request identified by the intention identification model 81 and the corresponding response information are not registered in the response information DB 82, for example. Various methods can be adopted as a method for determining that the management server 30 cannot acquire the response information corresponding to the request.

The smartphone 40, which has received the switching instruction from the management server 30 via the translation server 60, displays on the chat screen a message or the like inquiring whether to switch to the manned response of the operator, and inquires whether or not the user can switch. . In response to this inquiry, the user can switch to a manned response by, for example, performing a touch operation on a switching icon provided on the chat screen. When the answer affirming the switch is obtained from the user, the smartphone 40 transmits a request to switch to the manned response to the management server 30 via the translation server 60 .

The management server 30 that has received the switching request from the smartphone 40 acquires the date and time information at that time, and refers to the operator DB 83 stored in the storage unit 32 based on the acquired date and time information. The management server 30 determines an operator to request a manned response based on the date and time information and the operator DB 83, and transmits an instruction to perform a manned response to the operator terminal 70 used by the determined operator. Also, at this time, the management server 30 may determine the operator to switch to, taking into consideration the languages that the operator can handle.

In addition, the management server 30 transmits information about the history and the user, etc., and displays them on the operator terminal 70 before the operator starts manned responses. The history may include, for example, text information indicating the contents of voices input/output by the smart speaker 20, text information input/output by the automatic response of the chatbot, and the like. The management server 30 stores this information in the storage unit 32 as history information. The information about the user can be, for example, information about the user's sex, age, nationality, etc., and is acquired from another server device or the like that stores information about guests of the accommodation facility. By displaying such information on the operator terminal 70 in advance, the operator who makes a manned response can smoothly start a chat with the user.

After that, the management server 30 relays text information transmission and reception between the smartphone 40 and the operator terminal 70 to establish a text information chat between the user and the operator.

Further, in the present embodiment, after the manned response by the operator is finished, the management server 30 requests the operator who made the manned response to create training data for re-learning the intention identification model 81 . The management server 30, for example, inquires of the operator terminal 70, which has completed the manned response to the user, what the intention of this user's request was. At this time, the management server 30 may transmit the text information input by the user in the voice response or chatbot automatic response to the operator terminal 70 before the manned response, and display the text information on the operator terminal 70. . Based on the message or the like displayed on the operator terminal 70, the operator inputs the intention of the request of the user who made the manned response. The operator terminal 70 transmits information input by the operator to the management server 30 . The management server 30 receives information from the operator terminal 70, creates teacher data in which the text information of the request related to the user's input is associated with the intention of the request input by the operator, and stores it in the storage unit 32. Remember. The information input by the operator at the operator terminal 70 can be used not only for learning the intention identification model 81 but also for adding or correcting response information to the response information DB 82 .

FIG. 15 is a flowchart showing the procedure of chat processing performed by smartphone 40 according to the present embodiment. Chat processing unit 41b of processing unit 41 of smartphone 40 according to the present embodiment receives a request input by the user based on an operation on touch panel 44 (step S51). The chat processing unit 41b transmits the text information of the received request to the translation server 60 through the communication unit 43 (step S52).

After that, the chat processing unit 41b determines whether or not a response to the transmitted request has been received (step S53). If no response has been received (S53: NO), the chat processing unit 41b waits until a response is received. Here, the response received by the smartphone 40 is text information of the response to the request or an operator's instruction to switch to a manned response. If a response has been received (S53: YES), the chat processing unit 41b determines whether the received response is an instruction to switch to a manned response (step S54). If it is not a switching instruction (S54: NO), the chat processing unit 41b displays text information of the received response on the chat screen (step S55), and ends the process.

If the received response is a switching instruction (S54: YES), the chat processing unit 41b displays a message asking whether switching is possible or not, thereby inquiring of the user whether or not switching is possible (step S56). The chat processing unit 41b determines whether or not an affirmative answer to the inquiry has been obtained (step S57). If a response affirming the switch is obtained (S57: YES), the chat processing unit 41b transmits a request to switch to the manned response to the management server 30 via the translation server 60 (step S58), and executes the process. finish. If no answer affirming the switching is obtained (S57: NO), the chat processing unit 41b terminates the process without transmitting a switching request.

FIG. 16 is a flow chart showing the procedure of switching control processing performed by the management server 30 according to the present embodiment. The request acquisition unit 31a of the processing unit 31 of the management server 30 according to the present embodiment determines whether or not the communication unit 33 has received a request based on text information from the translation server 60 (step S71). If no request has been received (S71: NO), the request acquisition unit 31a waits until a request is received.

If a request has been received (S71: YES), the response processing unit 31b of the processing unit 31 inputs text information of the received request to the intention identification model 81 stored in the storage unit 32 (step S72). Next, the response processing unit 31b acquires the identification result of the intention of the request output by the intention identification model 81 (step S73). Based on the acquired identification result, the response processing unit 31b determines whether or not the intention of the request was identifiable (step S74). If the intention of the request can be identified (S74: YES), the response processing unit 31b refers to the response information DB 82 stored in the storage unit 32 based on the identified intention (step S75). The response processing unit 31b refers to the response information DB 82 and determines whether response information has been obtained (step S76). If the response information is obtained (S76: YES), the response processing unit 31b transmits the response information obtained from the response information DB 82 to the smartphone 40 that sent the request via the translation server 60 (step S77). exit.

If the intention of the request cannot be identified by the intention identification model 81 (S74: NO), or if response information for the intention of the request cannot be obtained from the response information DB 82 (S76: NO), the processing unit 31 The switching processing unit 31c transmits an instruction to switch from the automatic response of the chatbot to the manned response by the operator to the smartphone 40 that sent the request via the voice processing server 50 (step S78), and ends the process.

FIG. 17 is a flowchart showing a procedure of switching processing to manned response performed by the management server 30 according to the present embodiment. The switching processing unit 31c of the processing unit 31 of the management server 30 according to the present embodiment determines whether or not a request for switching to manned response has been received from the smartphone 40 via the translation server 60 (step S91). If a request for switching to manned response has not been received (S91: NO), the switching processing unit 31c waits until a request is received.

When receiving a request to switch to a manned response (S91: YES), the switching processing unit 31c acquires information on the date and time at that time (step S92). The date and time information can be obtained from, for example, an operating system operating on the management server 30 . The switching processing unit 31c refers to the operator DB 83 stored in the storage unit 32 based on the acquired date and time information (step S93). The switching processing unit 31c determines an operator for manned response based on the date and time information and the operator DB 83 (step S94). The switching processing unit 31c gives an instruction to respond to the operator terminal 70 used by the determined operator (step S95).

Thereafter, the processing unit 41 relays the request and the response by transmitting the user's request from the smartphone 40 to the operator terminal 70 and transmitting the operator's response from the operator terminal 70 to the smartphone 40 (step S96). . The processing unit 41 determines whether or not the manned response has ended (step S97). If the manned response has not ended (S97: NO), the processing unit 41 returns the process to step S96 and continues to relay requests and responses.

When the manned response has ended (S97: YES), the processing unit 41 inquires about the intention of the user's request to the operator terminal 70 that performed the manned response (step S98). Based on the operator's answer to this inquiry, the processing unit 41 generates teacher data in which the text information of the request and the intention of the request are associated, stores it in the storage unit 32 (step S99), and ends the process.

<Relearning the intention recognition model>
In the information providing system according to the present embodiment, learning (relearning) processing of the intention identification model 81 of the management server 30 is performed using teacher data created as a result of manned responses by the operator. Re-learning of the intention identification model 81 is performed periodically, such as once a month, or when the accumulated amount of teacher data exceeds a predetermined amount.

FIG. 18 is a flowchart showing the procedure of learning processing of the intention identification model 81 performed by the management server 30 according to this embodiment. The learning processing unit 31d of the processing unit 31 of the management server 30 according to the present embodiment determines whether or not it is time to re-learn the intention identification model 81 (step S111). If the timing for re-learning has not come yet (S111: NO), the learning processing unit 31d waits until the timing for re-learning.

When it is time to re-learn (S111: YES), the learning processing unit 31d reads the learning data stored in the storage unit 32 (step S112). The learning processing unit 31 d also reads out the intention identification model 81 stored in the storage unit 32 . The learning processing unit 31d performs processing for learning the intention identification model 81 using the read learning data (step S114). The learning processing unit 31d stores the learned intention identification model 81 for which the learning process has been completed in the storage unit 32 (step S115), and ends the process.

<Use before staying>
In the information providing system according to the present embodiment, information is provided not only to users (guests) who are staying at the accommodation facility, but also to users (for example, guests who make a reservation) before staying at the accommodation facility in a similar manner. is possible. FIG. 19 is a schematic diagram for explaining pre-stay information provision by the information provision system according to the present embodiment. For example, a user can make a reservation for lodging at an accommodation facility to the reservation server 80 . After the user completes the reservation for accommodation, the reservation server 80 transmits access information for accessing the information providing system that provides information on the accommodation facility to the user who made the reservation.

The transmission of the access information from the reservation server 80 to the user can be performed, for example, by sending an e-mail to the e-mail address registered as information relating to the user. The e-mail to the user may include, for example, a URL for accessing the information providing system or the management server 30 and identification information of the reserved accommodation facility as access information. The e-mail sent by the reservation server 80 is received by the user's smart phone 40, for example, and the user accesses the information providing system using the smart phone 40 based on the access information attached to the e-mail. In this case, the user can use the information provided by the automatic response of the chatbot provided by the management server 30 . Further, for example, an image of a two-dimensional code of access information may be attached to the e-mail. The user displays the two-dimensional code on the smartphone 40, reads the two-dimensional code using the camera 28 with the smart speaker 20 installed at home, and accesses the information providing system with the smart speaker 20. may In this case, the user can use the information provided by the voice response provided by the management server 30 through the smart speaker 20 .

Further, for example, the transmission of the access information from the reservation server 80 to the user may be by mail. A letter, card, or the like on which a two-dimensional code for access information is printed is enclosed in a guide for accommodation facilities or the like that is mailed to the user. The user can read the two-dimensional code with his/her smartphone 40 or smart speaker 20 and use the information providing system with his/her smartphone 40 or smart speaker 20 .

The management server 30 of the information providing system can make similar responses to requests from users regardless of whether the guest is a guest or a guest with a reservation. However, the management server 30 may limit the information given as a response to the user who made the reservation. The information providing system may also provide information to users other than guests and guests who make reservations, such as users who are considering lodging. In this case, the user can use the information providing system by accessing the website of the accommodation facility using his/her own smartphone 40 or smart speaker 20 and issuing a request such as an inquiry.

<Summary>
The information providing system according to the present embodiment having the above configuration acquires a request based on voice input from the user with the smart speaker 20 installed in the guest room of the accommodation facility, and transmits response information to the acquired request to the smart speaker 20. to output audio. When the response information for the acquired request cannot be output, the management server 30 switches to an automatic response by text input/output.

The case where the response information corresponding to the request cannot be output means that the response information corresponding to the user's request cannot be obtained from the response information DB 82 and an appropriate response cannot be given to the request. This is the case when it is not possible to make a unique response for each Specifically, for example, if the language spoken by the user is not included in the languages supported for speech processing by the speech processing server 50, or if the content of the user's speech cannot be appropriately converted into text information by the speech processing server 50, the user's request The case where the intention cannot be identified by the intention identification model 81, or the case where the response information corresponding to the response information DB 82 is not stored.

In addition, in response to a user's request, for example, when the intention of the request cannot be identified, a predetermined message such as "cannot respond to request" may be uniformly responded. Such a response is not a unique response for each request, nor is it an appropriate response to the request. Even when such a response is made, in the present embodiment, it is included in the case where the response information to the request cannot be output.

When switching from voice input/output to text input/output, the management server 30 causes the touch panel 25 of the smart speaker 20 to display access information 20a for accessing the automatic response by text input/output. This allows the user to use the access information 20a displayed on the smart speaker 20 to easily switch from voice input to text input/output.

In addition, management server 30 according to the present embodiment accepts access from smartphone 40 based on access information 20a displayed by smart speaker 20, acquires a request based on the text input to smartphone 40, and responds to the acquired request. The smartphone 40 is caused to output the response information as a text. This allows the user to input text using the familiar smartphone 40 and obtain a response to the request.

Moreover, the management server 30 according to the present embodiment switches to a manned response by an operator when response information cannot be output in response to a request based on the input of text information. As a result, the operator can reliably respond to user requests that cannot be responded to automatically.

In addition, the management server 30 according to the present embodiment stores the response of the operator according to the date and time in the operator DB 83, refers to the operator DB 83 based on the date and time information for switching, and determines the operator to which the manned response is switched. to decide. As a result, the management server 30 can appropriately determine the switching destination of the manned response according to, for example, the working hours of the operator center, and can reliably perform the manned response to the request from the user.

In addition, the management server 30 according to the present embodiment uses an intention identification model 81 that has been learned to accept information related to a user's request as an input, identify and output the intention of the request. The management server 30 acquires information related to the user's request and inputs it to the intention identification model 81, acquires the intention of the request output by the intention identification model 81, and transmits response information to the acquired intention of the request to the smart speaker 20 or The smartphone 40 is made to output. If response information cannot be output to the user's request and the operator switches to a manned response, the management server 30 re-learns the intention identification model 81 based on the result of this manned response. As a result, the identification ability of the intention identification model 81 can be improved, and the accuracy of the automatic response to the user's request can be improved.

Further, the management server 30 according to the present embodiment stores response information to requests in the response information DB 82, and the response information DB 82 is provided for each accommodation facility. The smart speaker 20 and smart phone 40 transmit identification information identifying the lodging facility and guest room along with the request entered by the user. Based on the identification information received together with the request from the user, the management server 30 uses the response information DB 82 corresponding to the accommodation facility to obtain response information to the request. The management server 30 causes the smart speaker 20 or the smartphone 40 to output the acquired response information. Thereby, the management server 30 can provide the user with response information suitable for the accommodation facility.

The information providing system according to the present embodiment also provides information to the user before lodging. For example, the reservation server 80 transmits access information to the smartphone 40 of the user who has reserved accommodation at the accommodation facility. The user accesses the information providing system using the smartphone 40 based on the access information, and the management server 30 receives access from the user's smartphone 40 based on the access information. Thereby, the information providing system can provide information not only to users who are staying at the accommodation facility but also to users who have not yet stayed at the accommodation facility. The user can obtain various information about the accommodation facility before lodging, and prepare for a trip or the like.

In the present embodiment, the user interface installed in the accommodation facility is the smart speaker 20, but the present invention is not limited to this. The user interface may be any device capable of inputting/outputting voice through a microphone and a speaker. For example, a personal computer to which a microphone and a speaker are connected can be used as a user interface. Also, in the present embodiment, the terminal device used by the user is the smart phone 40, but it is not limited to this. The terminal device may be, for example, a personal computer, a tablet terminal device, a mobile phone, etc., as long as it is capable of inputting and outputting text.

In addition, in this embodiment, before switching from voice response to chatbot automatic response and switching from chatbot automatic response to manned response by an operator, it is configured to inquire of the user whether or not to switch. However, it is not limited to this. The management server 30 may automatically perform these switching without making an inquiry to the user. Further, in the present embodiment, the smart speaker 20 is configured to display the access information 20a as a two-dimensional code, but the present invention is not limited to this. For example, the access information 20a may be displayed by the smart speaker 20 as character information including information such as URL, login ID and password. Any method may be used to output the access information 20a.

In the present embodiment, a plurality of server devices including the management server 30, the voice processing server 50, the translation server 60, and the reservation server 80 share processing to realize an information providing system. It is not limited to those mentioned above. For example, the plurality of server devices may be implemented as one server device, or the processing may be shared by, for example, five or more server devices. For example, the management server 30 may process the speech processing server 50 or the translation server 60 . Further, for example, the management server 30 may perform processing of the reservation server 80 . A server device that holds, learns, etc. the intention identification model 81 may be a device separate from the management server 30 .

In the present embodiment, the management server 30 is configured to respond to requests using the intention identification model 81 and the response information DB 82, but the configuration is not limited to this. For example, the management server 30 may use a learned model that receives input of text information of a request and outputs response information to the request. Alternatively, for example, the management server 30 may use a database in which request text information and response information are associated with each other, instead of using a trained model. Any method may be used by the management server 30 to respond to the request.

<Embodiment 2>
In the information providing system according to the above-described embodiment, the smart speaker 20 first performs a voice response to the user's request, and when it is determined that an appropriate voice response cannot be made, the response is switched to the chatbot, When it is further determined that the chatbot cannot give an appropriate response, the operator switches to a manned response. On the other hand, in the information providing system according to the second embodiment, when it is determined that the smart speaker 20 cannot make an appropriate voice response, the user selects whether to switch to the response by the chatbot or the manned response by the operator. can be selected.

In the information providing system according to Embodiment 2, a user staying at an accommodation facility makes a request by speaking to the smart speaker 20 provided in the guest room, and the information of the response to this request is transmitted by the voice of the smart speaker 20. can be obtained from the output. In this case, the voice response processing performed by the smart speaker 20, the management server 30, the voice processing server 50, etc. of the information providing system may be the same as that described in the first embodiment.

For example, when the language spoken by the user is not included in the corresponding language of the voice response, when it is determined that information cannot be provided by appropriate voice response, the management server 30 according to the second embodiment uses the chatbot Prompt the user to switch to the response by the operator or to the manned response by the operator. At this time, the management server 30 displays information for switching to a chatbot response and information for switching to a manned response on the display unit 25a of the smart speaker 20, and allows the user to select which response to switch to. .

FIG. 20 is a schematic diagram showing an example of access information displayed on the smart speaker 20 according to the second embodiment. In the information providing system according to the second embodiment, the smart speaker 20 displays on the touch panel 25 access information 20a for switching to a chatbot response and access information 20b for switching to a manned response at a call center. .

In the illustrated example, the smart speaker 20 displays a message "Please read the code of the desired response method" on the upper part of the substantially circular touch panel 25, and below this message is an access code for the chatbot. Information 20a and call center access information 20b are displayed side by side.

Note that the display method of the access information 20a and 20b shown in FIG. 20 is an example, and is not limited to this. For example, the smart speaker 20 displays options such as buttons or icons for accepting selection of either chatbot or call center on the touch panel 25, accepts selection of a response method from the user based on the touch operation for this option, and selects and responds. Only the access information 20 a and 20 b corresponding to the method may be displayed on the touch panel 25 . Further, for example, the smart speaker 20 may display information such as a telephone number for accessing the call center instead of the access information 20b.

The user can select a response method by reading either of the two pieces of access information 20 a and 20 b displayed on the smart speaker 20 with the camera of the smartphone 40 . When the access information 20a for the chatbot response is read, the smartphone 40 acquires information such as a URL from the access information 20a, accesses the chatbot, and provides the chatbot response service to the user. . Since the response by the chatbot is the same as in the above embodiment, detailed description is omitted.

When the access information 20b for manned response by the call center is read, the smartphone 40 acquires information such as a telephone number or URL from the access information 20b and accesses the call center (i.e., makes a call to the call center). ). In Embodiment 2, the operator of the call center and the user can perform voice communication (phone call) via the smartphone 40 by accessing the call center with the smartphone 40 . However, the operator and the user may chat with the operator via the operator terminal 70, the management server 30, etc. described in the above embodiments, instead of voice communication.

In the call center of the information providing system according to the second embodiment, in order to support multiple languages, multiple operators corresponding to each language are permanently stationed. The smart speaker 20 determines the language used by the user and displays access information 20b that enables voice communication with an operator corresponding to this language. The language used by the user may be determined by the smart speaker 20, by the management server 30, by the voice processing server 50, or by a device other than these.

Several methods may be employed to determine the language used by the user. A plurality of language determination methods will be described below, and the information providing system may use one or more of these methods for language determination.

(1) Determination from User's Speech The language used by the user is determined based on the content of words, words, sentences, or the like uttered by the user to the smart speaker 20 . Voice information related to the user's utterance is transmitted from the smart speaker 20 to the voice processing server 50, and the voice processing server 50 can convert the voice information into text information for the language to be processed. Therefore, the speech processing server 50 can determine whether or not the given speech information is in the language to be processed, and if it supports a plurality of languages, it can determine which language the speech information is in the language to be processed. can. If the voice information is not in the language to be processed, for example, the voice processing server 50 may only determine the language without converting it into text information. and the management server 30 may determine the language from the voice information, or the smart speaker 20 may determine the language from the voice information, for example.

To determine the language from the voice information, for example, a learning model that is machine-learned using teacher data in which the voice information and its language type are associated can be used. This learning model is trained for more languages than the languages to be processed by the speech processing server 50, receives speech information as an input, and outputs information regarding the language type of this speech information. The voice processing server 50, the management server 30, or the smart speaker 20 stores this learning model in advance, and can determine the language based on the voice information for languages not processed by the voice processing server 50.

(2) Judgment from guest information If the management server 30 or the computer of the accommodation facility has a database that manages information about guests, the language used by the user is judged based on the information stored in this database. be able to. For example, if the user's lodging information includes information on the country or place such as the user's address or hometown, etc., is stored in the database, the language corresponding to this country or place is determined to be the language used by the user. can be done.

A hotel management system called PMS (Property Management System) may be used for managing the guest information. The smart speaker 20 or management server 30 may communicate with the PMS's server to obtain information about the staying user and determine the language used by this user.

(3) Determination from smartphone information Information about the language set in the smartphone 40 owned by the user is acquired, and the language used by the user can be determined based on the acquired information. For example, the smart speaker 20 can exchange information with the smartphone 40 using wireless communication such as NFC (Near Field Communication), Bluetooth (registered trademark), or wireless LAN. The smartphone 40 is preset with the language used by the user, and the smart speaker 20 obtains the language setting information from the smartphone 40 by wireless communication to determine the language used by the user.

Note that when using the setting information of the smartphone 40 , the smart speaker 20 may not determine the language, but the smartphone 40 may determine the language. That is, the smartphone 40 that has read the access information 20b displayed on the smart speaker 20 determines the access destination according to the language set for itself when accessing the call center based on the access information 20b. may

(4) Selection by User The smart speaker 20 can receive a language selection operation from the user and determine the language used by the user. The smart speaker 20 displays, for example, a language selection box similar to the language selection box of the smartphone 40 shown in FIG. Acceptance of language selection by the smart speaker 20 may be performed at an appropriate timing. For example, the smart speaker 20 may accept language selection at the timing when the user visits an accommodation facility and uses the smart speaker 20 for the first time, and immediately before displaying the access information 20b for accessing the call center, for example. Language selection may be accepted at the timing of .

A device such as the smart speaker 20, the management server 30, the smartphone 40, or the voice processing server 50 according to Embodiment 2 adopts any one of the above (1) to (4) or appropriately combines a plurality of to determine the language used by the user. The smart speaker 20 displays access information 20b on the touch panel 25 according to the determination result of the used language. A user who has read this access information 20b with the smartphone 40 can use the access information 20b to access the call center, thereby performing voice communication with the operator in his/her own language.

Note that in the information providing system according to Embodiment 2, a device such as the smart speaker 20, the management server 30, the smartphone 40, or the voice processing server 50 determines the language used by the user before displaying the access information 20b. configuration, but not limited to this. For example, after the access information 20b displayed by the smart speaker 20 is read by the smartphone 40, and the user accesses the call center with the smartphone 40 using the access information 20b, the language used by the user may be determined. good. In this case, for example, a device such as a server that manages a call center determines the language used by the user by adopting any one of the above (1) to (4) or by appropriately combining a plurality of them, and the determination result voice communication between the operator and the user according to the

Further, the access information 20b displayed by the smart speaker 20 may display different information depending on, for example, each accommodation facility, each room of the accommodation facility, or each location of the accommodation facility. As a result, the call center can determine which lodging facility or the like the inquiry is from, according to the phone number dialed by the smartphone 40 . The access information 20b may display different information depending on, for example, the time or day of the week. may be displayed.

FIG. 21 is a flow chart showing the procedure of processing performed by the smart speaker 20 according to the second embodiment. The processing unit 21 of the smart speaker 20 according to the second embodiment transmits voice information regarding a request from the user to the voice processing server 50, and performs voice response processing of outputting voice information received from the voice processing server 50 from the speaker 27. (step S201).

The processing unit 21 determines whether or not the voice processing server 50 has given a switching instruction to switch from the voice response to the chatbot response or the call center response during the voice response process (step S202). If no switching instruction has been given (S202: NO), the processing unit 21 returns the process to step S201 and continues the voice response process.

If a switching instruction has been given (S202: YES), the user is asked whether switching is possible by outputting a message or the like inquiring about whether switching is possible from the speaker 27 or by displaying it on the touch panel 25 (step S203). The processing unit 21 determines whether or not an affirmative answer to the inquiry has been obtained (step S204). If no answer affirming the switching is obtained (S204: NO), the processing unit 21 returns the process to step S201 and continues the voice response process.

If an answer affirming the switching is obtained (S204: YES), the processing unit 21 determines the language used by the user (step S205). Note that the language determination may be performed at an earlier timing. The language determination may be performed by a device other than the smart speaker 20. In this case, the processing unit 21 of the smart speaker 20 obtains the determination result directly or indirectly from the device that performed the language determination.

Next, as shown in FIG. 11, the processing unit 21 sends access information 20a for switching to a response by a chatbot and access information 20b for switching to a manned response by a call center to the touch panel 25. is displayed (step S206), and the process ends.

FIG. 22 is a flowchart showing a procedure of voice communication start processing performed by smartphone 40 according to the second embodiment. The processing unit 41 of the smartphone 40 according to Embodiment 2 starts an application program, for example, according to a user's operation, and shifts to a mode of reading a two-dimensional code (step S211). In this mode, the access information acquisition unit 41a of the processing unit 41 captures the access information 20b displayed by the smart speaker 20 with the camera 45, thereby obtaining the access information 20b for switching to the manned response by the call center. is obtained (step S212). The access information acquisition unit 41a acquires information such as a telephone number included in the acquired access information 20b (step S213).

Next, the processing unit 41 accesses the call center based on the acquired information such as the telephone number (step S214). The processing unit 41 starts voice communication between the user and the call center (step S215), and ends the process.

The information providing system according to the second embodiment having the above configuration acquires a request based on the voice input from the user by the smart speaker 20 installed in the guest room of the accommodation facility, and transmits the response information to the acquired request to the smart speaker 20 to output audio. When response information to the acquired request cannot be output, the management server 30 and smart speaker 20 switch to either an automatic response by text input/output or a manned response by an operator.

At this time, the smart speaker 20 displays the access information 20a for switching to the automatic response by text input/output and the access information 20b for switching to the manned response by the operator, and prompts the user as to which switching is to be performed. let you choose. The user can select which response method to switch to by reading one of the access information 20 a and 20 b with his or her own smartphone 40 .

As a result, the user can select and switch between the response by the chatbot and the manned response by the operator, for example, when the desired result cannot be obtained with the voice response using the smart speaker 20 . Since the user can switch to a response method according to his or her preference, it can be expected that information suitable for the user will be provided.

Further, the information providing system according to the second embodiment determines the type of language used by the user, and switches to a manned response by an operator who can respond in the determined language. This allows the user to smoothly perform voice communication and the like with the operator.

In the information providing system according to Embodiment 2, the access information 20b displayed by the smart speaker 20 is read by the smartphone 40, and the user uses the smartphone 40 to perform voice communication with the operator of the call center. , but not limited to this. For example, when the user selects to switch to a manned response by a call center operator, the smart speaker 20 may access the call center and the smart speaker 20 may perform voice communication between the user and the operator.

Further, in the information providing system according to the second embodiment, it is not necessary to prepare for the operator's manned response by chat using the operator terminal 70 .

Further, other configurations of the information providing system according to Embodiment 2 are the same as those of the information providing system according to the previous embodiment, so the same parts are given the same reference numerals and detailed descriptions thereof are omitted. .

<Embodiment 3>
The information providing system according to the second embodiment described above switches to either an automatic response by text input/output or a manned response by an operator when response information to a request by voice cannot be output. ing. On the other hand, the information providing system according to Embodiment 3 first switches to an automatic response by text input/output when response information to a request by voice cannot be output. This can be done in the same procedure as in the information providing system according to the first embodiment.

The information providing system according to the first embodiment described above switches to a manned response by an operator using text input/output when an automatic response using text input/output cannot respond to a user's request. On the other hand, if the information providing system according to the third embodiment cannot respond to the user's request with an automatic response using text input/output, the same manned response using text input/output as the information providing system according to the first embodiment is performed. or to switch to voice manned response by a call center operator.

When the management server 30 of the information providing system according to the third embodiment cannot identify the intention of the request from the user, or when the response information for the intention of the request is not obtained, the automatic response of the chatbot , the operator transmits an instruction to switch to manned response to the smartphone 40 of the user. This corresponds to, for example, the processing shown in step S78 of the flowchart shown in FIG.

The smartphone 40 of the information providing system according to Embodiment 3 determines whether or not a switching instruction has been received from the management server 30 while the chatbot automatically responds by text input/output. This corresponds to, for example, the process shown in step S54 of the flowchart shown in FIG.

When a switching instruction is received from the management server 30, the smartphone 40 according to Embodiment 3 selects whether to switch to a manned response based on the operator's text input/output or to switch to a manned response based on the call center operator's voice input/output. accept. At this time, the smartphone 40 displays on the touch panel 44, for example, a message indicating switching to manned response, and a button labeled "chat (text input/output)" and a label "call center (voice input/output)". button with is displayed side by side. The smartphone 40 accepts selection of a manned response by text input/output or a manned response by voice input/output by receiving a touch operation on any button.

When switching to a manned response by text input/output is selected, the smartphone 40 transmits a request for switching to a manned response by text input/output to the management server 30 . This corresponds to, for example, the process shown in step S58 of the flowchart shown in FIG. The management server 30 that has received the request from the smartphone 40 makes a manned response by text input/output, for example, through the processing of the flowchart shown in FIG. 17 .

When switching to manned response by voice input/output is selected, smartphone 40 according to Embodiment 3 transmits a request for switching to manned response by voice input/output to management server 30 . Upon receiving this request, the management server 30 performs, for example, processing to determine the language used by the user, and determines a call center, operator, or the like suitable for the user. The management server 30 transmits to the smart phone 40 information such as a telephone number for making a call with the determined operator, or performs a process of establishing a call connection between the operator and the smart phone 40 . When receiving information such as a telephone number from the management server 30, the smartphone 40 accesses the call center based on the received information and starts voice communication with the operator.

The information providing system according to Embodiment 3 having the above configuration allows selection of switching to manned response by text input/output or manned response by voice input/output when the automatic response by text input/output cannot respond to the user's request. Receive from the user and switch to the selected manned response.

As a result, the user can switch to manned response by the operator when the desired result cannot be obtained from the automatic response by the chatbot. You can choose to do it. Since the user can switch to a response method according to his or her preference, it can be expected that information suitable for the user will be provided.

Since other configurations of the information providing system according to Embodiment 3 are the same as those of the information providing system according to the previous embodiment, the same parts are denoted by the same reference numerals, and detailed description thereof will be omitted.

The embodiments disclosed this time are illustrative in all respects and should be considered not restrictive. The scope of the present disclosure is indicated by the scope of the claims rather than the meaning described above, and is intended to include all changes within the scope and meaning equivalent to the scope of the claims.

10 information providing system 20 smart speaker (user interface)
20a Access information 21 Processing unit 22 Storage unit 22a Program 22b Identification information 23 Communication unit 25 Touch panel 25a Display unit 25b Input unit 26 Microphone 27 Speaker 28 Camera 30 Management server (information providing device)
31 processing unit 31a request acquisition unit 31b response processing unit 31c switching processing unit 31d learning processing unit 32 storage unit 32a server program 33 communication unit 40 smartphone (terminal device)
41 processing unit 41a access information acquisition unit 41b chat processing unit 42 storage unit 42a program 43 communication unit 44 touch panel 45 camera 50 voice processing server 60 translation server 70 operator terminal 80 reservation server 81 intention identification model (classifier)
82 Response information DB
83 Operator database
99 recording media

Claims (16)

retrieving voice-based requests entered at the property via the user interface;
Get the response information corresponding to the obtained request from the database,
causing the user interface to output the acquired response information by voice;
An information providing method, wherein when the response information cannot be obtained from the database, the user interface outputs access information for switching to an automatic response by text input/output. Receiving access from a terminal device based on the access information,
obtaining a request based on text entered into the terminal device;
Acquiring response information corresponding to the acquired request from the database;
causing the terminal device to output the acquired response information as text;
The information providing method according to claim 1. 3. The information providing method according to claim 2, wherein when response information corresponding to the text-based request cannot be obtained from the database, the response is switched to a manned response by an operator. Accepts selection of manned response by text input/output or manned response by voice input/output,
4. The information providing method according to claim 3, wherein switching to a manned response by an operator according to the accepted selection. 5. The information providing method according to claim 3, wherein the switching destination to the manned response is determined based on date, time or language information. Using a discriminator trained to accept information related to a user's request as an input and to discriminate and output the intention of the user's request,
inputting information related to a user's request to the discriminator and acquiring the intention of the request output by the discriminator;
Acquiring response information corresponding to the intent of the acquired request from the database;
Output the acquired response information,
When the operator switches to a manned response, re-learning the classifier based on the result of the manned response;
The information providing method according to any one of claims 3 to 5. Acquiring identification information that identifies the accommodation facility;
output response information corresponding to the obtained request and identification information;
The information providing method according to any one of claims 1 to 6. The database stores response information for each accommodation facility,
Acquiring response information corresponding to the accommodation facility from the database based on the identification information;
The information providing method according to claim 7. transmitting the access information to the user before staying at the accommodation facility;
Accepting access from the terminal device of the user based on the access information;
The information providing method according to any one of claims 1 to 8. When the response information cannot be obtained from the database, selectably display information for switching to an automatic response by inputting and outputting text and information for switching to a manned response by an operator.
The information providing method according to any one of claims 1 to 9. When the information for switching to one of the attended responses is selected,
determine the language used by the user;
Switching to manned response by an operator corresponding to the determined language,
The information providing method according to claim 10. a user interface;
a request acquisition unit that acquires a request based on voice input at an accommodation facility via the user interface; a response information acquisition unit that acquires from a database response information corresponding to the request acquired by the request acquisition unit; and the response information acquisition. A voice output processing unit that outputs the response information acquired by the unit to the user interface by voice, and an access for switching to an automatic response by text input/output when the response information cannot be acquired from the database. An information providing system comprising: an information providing device having a processing unit that performs processing for outputting information to the user interface. a request acquisition unit for acquiring a voice-based request entered at the accommodation facility via the user interface;
a response information acquisition unit that acquires response information corresponding to the request acquired by the request acquisition unit from a database;
A voice output processing unit that performs a process of outputting the response information acquired by the response information acquisition unit to the user interface by voice;
an information providing apparatus, comprising: a processing unit that performs a process of causing the user interface to output access information for switching to an automatic response by text input/output when the response information cannot be acquired from the database. to the computer,
retrieving voice-based requests entered at the property via the user interface;
Get the response information corresponding to the obtained request from the database,
causing the user interface to output the acquired response information by voice;
causing the user interface to output access information for switching to an automatic response by text input/output when the response information cannot be obtained from the database;
A computer program that causes a process to be performed. to the computer,
Acquire the access information output by the user interface at the accommodation facility,
Accessing an information providing device that provides information related to the accommodation facility based on the acquired access information,
get the request by text input,
Acquiring response information corresponding to the request from the information providing device and outputting it as text;
Accepting an operation related to switching to a manned response based on control from the information providing device;
A computer program that causes a process to be performed. 16. The computer program according to claim 15, causing a process of selecting a language for text input of said request and text output of said response information.

Images