Method for operating system for providing intelligent automated calling service

Abstract

A method for operating a system for providing an intelligent automated calling service is disclosed. The operation method comprises: a step in which a management server provides at least one user interface (UI) to a user terminal for inputting guidance information to be delivered to a target through an automated call; a step in which the management server transmits guidance information inputted through the user terminal to an automated calling system (ACS) server; and a step in which the ACS server executes an automated call to a target terminal and delivers the guidance information to the target terminal through the automated call. The present technology was developed through the "Development of AI Automated Calling System (ACS) Based on Startup Support Data" project of the Seoul Economic Promotion Agency under the 2023 Artificial Intelligence Technology Commercialization Support Program (CY230056).

Description

Method of operation of a system providing intelligent autocall service

[0001] The present disclosure relates to a method of operating a system, and more particularly, to a method of operating a system that provides an intelligent autocall service.

[0002] According to the startup trend announced by the Ministry of SMEs and Startups, 1.23 million companies are expected to be founded in 2023, and according to the Korea Startup Incubation Association's nationwide survey of startup incubation centers, there are 254 startup incubation centers, 87 of which are located in Seoul and the metropolitan area.

[0003] As such, the number of startup incubation centers is insufficient compared to the number of startup companies, resulting in low work efficiency due to inquiries and guidance calls regarding startup support programs. In addition, the lack of specialized personnel makes it difficult to compile the results of satisfaction surveys on events conducted by private companies.

[0004] Therefore, it is necessary to provide an automated response system to improve the efficiency of business processing operations and human resource management at startup incubation centers.

[0005] This technology was developed through the Seoul Economic Promotion Agency's 2023 Artificial Intelligence Technology Commercialization Support Project (CY230056) "Development of AI ACS (Auto Call System) Based on Startup Support Data."

[0006] The present disclosure seeks to increase the efficiency of administrative work at startup support organizations by digitally transforming telephone or consultation services at startup support organizations through the operating method of the system.

[0007] The purposes of the present disclosure are not limited to those mentioned above, and other purposes and advantages of the present disclosure not mentioned above can be understood through the following description and will be more clearly understood through the embodiments of the present disclosure. Furthermore, it will be readily apparent that the purposes and advantages of the present disclosure can be realized by the means and combinations thereof set forth in the claims.

[0008] According to one embodiment of the present disclosure, a method of operating a system providing an intelligent auto call service includes a step in which a management server provides a user terminal with at least one UI (User Interface) for inputting guidance information to be provided to a target through an auto call, a step in which the management server transmits guidance information input through the user terminal to an ACS (Auto Call System) server, and a step in which the ACS server executes an auto call for the target terminal and provides the guidance information to the target terminal through the auto call.

[0009] The method of operating the system may include a step in which the management server provides a first UI to the user's terminal for inputting at least one notice to be provided to the subject through the autocall, a step in which the management server transmits the notice input through the user's terminal to the ACS server, and a step in which the ACS server provides the notice to the subject's terminal through an autocall to the subject's terminal.

[0010] The method of operating the system may include a step in which the management server provides a second UI for inputting a plurality of items to be provided to the subject through the autocall and a provision scenario of the plurality of items to the user's terminal, a step in which the management server transmits the provision scenario input through the user's terminal to the ACS server, a step in which the ACS server provides at least one item among the plurality of items to the subject's terminal according to the provision scenario through an autocall to the subject's terminal, and a step in which the ACS server obtains a response from the subject to the item provided to the subject's terminal.

[0011] The step of providing at least one item to the terminal of the subject according to the provision scenario may include a step of the ACS server selecting one item from the plurality of items based on the provision scenario and providing the selected item to the terminal of the subject through a first chatbot, and the operating method of the system may include a step of the ACS server identifying a response item corresponding to the acquired response from among a plurality of response items matching the item provided to the terminal of the subject, and a step of the ACS server selecting one item from the plurality of items based on the provision scenario and the identified response item and providing the selected item to the terminal of the subject through the first chatbot when the response item corresponding to the acquired response is identified.

[0012] The method of operating the above system includes a step of providing consultation through a second chatbot linked with a generative model when a response item corresponding to the acquired response is not identified, and the step of providing consultation through the second chatbot includes: when a response item corresponding to the acquired response is not identified, inputting the acquired response into the generative model, and providing consultation through the second chatbot according to an output of the generative model.

[0013] The operating method of the above system may include a step of inputting the obtained response into a third artificial intelligence model for generating keyword-based guidance information when a response item corresponding to the obtained response is not identified, and a step of providing the outputted guidance information through a third chatbot linked with the third artificial intelligence model when guidance information matching a keyword included in the obtained response is output from the third artificial intelligence model.

[0014] The method of operating the above system may include a step of the ACS server extracting at least one word included in the obtained response when a response item corresponding to the obtained response is not identified, and a step of the ACS server providing the terminal of the subject with a supplementary service corresponding to the identified keyword when a keyword related to a supplementary service is identified among the extracted words.

[0015] The method of operating the above system may include a step in which the ACS server executes an autocall for terminals of a plurality of subjects to obtain a response from a terminal of at least one subject among the terminals of the plurality of subjects, and a step in which the ACS server transmits an autocall history including a response item corresponding to the obtained response and a start time and an end time of each autocall for the terminals of the plurality of subjects to the management server.

[0016] The method of operating the above system may include a step in which the management server, based on the autocall history, identifies a major autocall for which a response to a question set in the last order according to the provision scenario has been obtained among autocalls executed for each of the terminals of the plurality of subjects, a step in which the management server divides the plurality of autocalls into time zones according to the start time of the autocalls for each of the terminals of the plurality of subjects, and a step in which the management server calculates an autocall completion rate based on the ratio of major autocalls among autocalls matching each time zone, based on the time zone.

[0017] The method of operating the system according to one embodiment of the present disclosure can efficiently improve the administrative work of a startup support organization by performing various administrative tasks such as event participation surveys, satisfaction surveys, and questionnaires.

[0018] In addition, through the operating method of the system according to one embodiment of the present disclosure, artificial intelligence learning data can be constructed to activate AI business by using classification data based on speech data and intention.

[0019] FIG. 1 is a diagram illustrating the configuration of a system according to an embodiment of the present disclosure;

[0020] Figure 2 is a flowchart for explaining the operation process of the system according to one embodiment of the present disclosure;

[0021] Figures 3a to 3d are diagrams showing examples of UIs provided by a management server to a user's terminal according to one embodiment of the present disclosure;

[0022] FIG. 4 is a flowchart illustrating an operation in which an ACS server provides a question to a subject's terminal according to an embodiment of the present disclosure;

[0023] FIG. 5 is a diagram for explaining the first to third chatbots operated by the ACS server according to one embodiment of the present disclosure;

[0024] FIG. 6 is a diagram illustrating an example of a UI for a management server to provide statistical information on responses to questions according to one embodiment of the present disclosure;

[0025] FIG. 7 is a diagram illustrating an example of a UI for a management server to provide statistical information of autocall according to one embodiment of the present disclosure; and

[0026] FIG. 8 is a diagram illustrating an example of a format in which a management server provides a user with an autocall completion rate by time zone according to one embodiment of the present disclosure.

[0027] Before describing the present disclosure in detail, the description method of the specification and drawings will be described.

[0028] First, the terms used in this specification and claims are general terms selected based on their functions in the various embodiments of the present disclosure. However, these terms may vary depending on the intentions of those skilled in the art, legal or technical interpretations, and the emergence of new technologies. Furthermore, some terms may have been arbitrarily selected by the applicant. These terms may be interpreted according to the meanings defined in this specification. In the absence of a specific definition, they may be interpreted based on the overall content of this specification and common technical knowledge in the relevant field.

[0029] Additionally, the same reference numbers or symbols in each drawing attached to this specification represent parts or components that perform substantially the same functions. For convenience of explanation and understanding, the same reference numbers or symbols are used in different embodiments. In other words, even if components with the same reference numbers are all depicted in multiple drawings, the multiple drawings do not necessarily represent a single embodiment.

[0030] Additionally, terms including ordinal numbers, such as "first," "second," etc., may be used in this specification and claims to distinguish between components. These ordinal numbers are used to distinguish identical or similar components from each other, and the use of these ordinal numbers should not be interpreted in a limited manner. For example, components associated with these ordinals should not be restricted in their order of use or arrangement by their numbers. If necessary, each ordinal number may be used interchangeably.

[0031] In this specification, singular expressions include plural expressions unless the context clearly dictates otherwise. In this application, terms such as "comprise" or "consist of" are intended to indicate the presence of a feature, number, step, operation, component, part, or combination thereof described in the specification, but should be understood not to preclude the possibility of the presence or addition of one or more other features, numbers, steps, operations, components, parts, or combinations thereof.

[0032] In the embodiments of the present disclosure, terms such as "module," "unit," "part," etc. are terms used to refer to components that perform at least one function or operation, and such components may be implemented as hardware or software, or a combination of hardware and software. In addition, a plurality of "modules," "units," "parts," etc. may be integrated into at least one module or chip and implemented as at least one processor, except in cases where each needs to be implemented as a separate, specific hardware.

[0033] Additionally, in the embodiments of the present disclosure, when a part is said to be connected to another part, this includes not only a direct connection but also an indirect connection through another medium. Furthermore, unless specifically stated otherwise, the statement that a part includes a certain component does not exclude other components, but rather implies that other components may be included.

[0034] FIG. 1 is a diagram illustrating the configuration of a system according to one embodiment of the present disclosure.

[0035] Referring to FIG. 1, the system (1000) may be composed of a management server (100) and an ACS (Auto Call System) server (200).

[0036] The management server (100) is a server for providing a UI (User Interface) related to auto call to a user's terminal (300) or a target's terminal (400) or for managing information input through a UI provided to the user's terminal (300) or the target's terminal (400).

[0037] As an example, the management server (100) may provide a UI for inputting guidance information to be provided to a target through an autocall to the user's terminal (300), and may transmit the information input through the UI to the ACS server (200).

[0038] Autocall is a service that automatically makes a call for a specific purpose to deliver information to the recipient of the call or to receive information from the recipient. The system (1000) may provide an autocall service to deliver information to the caller who made the call to a number connected to the ACS server (200) or to receive information from the caller.

[0039] The ACS server (200) is a server for providing guidance information by executing an auto call to the target's terminal (400).

[0040] As an example, the ACS server (200) may execute an auto call to the terminal (400) of the subject based on the information of the subject acquired according to user input, and provide guidance information (e.g., notices, satisfaction survey questions, etc.) received from the management server (100) to the subject.

[0041] At this time, the ACS server (200) can obtain an audio signal including the subject's spoken voice by executing an autocall, and by transmitting the obtained audio signal to the management server (100), the management server (100) can provide the audio signal including the subject's spoken voice to the user's terminal (300).

[0042] The management server (100) and / or the ACS server (200) may be implemented as a server device or system including at least one computer. It is also possible for the management server (100) and / or the ACS server (200) to be implemented as a terminal device such as a desktop PC, a laptop PC, a tablet PC, or a smartphone.

[0043] The management server (100) may include memory, a processor, a communication unit, etc.

[0044] Additionally, the ACS server (200) may include memory, a processor, a communication unit, etc.

[0045] Memory is a configuration for storing an operating system (OS) that controls the overall operation of each server's components and at least one instruction or data related to each server's components.

[0046] Memory may include non-volatile memory such as ROM and flash memory, and may include volatile memory such as DRAM. Memory may also include a hard disk, a solid state drive (SSD), etc.

[0047] The processor is a configuration that controls each server as a whole.

[0048] In one embodiment, the processor may include a general-purpose processor such as a CPU (Central Processing Unit), an AP, a DSP (Digital Signal Processor), a graphics-only processor such as a GPU (Graphics Processor Unit), a VPU (Vision Processing Unit), or an AI-only processor such as an NPU (Neural Processing Unit). The AI-only processor may be designed with a hardware structure specialized for training or utilizing a specific AI model.

[0049] The communication department is a component for communicating with the outside world.

[0050] As an example, the management server (100) can transmit guidance information to the ACS server (200) through a communication unit included in the management server (100).

[0051] As an example, the ACS server (200) can transmit information about autocall history to the management server (100) through a communication unit included in the ACS server (200).

[0052] The communication unit may include circuits, modules, chips, etc. for performing communication using various wired and wireless communication methods. The communication unit may also be connected to external devices and servers via various networks.

[0053] Depending on the area or scale, a network may be a personal area network (PAN), a local area network (LAN), or a wide area network (WAN), and depending on the openness of the network, it may be an intranet, an extranet, or the Internet.

[0054] The communication unit can be connected to external devices and servers through various wireless communication methods such as LTE (long-term evolution), LTE-A (LTE Advance), 5G (5th Generation) mobile communication, CDMA (code division multiple access), WCDMA (wideband CDMA), UMTS (universal mobile telecommunications system), WiBro (Wireless Broadband), GSM (Global System for Mobile Communications), DMA (Time Division Multiple Access), WiFi (Wi-Fi), WiFi Direct, Bluetooth, BLE (Bluetooth Low Energy), NFC (near field communication), Zigbee, and LoRa.

[0055] Additionally, the communication unit can be connected to external devices and servers via wired communication methods such as Ethernet, optical network, Universal Serial Bus (USB), and ThunderBolt.

[0056] In addition, the communications department may be structured to utilize various new communication methods / technologies that will be developed in the future.

[0057] FIG. 2 is a flowchart for explaining the operation process of a system according to one embodiment of the present disclosure.

[0058] Referring to FIG. 2, the management server (100) can provide a UI (User Interface) to the user's terminal (300) for inputting guidance information to be provided to the subject through an auto call (S210).

[0059] As an example, the management server (100) may provide a first UI to the user's terminal (300) for inputting at least one notice to be provided to the target through an autocall.

[0060] For example, notices to be provided to the target audience may include, but are not limited to, information on startup support projects (e.g., number of selected participants, amount of support, project schedule, etc.), information on startup support centers (e.g., location, operating hours, etc.), and information on events (e.g., event schedule, event location, etc.).

[0061] As an example, the management server (100) may provide a second UI to the user's terminal (300) for inputting multiple items to be provided to the subject through an autocall and a provision scenario for the multiple items.

[0062] For example, multiple items may include, but are not limited to, items for a satisfaction survey on startup support programs.

[0063] The presentation scenario may include the order in which multiple items are presented, multiple response items for each of the multiple items, and at least one item to be presented based on the response items.

[0064] Meanwhile, the management server (100) may provide a third UI to the user's terminal (300) for entering information on the person to receive the auto call.

[0065] The information of the subject may include, but is not limited to, the name, contact information, address, etc. of each of the multiple subjects.

[0066] The management server (100) can receive guidance information input from the user's terminal (300) through the UI (S220).

[0067] For example, the management server (100) may receive guidance information including at least one notice from the user's terminal (300) by providing the first UI to the user's terminal (300).

[0068] For example, the management server (100) may receive guidance information including a plurality of items and a provision scenario of the plurality of items from the user's terminal (300) by providing a second UI to the user's terminal (300).

[0069] Meanwhile, the UI provided to the user's terminal (300) is not limited to the above-described example, and multiple UIs may be provided to the user's terminal (300).

[0070] The management server (100) can transmit guidance information entered through the UI to the ACS server (200) (S230).

[0071] Specifically, the management server (100) can transmit guidance information input by the user's terminal (300) through the UI to the ACS server (200) so that the ACS server (200) can perform an auto call to the target's terminal (400).

[0072] The ACS server (200) can execute an auto call to the target's terminal (400) (S240).

[0073] As an example, the ACS server (200) can execute an autocall for the target's terminal (400) by making a call to the target's terminal (400) based on the target's information obtained according to user input.

[0074] Meanwhile, the ACS server (200) can also obtain information on the subject by linking with an external database.

[0075] As an additional example, when a call made by the subject's terminal (400) is received by a number connected to the ACS server (200), the ACS server (200) may execute an autocall to the subject's terminal (400) to provide guidance information to the subject's terminal (400).

[0076] The ACS server (200) can provide guidance information to the target's terminal (400) through an auto call (S250).

[0077] As an example, when the ACS server (200) receives information about a notice from the management server (100), it can provide the notice to the target's terminal (400) through an auto call to the target's terminal (400).

[0078] As an additional example, when the ACS server (200) receives information about a plurality of items and a provision scenario of the plurality of items from the management server (100), it can provide at least one item to the subject's terminal (400) according to the received provision scenario.

[0079] At this time, the ACS server (200) can obtain the subject's response to the question provided to the subject's terminal (400).

[0080] For example, the ACS server (200) can obtain a response from the subject consisting of an audio signal including the subject's spoken voice by performing an autocall on the subject's terminal (400).

[0081] Meanwhile, if an input of an audio signal including the subject's spoken voice is identified while the question is being provided to the subject's terminal (400), the ACS server (200) can stop providing the question and obtain the audio signal.

[0082] At this time, the ACS server (200) may select a question to be provided to the subject's terminal (400) based on the acquired audio signal, or may provide consultation or an answer regarding the acquired audio signal through the second or third chatbot described later.

[0083] FIGS. 3A to 3D are diagrams illustrating examples of a UI provided by a management server to a user terminal according to one embodiment of the present disclosure.

[0084] Referring to FIGS. 3a and 3b, the management server (100) can provide a UI for inputting at least one item and multiple response items for the item to the user's terminal (300).

[0085] At this time, the management server (100) can provide a UI including settings for a mode (e.g., key input mode, voice input mode, etc.) for identifying a response item corresponding to the subject's response to the user's terminal (300).

[0086] Specifically, referring to FIG. 3a, the management server (100) may provide the user's terminal (300) with a UI for setting a key input mode that identifies a response item corresponding to the subject's response based on a number entered through an input unit (e.g., keypad, touchpad, etc.) of the subject's terminal (400) and inputting at least one item for the key input mode and a response item for the item.

[0087] Meanwhile, when the key input mode is selected, the management server (100) may provide a fourth UI for key input to the target's terminal (400) when an auto call to the target's terminal (400) is executed by the ACS server (200).

[0088] In addition, referring to FIG. 3b, the management server (100) may provide a user terminal (300) with a UI for inputting at least one question for the voice input mode, a response item for the question, and a keyword for each response item by identifying whether a keyword is included for each response item in the subject's response composed of an audio signal including the subject's spoken voice, and identifying a response item corresponding to the subject's response by setting a voice input mode, and inputting the keyword for each response item.

[0089] Referring to FIG. 3c, the management server (100) may provide a UI to the user's terminal (300) for setting an autonomous speech mode in which an audio signal including the subject's spoken voice is input as a response of the subject to a question, and for inputting at least one question for the autonomous speech mode.

[0090] When a subject's response to a question for which an autonomous speech mode is set is obtained, the ACS server (200) can match the subject's response with the question and add it to the autocall history without identifying the response item corresponding to the subject's response, and the ACS server (200) can transmit the autocall history to the management server (100), and the management server (100) can provide the autocall history to the user's terminal (300).

[0091] Referring to FIG. 3d, the management server (100) may provide a UI for inputting at least one notice (e.g., greeting, guidance message, event summary, etc.) and the order in which the notices are provided to the user's terminal (300).

[0092] In addition, the management server (100) may provide at least one UI including a first UI element for setting a connection item for each response item for each question to the user's terminal (300).

[0093] For example, when a response item for which a connection item is set according to a response of a subject is identified, the management server (100) may provide a UI including a first UI element for setting the connection item set for the response item to be provided to the user's terminal (300).

[0094] Accordingly, when a response item for which a connection item is set according to the subject's response is identified, the ACS server (200) can provide the subject's terminal (400) with the connection item set for the corresponding response item, rather than the item set in the next order of the item last provided to the subject's terminal (400).

[0095] For example, if a question asking about the reason for satisfaction with the response item of 'Very much so' among the response items for the question 'Overall, I was satisfied with the program' is set as a linked question according to user input, and if the response item of 'Very much so' is identified according to the subject's response to the question 'Overall, I was satisfied with the program', the ACS server (200) can provide the question asking about the reason for satisfaction with the response item of 'Very much so' set as a linked question to the subject's terminal (400), and if the response item of 'Average' is identified, the ACS server (200) can provide the question set in the next order of the question of 'Overall, I was satisfied with the program' to the subject's terminal (400) based on the provision scenario.

[0096] In addition, the management server (100) may provide at least one UI to the user's terminal (300) including a second UI element for inputting a reference time and a reference time for inputting a reference time for inputting a non-input comment notifying that the input time has been exceeded, in order to provide a non-input comment notifying that the input time has been exceeded if a response from the subject is not obtained within the reference time.

[0097] In addition, the management server (100) may provide at least one UI including a third UI element for inputting a number-exceeded comment to the user's terminal (300) to provide a number-exceeded comment to the user's terminal (400) to notify that the number of inputs has been exceeded when multiple response items are identified according to the subject's response.

[0098] Additionally, the management server (100) may provide at least one UI to the user's terminal (300) including a fourth UI element for inputting the speed at which guidance information is provided to the target's terminal (400).

[0099] The speed at which guidance information is provided to the subject's terminal (400) means the speed at which audio data including guidance information is provided through at least one of the first to third chatbots described below.

[0100] FIG. 4 is a flowchart illustrating an operation in which an ACS server provides a question to a subject's terminal according to one embodiment of the present disclosure.

[0101] Referring to FIG. 4, the ACS server (200) can select one item from among multiple items based on the provided scenario (S410).

[0102] Specifically, the ACS server (200) can select one item from among the multiple items based on the order set according to the provision scenario, as the management server (100) receives a plurality of items and a provision scenario of the multiple items obtained by providing a second UI to the user's terminal (300).

[0103] The ACS server (200) can provide the selected items to the subject's terminal (400) (S420).

[0104] As an example, the ACS server (200) can provide the selected item to the subject's terminal (400) through the first chatbot (10).

[0105] The first chatbot (10) is configured to provide at least one item to the terminal (400) of the subject. The first chatbot (10) can convert an item composed of text data into audio data through a TTS (Text To Speech) module and provide it to the terminal (400) of the subject.

[0106] More detailed information about this will be described later with reference to Figure 5.

[0107] The ACS server (200) can obtain the subject's response to the question from the subject's terminal (400) (S430).

[0108] As an example, the ACS server (200) can obtain a response from the subject consisting of an audio signal including the subject's spoken voice from the subject's terminal (400).

[0109] As an additional example, when a key input mode is selected by the user's terminal (300), the ACS server (200) may obtain a frequency (DTMF, Dual-Tone Multi-Frequency) for the input sound of the subject's terminal (400) through an autocall to the subject's terminal (400), and may obtain a response from the subject by identifying a number corresponding to the obtained frequency.

[0110] For this purpose, the frequency for the input sound by number may be stored in the ACS server (200).

[0111] As an additional example, when a key input mode is selected by the user's terminal (300) and the fourth UI is provided to the subject's terminal (400) by the management server (100), the ACS server (200) can receive the subject's response to the question from the subject's terminal (400) from the management server (100).

[0112] The ACS server (200) can identify a response item corresponding to the subject's response (S440).

[0113] In one embodiment, when a voice input mode is selected by a user's terminal (300), the ACS server (200) converts the subject's response, which is composed of an audio signal including the subject's spoken voice, into text data through the STT (Speech To Text) module included in the first chatbot (10), and identifies a response item corresponding to the subject's response based on the subject's response converted into text data and keywords for each response item.

[0114] Specifically, the ACS server (200) compares the vector extracted from the subject's response converted into text data and the vector of each keyword for each response item, and when a vector matching the vector of the keyword among the vectors extracted from the subject's response is identified, the response item matching the keyword can be identified as a response item corresponding to the subject's response.

[0115] To this end, the ACS server (200) can extract multiple vectors from text data and compare the extracted vectors to identify matching text.

[0116] For example, the ACS server (200) can measure the angle between the vector of the keyword for each response item and the vector extracted from the subject's response using cosine similarity, which divides the inner product between vectors by the product of the vector sizes, and if the measured angle is less than a threshold, it can be determined that the subject's response includes the keyword.

[0117] As an additional example, when the key input mode is selected by the user's terminal (300), the ACS server (200) can identify a response item corresponding to a number according to the subject's response.

[0118] The ACS server (200) can select one item from among multiple items based on the identified response items (S450).

[0119] Specifically, when a response item corresponding to the subject's response is identified among a plurality of response items matching the item provided to the subject's terminal (400), the ACS server (200) can select one item from among the plurality of items.

[0120] As an example, the ACS server (200) may select a question set in the next order of the question last provided through the first chatbot (10) among a plurality of questions based on the response items corresponding to the provided scenario and the target's response.

[0121] As an additional example, the ACS server (200) may select an item (linked item) set to be provided for a response item corresponding to the subject's response among a plurality of items based on the provided scenario and the response item corresponding to the subject's response.

[0122] Meanwhile, a case in which a response item corresponding to the subject's response is not identified among multiple response items matching the question provided to the subject's terminal (400) will be described later with reference to FIG. 5.

[0123] The ACS server (200) can provide selected items to the subject's terminal (400) based on the provided scenario and response items corresponding to the subject's response (S460).

[0124] Specifically, the ACS server (200) can provide selected items based on the provided scenario and response items corresponding to the subject's response to the subject's terminal (400) through the first chatbot (10).

[0125] FIG. 5 is a diagram illustrating first to third chatbots operated by an ACS server according to one embodiment of the present disclosure.

[0126] Referring to FIG. 5, the ACS server (200) may include at least one chatbot among the first chatbot (10), the second chatbot (20), and the third chatbot (30).

[0127] The first chatbot (10) is configured to provide at least one question to the target's terminal (400).

[0128] The first chatbot (10) may include, but is not limited to, a TTS (Speech To Text) module, an STT (Text To Speech) module, an NLP (Natural Language Processing) module, etc.

[0129] The STT module is a module for converting audio data into text data.

[0130] The STT module can generate text data by extracting features from audio data, identifying multiple phonemes based on the extracted features, and converting the identified phonemes into words and sentences.

[0131] The TTS module is a module for converting text data into audio data.

[0132] The TTS module can generate audio data by analyzing the grammatical structure of text data, dividing it into sentence units, converting each word into multiple phonemes, generating voice based on the converted phonemes, and converting it into a digital signal.

[0133] The NLP module is a module for identifying the structure and meaning of text data through natural language processing.

[0134] The NLP module performs a preprocessing process on text data (e.g., tokenization (dividing text into units such as words and sentences), removing stop words, etc.), and then extracts features from the preprocessed text data (e.g., frequency of word appearance, weighting based on importance of words (Term Frequency-Inverse Document Frequency), correlation between words, etc.) to identify the meaning of sentences and proper nouns contained in the text data, thereby enabling understanding of the context of the text data.

[0135] For example, an NLP module can perform a preprocessing step of dividing text data into sentence units, identifying the part of speech or sentence component of each word in each sentence included in the text data, and extracting words excluding stop words (e.g., particles, suffixes, adjectives, etc.).

[0136] At this time, the NLP module can extract features for text data by converting each extracted word into a vector form and determining that the closer the distance between the vectors of each word, the higher the frequency of co-occurrence, and determining that the smaller the angle between the vectors, the higher the correlation.

[0137] For example, the first chatbot (10) can convert a question composed of text data into audio data through a TTS module and provide it to the subject's terminal (400).

[0138] For example, the first chatbot (10) can convert the subject's response into text data through the STT module, and identify whether a keyword is included in each response item that matches the subject's response converted into text data, thereby identifying a response item corresponding to the subject's response.

[0139] Specifically, the first chatbot (10) can extract at least one vector from the subject's response converted into text data, measure the angle between the vector of the keyword for each response item and the vector extracted from the subject's response, and identify a response item that matches a keyword for which the measured angle between the vector extracted from the subject's response and the keyword is less than a threshold as a response item corresponding to the subject's response.

[0140] As an additional example, the first chatbot (10) can identify response items that match the subject's response through the STT module and the NLP module.

[0141] Specifically, the first chatbot (10) can convert the subject's response into text data through the STT module, and identify response items that match the subject's response by identifying the characteristics of the subject's response converted into text data through the NLP module.

[0142] As an additional example, the first chatbot (10) can identify requirements included in the subject's response (e.g., adjusting the playback speed of audio data providing guidance information or the volume providing guidance information) through the STT module and the NLP module.

[0143] For example, the first chatbot (10) can identify the needs of the subject by converting the subject's response into text data through the STT module and identifying the characteristics of the subject's response converted into text data through the NLP module, and can adjust the speed at which guidance information is provided or the volume at which guidance information is provided according to the identified needs.

[0144] Meanwhile, if the ACS server (200) does not obtain the subject's response to the guidance information provided to the subject's terminal (400) within a preset time, the first chatbot (10) may reduce the playback speed of audio data containing voice corresponding to the guidance information and provide it again to the subject's terminal (400).

[0145] The second chatbot (20) is configured to provide consultation to the target's terminal (400) based on the output of the generative model by linking with the generative model.

[0146] The second chatbot (20) may include, but is not limited to, an STT module, a TTS module, an NLP module, etc.

[0147] For example, the second chatbot (20) can convert a question composed of text data into audio data through a TTS module and provide it to the subject's terminal (400).

[0148] As an additional example, the second chatbot (20) can convert the subject's response into text data through the STT module and input it into a generative model.

[0149] Meanwhile, a generative model is a large language model (LLM) that performs natural language processing (NLP), and can be an artificial intelligence model trained to understand the context of data acquired according to user input, predict the next data in the sequence of data acquired according to user input based on the understood context, and generate result data (e.g., text, audio, etc.) that includes the predicted data.

[0150] At this time, the generative model may be a model that has been trained and stored in the ACS server (200), or may be a model acquired by the ACS server (200) connecting to an external server.

[0151] As an example, the ACS server (200) can provide consultation through a second chatbot (20) linked to a generative model when a response item corresponding to the subject's response to a question provided to the subject's terminal (400) is not identified.

[0152] Specifically, the ACS server (200) converts the subject's response to a question provided to the subject's terminal (400) into text data through the STT module and inputs the data into a generative model, and provides consultation to the subject's terminal (400) through the second chatbot (20) according to the output of the generative model.

[0153] Counseling is provided to the subject's terminal (400) in response to the subject's response to the question provided to the subject's terminal (400), and the ACS server (200) inputs the subject's response obtained by providing counseling to the subject's terminal (400) through a second chatbot (20) into a generative model and provides the output information to the subject's terminal (400) through a second chatbot (20) linked to the generative model to provide counseling.

[0154] Meanwhile, the ACS server (200) inputs the obtained response of the subject into a generative model by providing consultation through the second chatbot (20) to the subject's terminal (400), and if the subject is identified as satisfied based on the output information, the consultation can be terminated.

[0155] At this time, when the consultation through the second chatbot (20) is finished, the ACS server (200) can provide the items set in the next order of items provided to the subject's terminal (400) through the first chatbot (10) based on the provided scenario.

[0156] The third chatbot (30) is configured to be linked with a third artificial intelligence model for generating keyword-based guidance information and to provide information output by the third artificial intelligence model to the target's terminal (400).

[0157] The third chatbot (30) may include, but is not limited to, an STT module, a TTS module, an NLP module, etc.

[0158] For example, the third chatbot (30) can convert a question composed of text data into audio data through a TTS module and provide it to the subject's terminal (400).

[0159] As an additional example, the third chatbot (30) can convert the subject's response into text data through the STT module and input it into the third artificial intelligence model.

[0160] Meanwhile, the third artificial intelligence model may be a model trained to extract keywords included in input data and identify and output guidance information matching the extracted keywords.

[0161] Specifically, the third artificial intelligence model can extract multiple words from input data, convert each extracted word into a vector form, and identify the frequency of occurrence of words and the degree of association between words based on the vector of each word to extract at least one keyword.

[0162] For example, the third artificial intelligence model determines that the closer the distance between vectors, the higher the frequency of co-occurrence, and the smaller the angle between vectors, the higher the correlation. Therefore, words that frequently co-occur and have high correlation within text data can be selected through the vectors of each word and identified as keywords.

[0163] At this time, the third artificial intelligence model is trained based on the guidance information and can generate and output guidance information matching the keyword.

[0164] Specifically, the third artificial intelligence model can extract multiple words included in guidance information acquired through training or guidance information acquired based on user input, convert each extracted word into a vector form, and measure the angle between the vector of each keyword extracted from the subject's response and the vector of each multiple word included in the guidance information.

[0165] At this time, if the measured angle is less than the threshold, the third artificial intelligence model can determine that the keyword extracted from the subject's response matches the word included in the guidance information, and generate output data based on the guidance information that includes the word that matches the keyword extracted from the subject's response.

[0166] Meanwhile, the third artificial intelligence model may be a model that uses various learning algorithms such as Natural Language Processing (NLP), Convolutional Neural Networks (CNN), and Term Frequency-Inverse Document Frequency (TF-IDF), but is not limited thereto.

[0167] As an example, if a response item corresponding to the subject's response to a question provided to the subject's terminal (400) is not identified, the ACS server (200) may input the subject's response into a third artificial intelligence model for generating keyword-based guidance information.

[0168] At this time, when guidance information matching the keyword included in the response is output from the third artificial intelligence model, the ACS server (200) can provide the information output from the third artificial intelligence model to the subject's terminal (400) through the third chatbot (30) linked with the third artificial intelligence model.

[0169] Additionally, the first chatbot (10) can convert the subject's response, which is composed of audio data, into text data through the STT module, and then identify the meaning of the subject's response (e.g., the chatbot is frustrating, the question is interesting, etc.) through the NLP module.

[0170] At this time, if the meaning of the subject's response identified through the NLP module includes a negative response to the autocall (e.g., slow response, frustrating, inaccurate, that's not it, etc.), the ACS server (200) inputs the subject's response into the third artificial intelligence model, and when guidance information matching the keywords included in the subject's response is output from the third artificial intelligence model, the guidance information output by the third artificial intelligence model can be provided through the third chatbot (30).

[0171] As an additional example, the ACS server (200) may extract at least one word included in the subject's response if a response item corresponding to the subject's response is not identified.

[0172] At this time, if a keyword related to an additional service is identified among the words extracted from the subject's response, the ACS server (200) can provide an additional service corresponding to the identified keyword to the subject's terminal (400).

[0173] Specifically, the ACS server (200) can convert the subject's response into text data through the STT module included in the first chatbot (10) that provided the question to the subject's terminal (400) and extract at least one word from the subject's response.

[0174] At this time, the ACS server (200) can identify whether there is a match between the extracted word and a keyword set in relation to the additional service.

[0175] Specifically, the ACS server (200) can convert the extracted words and preset keywords into vector form and measure the angle between the vector of each word extracted from the subject's response and the vector of each preset keyword.

[0176] At this time, the ACS server (200) can provide an additional service corresponding to a vector of words extracted from the subject's response and a preset keyword for which an angle less than a threshold value is measured to the subject's terminal (400).

[0177] Additional services may include, but are not limited to, phone calls with users, chats with users, etc.

[0178] Meanwhile, as autocall is executed for the terminals of multiple subjects, the ACS server (200) records the number of times for which the subject's response to each question is not obtained within a preset time, and can identify a question for which the recorded number of times exceeds the preset number as a target question.

[0179] At this time, the ACS server (200) can identify whether the target item includes target words such as foreign languages, technical terms, and abbreviations.

[0180] For example, the ACS server (200) is connected to a server storing dictionary information for each language, and can identify whether each word included in the target item is listed in the dictionary, thereby identifying whether a target item includes a foreign language.

[0181] Additionally, the ACS server (200) can convert each of a plurality of words into a prototype through the NLP module and identify whether each word converted into a prototype is listed in the dictionary.

[0182] For example, the ACS server (200) is connected to a server that stores information on specialized terms by field, and can identify whether each word included in the target question is a specialized term, thereby identifying whether the target question includes a specialized term.

[0183] For example, the ACS server (200) is connected to a server that stores dictionary information in which abbreviations and full names are mapped, and can identify whether each word included in the target item is an abbreviation, thereby identifying whether the target item includes an abbreviation.

[0184] At this time, if it is identified that the target word is included in the target item, the ACS server (200) can obtain explanation information about the target word included in the target item and store it as an additional explanation for the target word.

[0185] Explanatory information for the target word may include translation information for foreign languages, explanatory information for technical terms, and original word information for abbreviations.

[0186] For example, the ASC server (200) can obtain translation information for a foreign language from a server in which dictionary information for each language is stored.

[0187] Additionally, the ASC server (200) can obtain explanatory information on specialized terms from a server that stores information on specialized terms by field.

[0188] Additionally, the ASC server (200) can obtain full-text information for an abbreviation from a server that stores dictionary information in which abbreviations and full-text are mapped.

[0189] As an additional example, the ACS server (200) may obtain information on whether a target word is included and explanation information for the target word by transmitting the target item to the user's terminal (300).

[0190] Afterwards, if the target item is provided to the subject's terminal but the subject's response is not obtained within a preset time, the ACS server (200) can provide an additional explanation for the target word.

[0191] For example, if a target item is provided to the subject's terminal (400) through the first chatbot (10) but the subject's response is not obtained within a preset time, the ACS server (200) may convert an additional explanation of the target word composed of text data into audio data through a TTS module and provide the same to the subject's terminal (400).

[0192] As a result, the system (1000) can provide detailed information about the items to increase the subject's understanding, thereby obtaining accurate responses to the items from the subject, thereby constructing reliable response data.

[0193] FIG. 6 is a diagram illustrating an example of a UI for a management server to provide statistical information on responses to questions according to one embodiment of the present disclosure.

[0194] Referring to FIG. 6, the management server (100) can calculate the number of responses (6) per response item based on the responses obtained by the ACS server (200) executing an auto call for the terminals of multiple subjects and provide the result to the user's terminal (300).

[0195] Specifically, the ACS server (200) can execute an auto call for terminals of multiple subjects and obtain a response from the terminal of at least one of the terminals of the multiple subjects.

[0196] At this time, the ACS server (200) can transmit an autocall history including a response item corresponding to the acquired response and the start and end times of each autocall for the terminals of multiple targets to the management server (100).

[0197] Accordingly, the management server (100) can calculate the number of responses (6) for each response item based on the number of responses for each response item included in the autocall history and provide the result to the user's terminal (300).

[0198] FIG. 7 is a diagram illustrating an example of a UI for a management server to provide statistical information of autocalls according to one embodiment of the present disclosure.

[0199] Referring to FIG. 7, the management server (100) can provide statistical information including at least one of the number of completed autocalls, the number of abandoned autocalls, the number of voice recognition failures, the number of calls in progress, and the number of unanswered autocalls to the user's terminal (300).

[0200] Specifically, the management server (100) can identify an autocall for which a response to a question set in the last order according to the provision scenario has been obtained based on the autocall history as a major autocall, and identify the number of major autocalls as the number of completed cases.

[0201] In addition, the management server (100) can identify the number of autocalls in which a response to the question set in the first order according to the provision scenario was obtained but a response to the question set in the last order was not obtained as the number of abandoned autocalls based on the autocall history.

[0202] In addition, the management server (100) can identify the number of autocalls in which an autocall was executed but a response to a question set in the first order according to the provision scenario was not obtained as the number of unanswered autocalls based on the autocall history.

[0203] In addition, the management server (100) can identify the number of autocalls that the ACS server (200) attempted to execute but failed to execute based on the autocall history as the number of calls in progress.

[0204] In addition, the management server (100) can identify the number of autocalls in which at least one item was provided according to the provision scenario but a response item corresponding to the subject's response was not identified as the number of voice recognition failures based on the autocall history.

[0205] FIG. 8 is a diagram illustrating an example of a format in which a management server provides a user with an autocall completion rate by time zone according to one embodiment of the present disclosure.

[0206] Referring to FIG. 8, the management server (100) can calculate an autocall completion rate based on the autocall history and provide it to the user's terminal (300) in a chart format.

[0207] Specifically, the management server (100) can classify multiple autocalls into time zones based on the autocall start time for each terminal of multiple targets based on the autocall history.

[0208] At this time, the management server (100) can calculate the autocall completion rate for each time zone based on the ratio of the main autocall among the autocalls matching each time zone.

[0209] As an additional example, the management server (100) may classify multiple autocalls by date based on the autocall start time for each terminal of multiple subjects based on the autocall history, and classify multiple autocalls by day of the week based on the day of the week information for each date.

[0210] For example, the management server (100) can obtain day-of-the-week information by user input.

[0211] At this time, the management server (100) can calculate the autocall completion rate for each day of the week based on the ratio of major autocalls among the autocalls matching each day of the week.

[0212] Meanwhile, the various embodiments described above may be implemented in a recording medium readable by a computer or similar device using software, hardware, or a combination thereof.

[0213] In terms of hardware implementation, the embodiments described in the present disclosure may be implemented using at least one of Application Specific Integrated Circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), processors, controllers, micro-controllers, microprocessors, and other electrical units for performing functions.

[0214] In some cases, the embodiments described herein may be implemented within the processor itself. In a software implementation, the embodiments described herein, such as the procedures and functions described herein, may be implemented as separate software modules. Each of the software modules described above may perform one or more of the functions and operations described herein.

[0215] Meanwhile, computer instructions for performing processing operations in electronic devices and the like according to the various embodiments of the present disclosure described above may be stored in a non-transitory computer-readable medium. When the computer instructions stored in such a non-transitory computer-readable medium are executed by a processor of a specific device, they cause the specific device to perform processing operations according to the various embodiments described above.

[0216] A non-transitory computer-readable medium refers to a medium that permanently stores data and can be read by a device, rather than a medium that stores data for a short period of time, such as a register, cache, or memory. Specific examples of non-transitory computer-readable media include CDs, DVDs, hard disks, Blu-ray discs, USBs, memory cards, and ROMs.

[0217] Although the preferred embodiments of the present disclosure have been illustrated and described above, the present disclosure is not limited to the specific embodiments described above, and various modifications may be made by a person having ordinary skill in the art to which the present disclosure pertains without departing from the gist of the present disclosure as claimed in the claims, and such modifications should not be understood individually from the technical idea or prospect of the present disclosure.

Claims

1. In the operating method of a system providing an intelligent auto call service, A step in which the management server provides at least one UI (User Interface) to the user's terminal for inputting guidance information to be provided to the target through an autocall; The step of the management server transmitting the guidance information entered through the user's terminal to the ACS (Auto Call System) server; and A method of operating a system, comprising: a step in which the ACS server executes an autocall for the terminal of the subject and provides the guidance information to the terminal of the subject through the autocall.

2. In paragraph 1, The method of operation of the above system is as follows: A step in which the management server provides a first UI to the user's terminal for inputting at least one notice to be provided to the subject through the autocall; A step in which the management server transmits a notice entered through the user terminal to the ACS server; and A method of operating a system, comprising: a step in which the ACS server provides the notice to the terminal of the subject through an autocall to the terminal of the subject.

3. In paragraph 1, The method of operation of the above system is as follows: A step in which the management server provides a second UI to the user's terminal for inputting a plurality of items to be provided to the subject through the autocall and a provision scenario of the plurality of items; A step in which the management server transmits the provision scenario input through the user terminal to the ACS server; The step of the ACS server providing at least one item among the plurality of items to the terminal of the subject according to the provision scenario through an auto call to the terminal of the subject; and A method of operating a system, comprising: a step of the ACS server obtaining a response from the subject to a question provided to the subject's terminal; 4. In paragraph 3, The step of providing at least one item to the terminal of the subject according to the provision scenario is: The above ACS server comprises a step of selecting one of the plurality of items based on the provided scenario and providing the selected item to the terminal of the subject through the first chatbot; The method of operation of the above system is as follows: The step of the ACS server identifying a response item corresponding to the acquired response among a plurality of response items matching the items provided to the terminal of the subject; and A method of operating a system, comprising: a step of the ACS server selecting one item from the plurality of items based on the provision scenario and the identified response item, and providing the selected item to the terminal of the subject through the first chatbot, when a response item corresponding to the acquired response is identified.

5. In paragraph 4, The method of operation of the above system is as follows: Including a step of providing consultation through a second chatbot linked to a generative model when a response item corresponding to the above-mentioned acquired response is not identified; The step of providing consultation through the above second chatbot is: A method of operating a system in which, if a response item corresponding to the acquired response is not identified, the acquired response is input into the generative model, and consultation is provided through the second chatbot according to the output of the generative model.

6. In paragraph 4, The method of operation of the above system is as follows: If a response item corresponding to the obtained response is not identified, a step of inputting the obtained response into a third artificial intelligence model for generating keyword-based guidance information; and A method of operating a system, comprising: a step of providing the outputted guidance information through a third chatbot linked with the third artificial intelligence model when guidance information matching a keyword included in the response obtained from the third artificial intelligence model is outputted.

7. In paragraph 4, The method of operation of the above system is as follows: If a response item corresponding to the acquired response is not identified, the ACS server extracts at least one word included in the acquired response; and A method of operating a system, comprising: a step in which, when a keyword related to an additional service is identified among the extracted words, the ACS server provides an additional service corresponding to the identified keyword to the terminal of the subject.

8. In paragraph 3, The method of operation of the above system is as follows: A step in which the ACS server executes an auto call to terminals of multiple subjects to obtain a response from at least one terminal of the multiple subjects; and A method of operating a system, comprising: a step in which the ACS server transmits, to the management server, a response item corresponding to the acquired response and an autocall history including the start time and end time of each autocall for the terminals of the plurality of targets.

9. In paragraph 8, The method of operation of the above system is as follows: A step in which the management server identifies, based on the autocall history, a major autocall for which a response to a question set in the last order according to the provision scenario has been obtained among the autocalls executed for each of the terminals of the plurality of subjects; A step in which the management server divides the plurality of autocalls into time zones according to the start time of the autocall for each terminal of the plurality of targets; and A method of operating a system, comprising: a step of the management server calculating an autocall completion rate based on the ratio of major autocalls among autocalls matching each time zone, for each time zone.

Images