KR20220081120A - Artificial intelligence call center system and method for providing system-based service - Google Patents

Abstract

In the service providing method based on the artificial intelligence call center system of the present invention, the voice recognition module performs voice recognition on the conversation voice between the civil servant and the counselor input from the communication equipment, and receives the voice recognition-based text data corresponding to the conversation voice. generating; extracting, by the dialogue analysis module, intention analysis data indicating the intention of the civil petitioner from the speech recognition-based text data using natural language processing technology; transmitting, by the reception content management module, the report type classification information generated by classifying the report type based on the speech recognition-based text data and the intention analysis data to an agent interface; and transmitting, by the agent interface, a corresponding manual mapped to the report type classification information by searching the integrated storage to the agent.

Description

Artificial intelligence call center system and system-based service providing method {Artificial intelligence call center system and method for providing system-based service}

The present invention relates to a call center system, and more particularly, to an artificial intelligence call center system for receiving reports based on voice recognition and responding to emergency situations.

Counselors working in the call center system for receiving reports and responding to emergencies summarize and record complaints from complaints, categorize reports/incident types, and appropriately identify complaints’ requests and situations, etc. Accordingly, it is necessary to respond promptly to the emergency situation of the complainant.

However, the task of summarizing and recording the content of the report and the task of classifying the report/incident type are handled manually by the counselor, and may take a long time depending on the experience of the counselor. This is a factor that makes it difficult to quickly respond to emergency situations.

SUMMARY OF THE INVENTION In order to solve the above problems, an object of the present invention is to use voice recognition technology to summarize and record report contents, classify report/incident types, and other necessary tasks in a call center system for receiving reports and responding to emergency situations. An object of the present invention is to provide an artificial intelligence call center system that can be automatically processed using the system and a service provision method based on the system.

The above and other objects, advantages and features of the present invention, and a method for achieving them will become apparent with reference to the embodiments described below in detail in conjunction with the accompanying drawings.

In a service providing method based on an artificial intelligence call center system according to an aspect of the present invention, a voice recognition module performs voice recognition on a conversation voice between a civil servant and a counselor input from a communication device, generating text data; extracting, by the dialogue analysis module, intention analysis data indicating the intention of the civil petitioner from the speech recognition-based text data using natural language processing technology; transmitting, by the reception content management module, the report type classification information generated by classifying the report type based on the speech recognition-based text data and the intention analysis data to an agent interface; and transmitting, by the agent interface, a corresponding manual mapped to the report type classification information by searching the integrated storage to the agent.

According to another aspect of the present invention, there is provided a service providing method based on an artificial intelligence call center system, in which a voice recognition module performs voice recognition on an agent's conversational voice input through an agent interface to generate text data based on the agent's voice. ; extracting, by the dialogue analysis module, intention analysis data indicating the intention of the civil servant from the counselor's voice-based text data using natural language processing technology; transmitting, by the reception content management module, report type classification information generated by classifying a report type based on the agent's voice-based text data and the intention analysis data to an agent interface; and transmitting, by the agent interface, a corresponding manual mapped to the report type classification information by searching the integrated storage to the agent.

The artificial intelligence call center system according to another aspect of the present invention performs voice recognition on a conversation voice between a civil servant and an agent input from the communication equipment or an agent's voice input from the agent UI, voice recognition-based text data a voice recognition module for generating; a dialogue analysis module for extracting intention analysis data representing the intention of the civil petitioner from the speech recognition-based text data using natural language processing technology; and a reception content management module that classifies the report type based on the speech recognition-based text data and the intention analysis data and transmits the report type classification information obtained to the agent interface, wherein the agent UI includes an integrated storage A corresponding manual mapped to the report type classification information is retrieved and delivered to the counselor.

The artificial intelligence call center system according to the present invention records the conversation between the complainant and the counselor in real time using voice recognition technology in the process of receiving a report, and provides all or part of the recorded contents to the counselor, thereby reducing the report reception processing time can be shortened

In addition, since the artificial intelligence call center system according to the present invention performs voice recognition centered on the voice of the counselor, when voice recognition is performed centered on the voice of the complainant in a situation where the call quality of the complainant is poor, the accuracy of voice recognition is lowered. It is possible to improve the accuracy of report/incident type classification through voice recognition of the agent's voice, and at the same time improve the accuracy of the corresponding code classification.

Ultimately, the artificial intelligence call center system of the present invention automatically processes related tasks using voice recognition technology, thereby reducing the burden of agents and reducing call traffic, as well as providing quick and quick response to customer service satisfaction. can be improved

1 is a block diagram showing the internal configuration of an artificial intelligence call center system according to a first embodiment of the present invention.
Fig. 2 is a block diagram showing the internal configuration of the reception content management module shown in Fig. 1;
3 is a block diagram illustrating an internal configuration of an artificial intelligence call center system according to a second embodiment of the present invention.
4 is a block diagram illustrating an internal configuration of an artificial intelligence call center system according to a third embodiment of the present invention.
5 is a block diagram illustrating an internal configuration of an artificial intelligence call center system according to a fourth embodiment of the present invention.
6 is a block diagram illustrating an internal configuration of an artificial intelligence call center system according to a fifth embodiment of the present invention.
7 is a flowchart illustrating a service providing method of an artificial intelligence call center system according to an embodiment of the present invention.
8 is a flowchart illustrating a service providing method of an artificial intelligence call center system according to another embodiment of the present invention.
9 is a block diagram of an exemplary computing device to which the service providing method of the artificial intelligence call center system of the present invention can be applied.

Specific structural or functional descriptions of the embodiments according to the concept of the present invention disclosed in this specification are only exemplified for the purpose of explaining the embodiments according to the concept of the present invention, and the embodiments according to the concept of the present invention are It may be implemented in various forms and is not limited to the embodiments described herein.

Since the embodiments according to the concept of the present invention may have various changes and may have various forms, the embodiments will be illustrated in the drawings and described in detail herein. However, this is not intended to limit the embodiments according to the concept of the present invention to specific disclosed forms, and includes changes, equivalents, or substitutes included in the spirit and scope of the present invention.

The terminology used herein is used only to describe specific embodiments, and is not intended to limit the present invention. The singular expression includes the plural expression unless the context clearly dictates otherwise. In the present specification, terms such as "comprise" or "have" are intended to designate that the described feature, number, step, operation, component, part, or combination thereof exists, and includes one or more other features or numbers, It should be understood that the possibility of the presence or addition of steps, operations, components, parts or combinations thereof is not precluded in advance.

1 is a block diagram showing the internal configuration of an artificial intelligence call center system according to a first embodiment of the present invention, and FIG. 2 is a block diagram showing the internal configuration of the reception content management module shown in FIG. 1 .

Referring to FIG. 1 , an artificial intelligence call center system 200 according to an embodiment of the present invention is connected to communicate with a user terminal 100 through a network 150 .

The user terminal 100 may be a wired or wireless communication terminal possessed by the civil servant, and the wireless communication terminal may be, for example, a smart phone, and if it can communicate with the artificial intelligence call center system 200, there is a special limitation on the type there is no

The network 150, as a medium connecting the user terminal 100 and the artificial intelligence call center system 200, is, for example, a public switched telephone network (PSTN), an Internet network (Internet Protocol network: IP network), etc. can

Artificial intelligence call center system 200, communication equipment 210, recording data storage 220, agent interface 230, artificial intelligence agent module 240, reception content management module 250 and integrated data storage 260 includes

The communication equipment 210 is configured to communicate with the user terminal 100 of the complainant through the network 150 , and provides a call connection with a counselor. The communication equipment 210 may be, for example, a predictive dialing system (PDS), a computer telephony integration (CTI), an interactive voice response (IVR), a private branch exchange (PBX), or the like. In addition, the communication device 210 acquires voice data by recording the voice conversation between the civil servant and the counselor. The voice data recorded by the communication device 210 is stored in a database built in the recording data storage 220 .

The agent interface 230 is configured to provide an interfacing function between the agent and the communication equipment 210, an interfacing function between the agent and the reception content management module 240, and an interfacing function between the agent and the storage 260, for example, It may be a desktop personal computer. A headset, a microphone speaker, etc. may be connected to the desktop personal computer for the counselor to consult with the civil servant.

The artificial intelligence agent module 240 includes a voice recognition module 241 , a conversation analysis module 242 , and a conversation management module 243 .

The voice recognition module 241 outputs a voice recognition result obtained by performing voice recognition on voice data input in real time through a communication device to the conversation analysis module 242 . Here, the voice recognition result may be text data converted from the voice including the voice of the civil servant and the voice of the counselor.

The voice recognition module 241 may be, for example, a software module or a hardware module executing the software module. Here, the software module may be, for example, a well-known speech recognition algorithm or a speech recognition model trained in advance through machine learning.

The speech recognition model may be an artificial neural network model trained in advance by using a large amount of data classified for each domain as training data (or training data) in order to improve speech recognition performance.

The dialog analysis module 242 processes text data (conversation contents between the complainant and the counselor) input from the voice recognition module 241 according to natural language processing technology, and analyzes the request and intention of the complainant. For example, the conversation analysis module 242 uses a conversation corpus in which intention values and/or request values specific to the call center system are tagged, and the request of the complainant and/or the counselor from the text data representing the conversation between the complainant and the counselor and intention analysis.

The conversation management module 243 collects analysis result data (hereinafter referred to as intention analysis data of the civil servant and/or counselor) obtained by analyzing the dialogue analysis module 242, and receives the collected intention analysis data from the reception content management module (250) is output. In addition, the conversation management module 243 may collect text data converted from the conversation voice between the civil servant and the counselor through the voice recognition process of the voice recognition module 241 , and output it to the reception content management module 250 . .

Here, the text data acquired by the voice recognition module 241 is used as information representing the entire conversation between the complainant and the counselor, and the intention analysis data acquired by the conversation analysis module 242 is the entire conversation content between the complainant and the counselor. It can also be used as summary information.

Meanwhile, in FIG. 1 , the voice recognition module 241 , the conversation analysis module 242 , and the conversation management module 243 included in the artificial intelligence agent module 240 are separated, but may be integrated into one software module. For example, the dialogue analysis module 242 and the dialogue management module 243 are integrated into one artificial neural network model, or the speech recognition module 241 , the dialogue analysis module 242 and the dialogue management module 243 are all in one. can be integrated into the artificial neural network model of

The reception content management module 250 stores the text data input from the artificial intelligence agent module 240 in the consultation conversation database built in the integrated data storage 260 .

In addition, the reception content management module 250 automatically classifies the report/event type based on the text data and/or intention analysis data input from the artificial intelligence agent module 240, and the classification result (report type classification information) It may be provided to the agent through the agent interface 230 .

In addition, the reception content management module 250 uses the text data input from the agent interface 230 to the text data input from the AI agent module 240 (hereinafter, referred to as 'speech recognition-based text data'). can be corrected Here, the text data input from the agent interface 230 (text data based on the agent UI) is written (inputted) by the agent using an input means (eg, keyboard, etc.) provided in the agent interface 230 . It could be a conversation.

The reception contents management module 250 is configured to perform automatic classification of report/incident types and correction of text data input from the agent interface 230 (conversation contents of civil servants directly written and input by the agent). 2 , it may be configured to include a classifier 252 and a corrector 254 .

The classifier 252 automatically classifies the report/event type using the speech recognition-based text data and intention analysis data input from the artificial intelligence agent module 240 . The classifier 252 may be implemented as an artificial neural network model trained in advance using, for example, training text data and training intention analysis data as inputs. Here, the artificial neural network model may be, for example, an artificial neural network (ANN), a recurrent neural network (RNN), a convolutional neural network (CNN), or a deep neural network (DNN). have.

The corrector 254 corrects the text data input by the agent inputted from the agent interface 230 by using the voice recognition-based text data input from the artificial intelligence agent module 240 . The corrector 254 may be implemented as a pre-trained artificial neural network model using, as inputs, voice recognition-based training text data and training text data written by a counselor. Here, the artificial neural network model may be, for example, an artificial neural network (ANN), a recurrent neural network (RNN), a convolutional neural network (CNN), or a deep neural network (DNN). have.

The correction method using the artificial neural network model, for example, calculates the similarity between words included in text data written by an agent and words included in speech recognition-based text data, and then text data directly input by the agent. It may be to replace a word having a similarity less than or equal to a threshold among words included in the . Here, for calculating the similarity between words, a word similarity calculation based on a word embedding vector (eg, word2vector) or an algorithm related to the sentence similarity calculation may be used.

After correcting the text data written by the counselor, the reception content management module 250 may store the corrected text data in the counseling conversation database built in the integrated data storage 260 .

In addition, the reception content management module 250 transmits the corrected text data back to the conversation analysis module 242 of the artificial intelligence agent module 240, and transmits the corrected text (corrected text data based on the agent UI). The received conversation analysis module 242 may re-analyze the request and intention of the complainant and/or the counselor based on the corrected text data.

Thereafter, the dialogue analysis module 242 transmits the intention analysis data re-analyzed through the dialogue management module 243 to the reception contents management module 250 again, and the reception contents management module 250 is based on the re-analyzed intention analysis data to automatically classify the report/incident type again.

On the other hand, the voice recognition-based text data input from the AI counselor module 240 includes text data obtained by recognizing the voice of the civil servant and text data obtained by recognizing the voice of the counselor.

Due to the call quality of the user terminal 100 , the text data obtained by recognizing the voice of the civil servant may have lower accuracy than the text data obtained by recognizing the voice of the counselor.

Accordingly, when the reception content management module 250 corrects the text data written (or inputted) by the counselor, among the text data obtained by recognizing the civil servant's voice and the text data obtained by recognizing the counselor's voice, the counselor By applying a higher weight to the text data obtained by recognizing the voice of may be

On the other hand, the agent accesses the scenario database built in the integrated data storage 260 using the agent interface 230 and delivers necessary information to respond to an emergency situation to the civil servant according to the response manual stored in the scenario database.

3 is a block diagram illustrating an internal configuration of an artificial intelligence call center system according to a second embodiment of the present invention.

Referring to FIG. 3 , the internal components 210 to 260 of the artificial intelligence call center system 200 according to the second embodiment are the internal components 210 of the artificial intelligence call center system according to the first embodiment shown in FIG. 1 . 260) is almost the same. However, in that the artificial intelligence agent module 240A included in the artificial intelligence call center system 200 according to the second embodiment performs voice recognition only for the agent's voice, the artificial intelligence agent module 240 according to the first embodiment ) is different from

In the system configuration of FIG. 1 , the call quality of the user terminal 100 carried by the civil petitioner may be low. In this case, in order to respond based on accurate information, the counselor will proceed with the dialogue with the complainant by asking the complainant's query again and answering the question. For this reason, it is possible to analyze the request and intention of the civil petitioner only by performing voice recognition on the voice uttered by the counselor.

The artificial intelligence agent module 240A according to the second embodiment shown in FIG. 3 performs voice recognition only on the agent's voice and analyzes the voice recognition result, so the artificial intelligence agent module 240 shown in FIG. 1 and Alternatively, it is not connected to the communication equipment 210, but is connected to the agent UI 230 in order to use only the agent's voice as an input.

Similar to the artificial intelligence agent module 240 shown in FIG. 1 , the voice recognition module 241A included in the artificial intelligence agent module 240A according to the second embodiment performs voice recognition on the agent's voice to perform text The data is generated, and the conversation analysis module 242A analyzes text data corresponding to the counselor's voice and generates intention analysis data obtained by analyzing the request and intention of the civil petitioner.

The conversation management module 243A according to the second embodiment collects the text data for the agent's voice acquired by the voice recognition module 241A and the intention analysis data acquired by the dialog analysis module 242A, and receives it It is transmitted to the management module 250 .

In addition, the conversation management module 243A according to the second embodiment may summarize the contents of the conversation between the complainant and the counselor by using the collected text data and intention analysis data.

In order to summarize the conversation between the complainant and the counselor, the conversation between the complainant and the counselor can be summarized from text data and intention analysis data using a template specialized for call service. For example, the generation of summary data summarizing the conversation between the civil servant and the counselor is achieved by extracting key keywords from among the words included in the text data using intention analysis data, and inputting the extracted keywords into the template. It is possible. Here, the summary data may be referred to as template data in that a template specialized for a call service is used to summarize the conversation between the civil servant and the counselor.

Such summary data or template data may be generated using a pre-trained artificial neural network model. Here, the artificial neural network model for generating the summary data or template data may be previously trained (trained) to generate summary data or template data using text data for training and intention analysis data for training as inputs.

The reception contents management module 250A according to the second embodiment similarly to the reception contents management module 250 according to the first embodiment, the text data input from the artificial intelligence agent module 240A is stored in the integrated storage 260 . It is saved in the established counterpart conversation DB.

In addition, the reception content management module 250A according to the second embodiment uses the intention analysis data and summary data (template data) input from the artificial intelligence agent module 240A together with the text data to determine the report/incident reception type. automatically classify.

In addition, the reception content management module 250A according to the second embodiment uses text data obtained by performing voice recognition on only the agent's voice to directly write (input) text by the agent using the agent UI 230 . Correct the data.

4 is a block diagram illustrating an internal configuration of an artificial intelligence call center system according to a third embodiment of the present invention.

Referring to FIG. 4 , in the artificial intelligence call center system 200 according to the third embodiment, the first and second artificial intelligence agent modules 240B are configured to further include a customer-facing knowledge/information generation module 244B. There is a difference from the artificial intelligence agent modules 240 and 240A according to the embodiment.

For the remaining components except for the customer-facing knowledge/information generating module 244B, the descriptions of the components shown in FIGS. 1 and 3 will be substituted, and hereinafter, the customer-facing knowledge/information generating module 244B will be focused and described.

The speech recognition-based text data, intention analysis data, and summary data (template data) collected and generated by the conversation management module 243B are the reception contents management module 250 or 250A and the reception contents customer response knowledge/information generation module (244B).

The customer response knowledge/information generation module 244B includes voice recognition-based text data, intention analysis data and summary data provided from the conversation management module 243B, and corrected text data and reports/events provided from the reception content management module 250B. Based on the type classification result, the scenario database built in the integrated storage 260 is searched to search for information for customer service in real time, and the retrieved information is provided to the agent through the agent UI 230 .

The agent UI 230 can search for information in the integrated storage (domain knowledge DB, scenario DB, counseling conversation DB, etc. built in 260), and this information providing log remains and can be recommended when responding to civil servants next time.

5 is a block diagram illustrating an internal configuration of an artificial intelligence call center system according to a fourth embodiment of the present invention.

Referring to FIG. 5 , in the artificial intelligence call center system 200 according to the fourth embodiment, the artificial intelligence agent module 240C further includes a response generating module 244C and a voice utterance module 245C serving as a voice chatbot. It is different from the above-described embodiments in that it is configured to include.

First, the reception contents management module 250 finally determines the contents of the report reception and categorizes the report/incident type. Here, the final determined report reception content may be data obtained by correcting text data directly written by the agent through the agent UI using text data generated by the artificial intelligence agent module 240C based on voice recognition.

According to the fourth embodiment, the counselor may receive the final determined report reception content through the counselor UI 230 and respond to the complaints' inquiries by using the voice chatbots 244C and 245C.

The response generation module 244C is built in the integrated storage 260 based on the information provided by the reception content management module 250 and the conversation management module 243B according to the agent input command transmitted through the agent UI 230 . Searches the domain knowledge database and scenario database to generate appropriate response data. Here, the response data may be text data. Here, the domain knowledge base may store response manuals classified according to domains, and the scenario database may store response manuals classified according to report/incident reception types.

The voice utterance module 245C converts the response data (such as a guide message) generated by the response generation module 244C into voice and transmits it to the user terminal 100 of the complainant through the communication device 210 . Here, the speech speech module 245C may be, for example, a text-to-speech (TTS).

For response data, for example, in case of 112, by incident type (theft, assault, etc.), in case of 119, by type of disaster (fire, rescue, first aid, etc.), the conversation situation is recognized and an appropriate response guide message is delivered.

Information provided to the agent UI 230 through the customer response knowledge/information generation module 244B shown in FIG. and a guide message selected in the process of responding to civil servants through the function menu can be used.

When the artificial intelligence agent module 240C according to the fourth embodiment analyzes the request and intention of the civil servant using only the agent's voice, the artificial intelligence agent module 240C is the agent UI ( 230) is connected.

In addition, the artificial intelligence agent module 240C according to the fourth embodiment transmits the guidance voice artificially generated by the voice utterance module 245C to the user terminal 100 of the civil servant. The artificial intelligence agent module 240C communicates It is additionally connected to the equipment 210 .

6 is a block diagram illustrating an internal configuration of an artificial intelligence call center system according to a fifth embodiment of the present invention.

Referring to FIG. 6 , in the artificial intelligence call center system 200 according to the fifth embodiment, the artificial intelligence agent module 240D is configured to further include a duplicate report determination module 246D, which is different from the embodiments described above. have.

When an emergency situation occurs in the same space (region), a plurality of overlapping reports may be received in the call center system. This hinders the smooth receipt of other reports.

In the artificial intelligence call center system 200 according to the fifth embodiment, in order to minimize duplicate report reception, the artificial intelligence agent module 240D is configured to further include a duplicate report determination module 246D.

First, the reception contents management module 250 finally determines the contents of the report reception, as in the above-described embodiments, and automatically classifies the report/incident type.

The duplicate report determination module 246D included in the artificial intelligence agent module 240D according to the fifth embodiment compares the previous information collected by the conversation management module 243B with the current information, It is determined whether the contents of the complaint and the report of other complainants overlap. Here, the previous information collected by the conversation management module 243B includes the previous speech recognition-based text data, the previous intention analysis data, and the previous summary data, and the current information collected by the conversation management module 243B includes current speech recognition-based text data, current intention analysis data, and current summary data.

The duplicate report determination module 246D is configured to recognize at least one of the previous voice recognition-based text data included in the previous information, the previous intention analysis data, and the previous summary data and the current voice recognition included in the current information. By comparing at least one of the text data, the current intention analysis data, and the current summary data, it is determined whether the complainant's report and the other complainant's report overlap.

If it is a duplicate report, for example, "The content of ~~ (summary) reported by the complainant was received and processed as a report/incident type of ~~ at OO o'clock on O/O/O/OO, OO." Alternatively, following this comment, it is also possible to guide the situation, such as "~~ In a situation, the following actions are suggested. First, ~~~~ ??".

Guidance for action tips, etc. may be repeatedly reproduced through a function such as listening again provided by the communication device 210 . Such a guide message may be any one of a message stored in advance for a call consultation service, a response message previously provided by an agent in a report reception, and a message generated by the artificial intelligence agent module 240D. Here, the comment generated by the artificial intelligence agent module 240D may be generated by the response generating module 244C and the voice utterance module 245D serving as a voice chatbot.

The duplicate report determination module 246D may be implemented as a pre-trained artificial neural network model to determine whether or not to report a duplicate report. Here, the artificial neural network model may be a model trained in advance to calculate the similarity between the previous information and the current information collected by the conversation management module.

As an example, the duplicate report determination module 246D implemented as an artificial neural network model is a word or sentence included in the speech recognition-based text data previously collected by the conversation management module 243B and the currently collected speech recognition-based The degree of similarity between words or sentences included in text data is calculated, and if the calculated similarity is greater than or equal to a threshold, the complaints reported by the complainant corresponding to the currently collected speech recognition-based text data are previously collected speech recognition-based text data. It is discriminated as the content of the report that is duplicated with the report of the complainant who responds to the In this case, the duplicate report determination module receives the previous text data and the current text data corresponding to the voice recognition result directly from the voice recognition module 241B without going through the conversation management module 243B, and calculates their similarity. have. In addition, when the voice recognition module 241B performs voice recognition only for the agent's voice, the text data used for calculating the similarity in the duplicate report determination module 246D may be the voice recognition result of the agent's voice, not the complainant's voice. .

In the artificial intelligence call center system according to the first to fifth embodiments described above, in order to help the understanding of the description, the artificial intelligence agent module (240, 240A, 240B, 240C or 240D) and the reception content management module (250, 250A or 250B) has been described as separate, but they may be integrated into one module. For example, the reception content management module 250 , 250A or 250B may be embedded in the artificial intelligence agent module 240 , 240A, 240B, 240C or 240D.

In addition, in each embodiment, the components included in the artificial intelligence agent module 240, 240A, 240B, 240C or 240D are also divided into functional units to help the understanding of the description, and may be integrated into one module. is self-evident

In addition, although not clearly shown in FIGS. 1 to 6 for explaining each embodiment, the artificial intelligence agent module (240, 240A, 240B, 240C or 240D) and the reception content management module (250, 250A or 250B) are a processor and a memory It will be fully understood by those skilled in the art that it may be implemented as an artificial neural network algorithm executed by a computing device comprising the same.

7 is a flowchart illustrating a service provision method of an artificial intelligence call center system according to an embodiment of the present invention, and FIG. 8 is a flowchart illustrating a service provision method of an artificial intelligence call center system according to another embodiment of the present invention.

The service providing method of the artificial intelligence call center system shown in FIG. 7 is a service providing method based on a voice recognition result for a conversation voice including both the civil servant's voice and the counselor's voice, whereas the service providing method shown in FIG. The method differs in the service provision method based on the voice recognition result obtained by performing voice recognition only on the counselor's voice.

First, referring to FIG. 7 , in step 710 , in the voice recognition module 241 , voice recognition is performed on the conversation voice between the civil servant and the counselor input from the communication equipment 210 , and the voice recognition corresponding to the conversation voice is performed. The process of generating text data based on it is performed.

Next, in step 720, a process of extracting intention analysis data representing the intention of the civil petitioner from the speech recognition-based text data by the dialogue analysis module 242 using natural language processing technology is performed. Here, the extraction of the intention analysis data may be, for example, extracting the intention analysis data from the speech recognition-based text data using a conversation corpus.

Next, in step 730, the process of the reception content management module 250 automatically classifying the report type based on the speech recognition-based text data and the intention analysis data and delivering the report type classification information generated to the agent interface is carried out Here, for the automatic classification of the report type, an artificial neural network model trained in advance by using the training text data and the training intention analysis data as inputs may be used.

In step 730, the process of receiving the agent UI-based text data corresponding to the conversation voice between the civil servant and the agent directly written by the agent from the agent interface ( ), and the agent UI-based text data using the voice recognition-based text data A process of correcting the text data of may be further performed. Here, the correction of the agent UI-based text data is performed, for example, after calculating the similarity between the words included in the agent UI-based text data and the words included in the voice recognition-based text data, and the Among words included in the agent UI-based text data, a word having a similarity less than or equal to a threshold may be replaced with a corresponding word included in the voice recognition-based text data.

Next, in step 740 , the agent interface 230 searches the integrated storage 260 and delivers a corresponding manual mapped to the report type classification information to the agent. Here, the process of delivering the corresponding manual to the counselor may be a process of displaying the corresponding manual through the display screen provided on the counselor interface 230 .

Additionally, in the service providing method of the artificial intelligence call center system according to an embodiment, the reception content management module 240 uses the voice recognition-based text data to allow the agent to directly use the agent UI-based text through the agent interface. In the process of compensating data, the conversation analysis module 242 extracts the intention analysis data again based on the corrected agent UI-based text data, and the re-extracted intention analysis data is transferred to the reception content management module 250 In the process of delivering to and the reception content management module 250, the corrected report type classification information generated by reclassifying the report type based on the corrected agent UI-based text data and the re-extracted intention analysis data The method may further include transmitting the information to the agent interface 230 . In the system configuration of FIG. 1 for implementing the method according to this additional embodiment, the dialogue analysis module 242 and the dialogue management module 243 are separated, but the dialogue management module 243 is included in the dialogue analysis module 242 . may be integrated as much as possible. In this case, the service providing method does not need to be mentioned about the process performed by the conversation management module 243 .

Hereinafter, a service providing method based on a voice recognition result obtained by performing voice recognition on only the counselor's voice will be described with reference to FIG. 8 .

Referring to FIG. 8, first, in step 810, the voice recognition module 241A performs voice recognition on the agent's conversation voice input through the agent interface 230 to generate text data based on the agent's voice. The process is carried out.

Next, in step 820, the conversation analysis module 242A extracts intention analysis data representing the intention of the civil servant from the text data based on the counselor's voice by using natural language processing technology.

Next, in step 830, the reception content management module 250A transmits the report type classification information generated by classifying the report type based on the agent voice-based text data and the intention analysis data to the agent interface 230. The process is carried out.

Step 830 is a process of receiving, from the agent interface, text data based on the agent UI corresponding to the conversation voice between the complainant and the agent directly written by the agent through the agent interface, and the agent UI using the text data based on the agent's voice. It may further include a process of correcting the based text data. Here, the correction of the agent UI-based text data is performed after calculating the similarity between words included in the agent UI-based text data and the words included in the agent voice-based text data, and then, the agent UI-based text data Among words included in the data, a word having a similarity of less than or equal to a threshold may be replaced with a corresponding word included in the agent's voice-based text data. In this embodiment, although the replacement target unit is limited to a word, it is natural that the unit may be extended to a phrase or clause sentence unit.

Next, in step 840, the agent interface 230 searches the integrated storage 260 and delivers a corresponding manual mapped to the report type classification information to the agent.

In the service providing method according to another embodiment of FIG. 8 , additionally, after the reception content management module corrects the agent UI-based text data using the agent voice-based text data, the corrected agent UI-based text data The process of delivering text data to the dialog analysis module, the dialog analysis module extracts the intention analysis data again based on the corrected agent UI-based text data, and the re-extracted intention analysis data is transferred to the reception content management module The process of delivering to and the reception content management module transmits the corrected report type classification information generated by reclassifying the report type based on the corrected agent UI-based text data and the re-extracted intention analysis data to the agent interface It may further include a process of transferring to.

9 is a block diagram of an exemplary computing device to which the service providing method of the artificial intelligence call center system of the present invention can be applied.

Referring to FIG. 9 , computing device 500 may be used to perform various procedures such as those discussed herein.

Computing device 500 may be any of a variety of computing devices, such as desktop computers, notebook computers, server computers, portable computers, tablet computers, and the like.

Computing device 500 includes one or more processors (CPU, GPU) 402 , one or more memory device(s) 404 , one or more interface(s) 406 , one or more mass storage device(s) 408 . , one or more input/output (I/O) device(s) 410 , and a display device 430 , all connected to bus 412 .

Processor(s) 402 includes one or more processors or controllers that execute instructions stored in memory device(s) 404 and/or mass storage device(s) 408 . Processor(s) 402 may also include various types of computer-readable media, such as cache memory.

Here, the processor(s) 402 executes and operates the configurations 210, 220, 230, 240, 240A, 240B, 240C, 240D, 250, 250A, 250B, 260 shown in FIGS. can be the subject.

Memory device(s) 404 may include volatile memory (eg, random access memory (RAM) 414 ) and/or non-volatile memory (eg, read-only memory (ROM) 416 ). a variety of computer-readable media. Memory device(s) 404 may also include rewritable ROM, such as flash memory.

Mass storage device(s) 408 include a variety of computer-readable media, such as magnetic tape, magnetic disks, optical disks, solid-state memory (eg, flash memory), and the like. As shown in FIG. 4 , a particular mass storage device is a hard disk drive 424 . Various drives may also be included in mass storage device(s) 408 to enable reading from and/or writing to various computer readable media. Mass storage device(s) 408 include removable storage 426 and/or non-removable media.

Input/output (I/O) device(s) 410 include various devices that allow data and/or other information to be input to or retrieved from computing device 400 . Exemplary I/O device(s) 410 include a cursor control device, keyboard, keypad, microphone, monitor or other display device, speaker, printer, network interface card, modem, lens, CCD or other image capture device, etc. include

Display device 430 includes any type of device capable of displaying information to one or more users of computing device 400 . Examples of display devices 430 include monitors, display terminals, video projection devices, and the like.

A graphics-processing unit (GPU) 432 may be coupled to the processor(s) 402 and/or the display device 430 . The GPU may be operable to render computer-generated images and perform other graphics processing.

A GPU may include some or all of the functionality of a general purpose processor, such as processor(s) 402 . The GPU may also include additional functions specific to graphics processing. The GPU may include hard-coded and/or hard-wired graphics functions related to coordinate transformations, shading, texturing, rasterization, and other functions useful for rendering computer-generated images.

Interface(s) 406 include various interfaces that enable computing device 400 to interact with other systems, other devices, or other computing environments.

Exemplary interface(s) 406 include any number of different network interfaces 420 , such as interfaces to local area networks (LANs), wide area networks (WANs), wireless networks, and the Internet.

Other interface(s) include user interface 418 and peripheral device interface 422 .

Interface(s) 406 may also include one or more user interface elements 418 . Interface(s) 406 may also include one or more peripheral interfaces, such as interfaces for printers, pointing devices (mouse, trackpad, etc.), keyboards, and the like.

Bus 412 includes processor(s) 402 , memory device(s) 404 , interface(s) 406 , mass storage device(s) 408 , and I/O device(s) 410 . ) to communicate with each other, as well as other devices or components connected to the bus 412 .

Bus 412 represents one or more of several types of bus structures, such as a system bus, a PCI bus, an IEEE 1394 bus, a USB bus, and the like.

For purposes of illustration, although programs and other executable program components are shown as separate blocks herein, such programs and components may reside at various times in different storage components of computing device 400 , and the processor(s) ) 402 . Alternatively, the systems and procedures described herein may be implemented in hardware, or hardware, or a combination of hardware, software and/or firmware. For example, one or more application specific integrated circuits (ASICs) may be programmed to perform one or more systems and procedures described herein.

So far, the present invention has been looked at focusing on examples. Those of ordinary skill in the art to which the present invention pertains will understand that the present invention may be implemented in variously changed or modified forms without departing from the essential characteristics of the present invention. Therefore, the disclosed embodiments are to be considered in terms of illustrative rather than restrictive views. The scope of the present invention is indicated in the claims rather than the foregoing description, and all differences within the scope equivalent thereto should be construed as being included in the present invention.

In addition, in the above method, examples of the computer-readable medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical recording media such as CD-ROMs and DVDs, magneto-optical media such as floppy disks; and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like.

Examples of program instructions include not only machine code generated by a compiler, but also high-level language code that can be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate as one or more software modules to perform the operations of the present invention, and vice versa.

The protection scope of the present invention is not limited to the description and expression of the embodiments explicitly described above. In addition, it is added once again that the protection scope of the present invention cannot be limited due to obvious changes or substitutions in the technical field to which the present invention pertains.

Claims (16)

generating, by the voice recognition module, voice recognition-based text data corresponding to the conversation voice by performing voice recognition on the conversation voice between the civil servant and the counselor input from the communication equipment;
extracting, by the dialogue analysis module, intention analysis data indicating the intention of the civil petitioner from the speech recognition-based text data using natural language processing technology;
transmitting, by the reception content management module, the report type classification information generated by classifying the report type based on the speech recognition-based text data and the intention analysis data to an agent interface; and
transmitting, by the agent interface, a corresponding manual mapped to the report type classification information by searching the integrated repository to the agent
A method of providing a service based on an artificial intelligence call center system comprising a. In claim 1,
The step of extracting the intention analysis data,
The service providing method based on an artificial intelligence call center system, wherein the terminal extracts the intention analysis data from the speech recognition-based text data using a conversation corpus. In claim 1,
The steps of passing the reporting type classification information to the agent interface are:
The method for providing a service based on an artificial intelligence call center system is the step of classifying the report type using an artificial neural network model trained in advance using text data for training and intention analysis data for training as inputs. In claim 1,
The step of delivering the report type classification information to the agent interface comprises:
receiving, from the counselor interface, text data based on a counselor UI corresponding to a conversation voice between a civil petitioner and a counselor directly written by the counselor; and
correcting the agent UI-based text data using the voice recognition-based text data;
A method of providing a service based on an artificial intelligence call center system comprising a. In claim 4,
The step of correcting the text data based on the agent UI,
After calculating the similarity between the words included in the agent UI-based text data and the words included in the voice recognition-based text data, the similarity below a threshold among the words included in the agent UI-based text data is determined replacing the words with the corresponding words included in the speech recognition-based text data;
An artificial intelligence call center system-based service provision method. In claim 1,
correcting, by the reception content management module, the agent UI-based text data directly through the agent interface by the agent using the voice recognition-based text data;
re-extracting, by the dialog analysis module, the intention analysis data based on the corrected agent UI-based text data, and transmitting the re-extracted intention analysis data to the reception content management module; and
transmitting, by the reception content management module, the corrected report type classification information generated by reclassifying the report type based on the corrected agent UI-based text data and the re-extracted intention analysis data to the agent interface;
An artificial intelligence call center system-based service provision method further comprising a. performing, by the voice recognition module, voice recognition on the agent's conversational voice input through the agent interface, and generating text data based on the agent's voice;
extracting, by the dialogue analysis module, intention analysis data indicating the intention of the civil servant from the text data based on the counselor's voice using natural language processing technology;
transmitting, by the reception content management module, report type classification information generated by classifying a report type based on the agent's voice-based text data and the intention analysis data to an agent interface; and
transmitting, by the agent interface, a corresponding manual mapped to the report type classification information by searching the integrated repository to the agent
A method of providing a service based on an artificial intelligence call center system comprising a. In claim 7,
The step of delivering the report type classification information to the agent interface comprises:
receiving, from the counselor interface, text data based on a counselor UI corresponding to a conversation voice between a civil petitioner and a counselor directly written by the counselor through the counselor interface; and
correcting the agent UI-based text data using the agent voice-based text data;
A service provision method based on an artificial intelligence call center system further comprising a. In claim 8,
The step of correcting the text data based on the agent UI,
After calculating the similarity between the words included in the agent UI-based text data and the agent voice-based text data, the similarity below a threshold among the words included in the agent UI-based text data is calculated replacing the words with the corresponding words included in the text data based on the agent's voice
An artificial intelligence call center system-based service provision method. In claim 7,
after the reception content management module corrects the agent UI-based text data using the agent voice-based text data, transmitting the corrected agent UI-based text data to the conversation analysis module;
re-extracting, by the dialog analysis module, the intention analysis data based on the corrected agent UI-based text data, and transmitting the re-extracted intention analysis data to the reception content management module; and
transmitting, by the reception content management module, the corrected report type classification information generated by reclassifying the report type based on the corrected agent UI-based text data and the re-extracted intention analysis data to the agent interface;
A service provision method based on an artificial intelligence call center system further comprising a. an agent interface that communicates with the user terminal of the complainant through communication equipment (hereinafter referred to as 'counselor UI');
a voice recognition module for generating voice recognition-based text data by performing voice recognition on a conversation voice between a civil servant and a counselor input from the communication device or an agent voice input from the counselor UI;
a dialogue analysis module for extracting intention analysis data representing the intention of a civil petitioner from the speech recognition-based text data using natural language processing technology; and
and a reception content management module that classifies the report type based on the speech recognition-based text data and the intention analysis data and transmits the report type classification information obtained to the agent interface,
The agent UI is
An artificial intelligence call center system that searches the integrated storage and delivers a response manual mapped to the report type classification information to the agent. In claim 11,
The dialogue analysis module,
An artificial intelligence call center system that extracts the intention analysis data from the speech recognition-based text data using a conversation corpus tagged with an intention value specialized for the call center system. In claim 11,
The reception content management module,
The agent receives, from the agent UI, the agent UI-based text data written by summarizing the conversation voice between the complainant and the agent, and corrects the agent UI-based text data using the voice recognition-based text data, the correction An artificial intelligence call center system that stores the agent UI-based text data in the counseling conversation database built in the integrated storage. In claim 13,
The reception content management module,
An artificial intelligence call center system for correcting the agent UI-based text data using the voice recognition-based text data generated by performing voice recognition on the conversation voice between the civil servant and the agent. In claim 13,
The reception content management module,
An artificial intelligence call center system for correcting the agent UI-based text data using the voice recognition-based text data generated by performing voice recognition on the agent's voice. In claim 13,
The dialogue analysis module,
extracting corrected intention analysis data from the corrected agent UI-based text data transmitted from the reception contents management module, and transferring the corrected intention analysis data to the reception contents management module;
The reception content management module,
The artificial intelligence call center system to reclassify the report type based on the corrected agent UI-based text data and the corrected intention analysis data.

Images