Manual customer service auxiliary system and method based on large model

Abstract

The invention relates to the technical field of data processing and artificial intelligence, in particular to an artificial customer service auxiliary system and method based on a large model. The system comprises a user side, a customer service side, a data processing module, a visual window, an intelligent extraction module, an intelligent quality inspection module, an intelligent assistant module and an appeal analysis module. The intelligent extraction module is used for converting voice information into text information, core content is generated to be displayed on a visual window, the intelligent assistant module is used for providing real-time assistance for customer service, the service efficiency and accuracy are improved, the appeal analysis module is used for deeply analyzing the appeal of a user, a basis is provided for accurate service, and in the call process, the user experience is improved. The call quality and the service quality of the voice information sent by the customer service terminal are monitored in real time by using the intelligent quality inspection module, and the prompt information is sent to the customer service terminal, so that the customer service terminal can correct in time, and the response speed, the reply accuracy and the service experience of the manual customer service terminal can be improved by adopting the technical scheme.

Description

Technical Field

[0001] The present invention relates to the field of data processing and artificial intelligence technology, and in particular to an artificial customer service assistance system and method based on a large model. Background Art

[0002] In daily work, manual customer service bears the important responsibilities of answering public inquiries, providing statistical data and analysis reports.

[0003] However, with the continuous increase in the number of consultations and the increasing complexity of consultation content, the traditional manual customer service model faces many challenges, such as slow response speed, low answer accuracy, and poor service experience. Summary of the invention

[0004] The purpose of the present invention is to provide a manual customer service assistance system and method based on a large model to solve the problems that the traditional manual customer service model faces many challenges as the number of consultations continues to increase and the content of consultations becomes increasingly complex, such as slow response speed, low answer accuracy, and poor service experience.

[0005] To achieve the above-mentioned purpose, the present invention provides an artificial customer service assistance system based on a large model, the artificial customer service assistance system based on a large model comprises a user end, a customer service end, a data processing module, a visualization window, an intelligent extraction module, an intelligent quality inspection module, an intelligent assistant module and a demand analysis module, the user end is communicatively connected with the customer service end, and the customer service end is provided with the data processing module, the intelligent extraction module, the intelligent quality inspection module, the intelligent assistant module, the demand analysis module and the visualization window;

[0006] The data processing module is used to pre-process the voice information transmitted by the user to the customer service end. The intelligent extraction module is used to convert the voice information into text information, and after semantic understanding and keyword extraction based on the large model, generate core content for display on the visualization window. The intelligent quality inspection module is used to monitor the call quality and service quality of the voice information sent by the customer service end in real time, and send prompt information to the customer service through the visualization window. The intelligent assistant module is used to provide real-time assistance to the customer service through the visualization window to improve service efficiency and accuracy. The demand analysis module is used to deeply analyze the user's demands, provide a basis for accurate service, and display it through the visualization window.

[0007] Among them, the visualization window is provided with a summary text icon, a service quality reminder icon, an intelligent auxiliary icon and a core demand icon. The summary text icon is associated with the core content extracted by the intelligent extraction module, the service quality reminder icon is associated with the prompt information sent by the intelligent quality inspection module, the intelligent auxiliary icon is associated with the auxiliary content provided by the intelligent assistant module, and the core demand icon is associated with the user's demands analyzed by the demand analysis module.

[0008] Among them, the intelligent extraction module includes a speech recognition submodule, a semantic understanding submodule, a keyword extraction submodule and an intelligent summary generation submodule. The speech recognition submodule is used to convert the voice call content sent by the user terminal into text data. The semantic understanding submodule uses a pre-trained large model to encode the text, capture contextual information, and achieve deep semantic understanding. The keyword extraction submodule is based on the TF-IDF algorithm and combines the semantic understanding ability of the large model to extract keywords in the text. The intelligent summary generation submodule is based on the Transformer neural network model to automatically generate a core summary of the call content and display it on the visualization window.

[0009] Among them, the intelligent quality inspection module includes real-time quality inspection and early warning submodule, full quality inspection submodule, quality inspection result visualization submodule, service problem automatic classification submodule, quality inspection model optimization submodule, quality inspection case library construction submodule and cross-channel quality inspection integration submodule. The real-time quality inspection and early warning submodule is used to monitor the customer service's speaking speed, silent time and language use during the call, and issue early warning prompts in time through the visualization window. The full quality inspection submodule is used to perform 100% quality inspection on all calls, deeply explore business problems, user needs and potential risks in the call content, and the quality inspection result visualization submodule is used to display the quality inspection results. The results are presented in the form of charts and reports, which is convenient for managers to quickly understand the service quality status. The automatic classification submodule of service problems is used to automatically classify and generate work orders according to the types of problems found in quality inspection, and transfer them to relevant personnel for processing. The quality inspection model optimization submodule is used to continuously optimize the quality inspection model and rules according to business development and changes in service standards, and improve the accuracy of quality inspection. The quality inspection case library construction submodule is used to establish a quality inspection case library, organize and archive typical quality inspection cases for customer service and management personnel to learn and refer to. The cross-channel quality inspection integration submodule is used for quality inspection integration of multiple channels to ensure the consistency and coherence of service quality.

[0010] Among them, the intelligent assistant module includes a dynamic knowledge push sub-module and a knowledge base update sub-module. The dynamic knowledge push sub-module is used to analyze the conversation content in real time and automatically push relevant knowledge points and solutions to the visualization window. The knowledge base update sub-module is used to monitor changes in policies and regulations in real time and automatically capture, organize and update the knowledge base content.

[0011] Among them, the demand analysis module includes a real-time demand insight submodule, a demand labeling submodule, a hot demand monitoring submodule, a personalized demand analysis submodule, a business process association analysis submodule and a demand data mining submodule. The real-time demand insight submodule is used to quickly identify the user's core demands during a call. The demand labeling submodule is used to automatically classify and label user demands to facilitate subsequent statistical analysis. The hot demand monitoring submodule is used to track the hot trends of user demands in real time and issue early warnings in time. The personalized demand analysis submodule is used to combine user basic information and historical interaction records to perform personalized demand analysis. The business process association analysis submodule is used to analyze the relationship between user demands and existing business processes to find optimization space. The demand data mining submodule is used to use big data analysis technology to explore the potential laws and trends of user demands.

[0012] Among them, the artificial customer service assistance system based on the big model also includes an intelligent summary module, which is used to automatically generate summary content after the call ends, helping customer service to quickly organize and review the key points of the call.

[0013] The present invention also provides a method for assisting human customer service based on a large model, which is applied to the human customer service assisting system based on a large model as described above, and comprises the following steps:

[0014] When the user is talking to the customer service through the user terminal, the voice information is pre-processed by the data processing module;

[0015] The intelligent extraction module is used to convert voice information into text information, and after semantic understanding and keyword extraction based on the big model, the core content is generated to be displayed on the visualization window, so that the customer service can understand the important content in the user's voice information;

[0016] At the same time, the intelligent assistant module is used to provide real-time assistance to customer service to improve service efficiency and accuracy;

[0017] And use the demand analysis module to deeply analyze the user's demands to provide a basis for accurate service;

[0018] During the call, the intelligent quality inspection module is used to monitor the call quality and service quality of the voice information sent by the customer service end in real time, and a prompt message is sent to the customer service through the visualization window to facilitate the customer service to make corrections in time.

[0019] The present invention discloses an artificial customer service assistance system and method based on a large model, comprising a user end, a customer service end, a data processing module, a visualization window, an intelligent extraction module, an intelligent quality inspection module, an intelligent assistant module and a demand analysis module. In the process of a user talking to a customer service through the user end, the data processing module is used to preprocess the voice information, the intelligent extraction module is used to convert the voice information into text information, and after semantic understanding and keyword extraction based on the large model, the core content is generated to be displayed on the visualization window, so that the customer service can understand the important content in the user's voice information. At the same time, the intelligent assistant module is used to provide real-time assistance to the customer service to improve service efficiency and accuracy, and the demand analysis module is used to deeply analyze the user's demands to provide a basis for accurate service. During the call, the intelligent quality inspection module is used to monitor the call quality and service quality of the voice information sent by the customer service end in real time, and prompt information is sent to the customer service through the visualization window, so that the customer service can make corrections in time. The technical solution can improve the response speed, reply accuracy and service experience of artificial customer service. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0021] Figure 1 It is a principle block diagram of an artificial customer service assistance system based on a large model provided by the present invention.

[0022] Figure 2 It is a connection diagram of the intelligent extraction module, the intelligent assistant module and the visualization window provided by the present invention.

[0023] Figure 3 It is a connection diagram of the intelligent quality inspection module, the demand analysis module and the visualization window provided by the present invention.

[0024] Figure 4 It is a step flow chart of an artificial customer service assistance method based on a large model provided by the present invention.

[0025] 101-user end, 102-client end, 103-data processing module, 104-visualization window, 105-intelligent extraction module, 106-intelligent quality inspection module, 107-intelligent assistant module, 108-demand analysis module, 109-summary text icon, 110-service quality reminder icon, 111-intelligent auxiliary icon, 112-core demand icon, 113-speech recognition submodule, 114-semantic understanding submodule, 115-keyword extraction submodule, 116-intelligent summary generation submodule, 117-real-time quality inspection and warning submodule, 118-full quality inspection submodule Module, 119-quality inspection result visualization submodule, 120-service problem automatic classification submodule, 121-quality inspection model optimization submodule, 122-quality inspection case library construction submodule, 123-cross-channel quality inspection integration submodule, 124-dynamic knowledge push submodule, 125-knowledge base update submodule, 126-real-time demand insight submodule, 127-demand labeling submodule, 128-hot demand monitoring submodule, 129-personalized demand analysis submodule, 130-business process association analysis submodule, 131-demand data mining submodule, 132-intelligent summary module. DETAILED DESCRIPTION

[0026] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present invention, and should not be construed as limiting the present invention.

[0027] See also Figures 1 to 3 The present invention provides an artificial customer service assistance system based on a large model, the artificial customer service assistance system based on a large model comprises a user terminal 101, a customer service terminal 102, a data processing module 103, a visualization window 104, an intelligent extraction module 105, an intelligent quality inspection module 106, an intelligent assistant module 107 and a demand analysis module 108, the user terminal 101 is connected to the customer service terminal 102 for communication, and the customer service terminal 102 is provided with the data processing module 103, the intelligent extraction module 105, the intelligent quality inspection module 106, the intelligent assistant module 107, the demand analysis module 108 and the visualization window 104;

[0028] The data processing module 103 is used to pre-process the voice information transmitted by the user to the customer service end 102. The intelligent extraction module 105 is used to convert the voice information into text information, and after semantic understanding and keyword extraction based on the large model, generate core content for display on the visualization window 104. The intelligent quality inspection module 106 is used to monitor the call quality and service quality of the voice information sent by the customer service end 102 in real time, and send prompt information to the customer service through the visualization window 104. The intelligent assistant module 107 is used to provide real-time assistance to the customer service through the visualization window 104 to improve service efficiency and accuracy. The demand analysis module 108 is used to deeply analyze the user's demands, provide a basis for accurate service, and display it through the visualization window 104.

[0029] In this embodiment, when a user is talking to customer service through the user terminal 101, the data processing module 103 is used to pre-process the voice information, the intelligent extraction module 105 is used to convert the voice information into text information, and after semantic understanding and keyword extraction based on the large model, the core content is generated to be displayed on the visualization window 104, so that the customer service can understand the important content in the user's voice information. At the same time, the intelligent assistant module 107 is used to provide real-time assistance to the customer service to improve service efficiency and accuracy, and the demand analysis module 108 is used to deeply analyze the user's demands to provide a basis for accurate service. During the call, the intelligent quality inspection module 106 is used to monitor the call quality and service quality of the voice information sent by the customer service terminal 102 in real time, and a prompt message is sent to the customer service through the visualization window 104 to facilitate the customer service to make corrections in time. The use of this technical solution can improve the response speed, reply accuracy and service experience of manual customer service.

[0030] Among them, the visualization window 104 is provided with a summary text icon 109, a service quality reminder icon 110, an intelligent auxiliary icon 111 and a core demand icon 112. The summary text icon 109 is associated with the core content extracted by the intelligent extraction module 105, the service quality reminder icon 110 is associated with the prompt information sent by the intelligent quality inspection module 106, the intelligent auxiliary icon 111 is associated with the auxiliary content provided by the intelligent assistant module 107, and the core demand icon 112 is associated with the user's demand analyzed by the demand analysis module 108.

[0031] Among them, the intelligent extraction module 105 includes a speech recognition submodule 113, a semantic understanding submodule 114, a keyword extraction submodule 115 and an intelligent summary generation submodule 116. The speech recognition submodule 113 is used to convert the voice call content sent by the user terminal 101 into text data. The semantic understanding submodule 114 uses a pre-trained large model to encode the text, capture contextual information, and achieve deep semantic understanding. The keyword extraction submodule 115 is based on the TF-IDF algorithm and combines the semantic understanding ability of the large model to extract keywords from the text. The intelligent summary generation submodule 116 is based on the Transformer neural network model to automatically generate a core summary of the call content and display it on the visualization window 104.

[0032] Among them, the intelligent quality inspection module 106 includes a real-time quality inspection and early warning submodule 117, a full quality inspection submodule 118, a quality inspection result visualization submodule 119, a service problem automatic classification submodule 120, a quality inspection model optimization submodule 121, a quality inspection case library construction submodule 122 and a cross-channel quality inspection integration submodule 123. The real-time quality inspection and early warning submodule 117 is used to monitor the customer service's speaking speed, silent time and language use during the call, and issue early warning prompts in time through the visualization window 104. The full quality inspection submodule 118 is used for 100% quality inspection of all calls, and deeply explores the business problems, user needs and potential risks in the call content. The quality inspection result visualization Submodule 119 is used to present the quality inspection results in the form of charts and reports, so that management personnel can quickly understand the service quality status. The automatic classification submodule 120 of service problems is used to automatically classify and generate work orders according to the types of problems found in the quality inspection, and transfer them to relevant personnel for processing. The quality inspection model optimization submodule 121 is used to continuously optimize the quality inspection model and rules to improve the accuracy of quality inspection according to business development and changes in service standards. The quality inspection case library construction submodule 122 is used to establish a quality inspection case library, organize and archive typical quality inspection cases for customer service and management personnel to learn and refer to. The cross-channel quality inspection integration submodule 123 is used for quality inspection integration of multiple channels to ensure the consistency and coherence of service quality.

[0033] Among them, the intelligent assistant module 107 includes a dynamic knowledge push sub-module 124 and a knowledge base update sub-module 125. The dynamic knowledge push sub-module 124 is used to analyze the conversation content in real time and automatically push relevant knowledge points and solutions to the visualization window 104. The knowledge base update sub-module 125 is used to monitor changes in policies and regulations in real time and automatically capture, organize and update the knowledge base content.

[0034] Among them, the demand analysis module 108 includes a real-time demand insight submodule 126, a demand labeling submodule 127, a hot demand monitoring submodule 128, a personalized demand analysis submodule 129, a business process association analysis submodule 130 and a demand data mining submodule 131. The real-time demand insight submodule 126 is used to quickly identify the user's core demands during a call, the demand labeling submodule 127 is used to automatically classify and label user demands for subsequent statistical analysis, the hot demand monitoring submodule 128 is used to track the hot trends of user demands in real time and issue early warnings in time, the personalized demand analysis submodule 129 is used to combine user basic information and historical interaction records to perform personalized demand analysis, the business process association analysis submodule 130 is used to analyze the relationship between user demands and existing business processes to find optimization space, and the demand data mining submodule 131 is used to use big data analysis technology to explore the potential laws and trends of user demands.

[0035] The artificial customer service assistance system based on the large model further includes an intelligent summary module 132, and the intelligent summary module 132 is used to automatically generate summary content after the call is over, so as to help customer service staff to quickly organize and review the key points of the call.

[0036] See also Figure 4 The present invention also provides a method for assisting human customer service based on a large model, which is applied to the human customer service assisting system based on a large model as described above, and comprises the following steps:

[0037] S1: When a user is talking to a customer service representative via the user terminal 101, the data processing module 103 is used to pre-process the voice information;

[0038] S2: The voice information is converted into text information by using the intelligent extraction module 105, and after semantic understanding and keyword extraction based on the big model, the core content is generated to be displayed on the visualization window 104, so that the customer service can understand the important content in the user's voice information;

[0039] S3: At the same time, the intelligent assistant module 107 is used to provide real-time assistance to customer service to improve service efficiency and accuracy;

[0040] S4: and using the demand analysis module 108 to deeply analyze the user's demands to provide a basis for accurate service;

[0041] S5: During the call, the intelligent quality inspection module 106 is used to monitor the call quality and service quality of the voice information sent by the customer service end 102 in real time, and a prompt message is sent to the customer service through the visualization window 104 to facilitate the customer service to make corrections in time.

[0042] In this embodiment, when a user is talking to customer service through the user terminal 101, the data processing module 103 is used to pre-process the voice information, the intelligent extraction module 105 is used to convert the voice information into text information, and after semantic understanding and keyword extraction based on the large model, the core content is generated to be displayed on the visualization window 104, so that the customer service can understand the important content in the user's voice information. At the same time, the intelligent assistant module 107 is used to provide real-time assistance to the customer service to improve service efficiency and accuracy, and the demand analysis module 108 is used to deeply analyze the user's demands to provide a basis for accurate service. During the call, the intelligent quality inspection module 106 is used to monitor the call quality and service quality of the voice information sent by the customer service terminal 102 in real time, and a prompt message is sent to the customer service through the visualization window 104 to facilitate the customer service to make corrections in time. The use of this technical solution can improve the response speed, reply accuracy and service experience of manual customer service.

[0043] What is disclosed above is only a preferred embodiment of the present invention, and it certainly cannot be used to limit the scope of rights of the present invention. Ordinary technicians in this field can understand that all or part of the processes of the above embodiment and equivalent changes made according to the claims of the present invention still fall within the scope of the invention.

Claims

1. A manual customer service assistance system based on a large model, characterized in that: It includes a user end, a customer service end, a data processing module, a visualization window, an intelligent extraction module, an intelligent quality inspection module, an intelligent assistant module and a demand analysis module. The user end is connected to the customer service end in communication, and the customer service end is provided with the data processing module, the intelligent extraction module, the intelligent quality inspection module, the intelligent assistant module, the demand analysis module and the visualization window; The data processing module is used to pre-process the voice information transmitted by the user to the customer service end. The intelligent extraction module is used to convert the voice information into text information, and after semantic understanding and keyword extraction based on the large model, generate core content for display on the visualization window. The intelligent quality inspection module is used to monitor the call quality and service quality of the voice information sent by the customer service end in real time, and send prompt information to the customer service through the visualization window. The intelligent assistant module is used to provide real-time assistance to the customer service through the visualization window to improve service efficiency and accuracy. The demand analysis module is used to deeply analyze the user's demands, provide a basis for accurate service, and display it through the visualization window.

2. The artificial customer service assistance system based on a large model as claimed in claim 1, characterized in that: The visualization window is provided with a summary text icon, a service quality reminder icon, an intelligent assistance icon and a core demand icon. The summary text icon is associated with the core content extracted by the intelligent extraction module, the service quality reminder icon is associated with the prompt information sent by the intelligent quality inspection module, the intelligent assistance icon is associated with the auxiliary content provided by the intelligent assistant module, and the core demand icon is associated with the user's demand analyzed by the demand analysis module.

3. The artificial customer service assistance system based on a large model as claimed in claim 2, characterized in that: The intelligent extraction module includes a speech recognition submodule, a semantic understanding submodule, a keyword extraction submodule and an intelligent summary generation submodule. The speech recognition submodule is used to convert the voice call content sent by the user terminal into text data. The semantic understanding submodule uses a pre-trained large model to encode the text, capture context information, and achieve deep semantic understanding. The keyword extraction submodule is based on the TF-IDF algorithm and combines the semantic understanding ability of the large model to extract keywords in the text. The intelligent summary generation submodule is based on the Transformer neural network model to automatically generate a core summary of the call content and display it on the visualization window.

4. The artificial customer service assistance system based on a large model as claimed in claim 3, characterized in that: The intelligent quality inspection module includes a real-time quality inspection and early warning submodule, a full quality inspection submodule, a quality inspection result visualization submodule, a service problem automatic classification submodule, a quality inspection model optimization submodule, a quality inspection case library construction submodule and a cross-channel quality inspection integration submodule. The real-time quality inspection and early warning submodule is used to monitor the customer service's speaking speed, silent time and language use during the call, and issue early warning prompts in time through the visualization window. The full quality inspection submodule is used to perform 100% quality inspection on all calls, and deeply explore the business problems, user needs and potential risks in the call content. The quality inspection result visualization submodule is used to display the quality inspection results. It is presented in the form of charts and reports, which makes it easy for managers to quickly understand the service quality status. The automatic classification submodule of service problems is used to automatically classify and generate work orders according to the types of problems found in quality inspection, and transfer them to relevant personnel for processing. The quality inspection model optimization submodule is used to continuously optimize the quality inspection model and rules according to business development and changes in service standards, and improve the accuracy of quality inspection. The quality inspection case library construction submodule is used to establish a quality inspection case library, organize and archive typical quality inspection cases for customer service and management personnel to learn and refer to. The cross-channel quality inspection integration submodule is used for quality inspection integration of multiple channels to ensure the consistency and coherence of service quality.

5. The artificial customer service assistance system based on a large model as claimed in claim 4, characterized in that: The intelligent assistant module includes a dynamic knowledge push submodule and a knowledge base update submodule. The dynamic knowledge push submodule is used to analyze the conversation content in real time and automatically push relevant knowledge points and solutions to the visualization window. The knowledge base update submodule is used to monitor changes in policies and regulations in real time and automatically capture, organize and update the knowledge base content.

6. The artificial customer service assistance system based on a large model as claimed in claim 5, characterized in that: The demand analysis module includes a real-time demand insight submodule, a demand labeling submodule, a hot demand monitoring submodule, a personalized demand analysis submodule, a business process association analysis submodule and a demand data mining submodule. The real-time demand insight submodule is used to quickly identify the user's core demands during a call. The demand labeling submodule is used to automatically classify and label user demands to facilitate subsequent statistical analysis. The hot demand monitoring submodule is used to track the hot trends of user demands in real time and issue early warnings in time. The personalized demand analysis submodule is used to combine user basic information and historical interaction records to perform personalized demand analysis. The business process association analysis submodule is used to analyze the relationship between user demands and existing business processes to find optimization space. The demand data mining submodule is used to use big data analysis technology to explore the potential laws and trends of user demands.

7. The artificial customer service assistance system based on a large model as claimed in claim 6, characterized in that: The artificial customer service assistance system based on the large model also includes an intelligent summary module, which is used to automatically generate summary content after the call is over, helping customer service to quickly organize and review the key points of the call.

8. A method for assisting human customer service based on a large model, applied to a system for assisting human customer service based on a large model as claimed in claim 1, characterized in that: The steps include: When the user is talking to the customer service through the user terminal, the voice information is pre-processed by the data processing module; The intelligent extraction module is used to convert the voice information into text information, and after semantic understanding and keyword extraction based on the big model, the core content is generated to be displayed on the visualization window, so that the customer service can understand the important content in the user's voice information; At the same time, the intelligent assistant module is used to provide real-time assistance to customer service to improve service efficiency and accuracy; And use the demand analysis module to deeply analyze the user's demands to provide a basis for accurate service; During the call, the intelligent quality inspection module is used to monitor the call quality and service quality of the voice information sent by the customer service end in real time, and a prompt message is sent to the customer service through the visualization window to facilitate the customer service to make corrections in time.

Images