Customer service system

Abstract

The invention discloses a customer service system, and relates to the technical field of customer service systems, the customer service system comprises a voice robot module, the voice robot module is used for responding to a customer call, and a voice recognition unit and a natural language processing unit are integrated in the voice robot module; and the geographic information system module is used for acquiring the geographic position information of the incoming call client and storing a contact information database of investment attraction managers in each region. According to the invention, an end-to-end deep learning model is adopted, speech recognition and intention recognition are fused in one model for joint optimization, the problem of error accumulation and amplification in a traditional serial model is avoided, and the recognition accuracy is improved through an attention mechanism and context semantic understanding. The intention recognition accuracy and robustness in complex scenes such as a noisy environment and various accents are remarkably improved, and the judgment reliability of the incoming call intention of the customer is ensured.

Description

A customer service system Technical Field

[0001] This application relates to the technical field of customer service systems, and in particular to a customer service system. Background Technology

[0002] In traditional franchise customer service systems, handling franchise inquiries typically relies on human agents for initial answering and assessment. This process has several significant problems: First, human agents need to communicate with the customer to determine whether the call is for franchise inquiries or other customer service issues, which consumes valuable initial response time. Once a franchise inquiry is confirmed, the agent needs to verbally ask for and record the customer's basic information, and then manually transfer the call to the appropriate regional franchise manager based on memory or by checking the contact list. This traditional model has the following inherent flaws: intent identification, information recording, and call transfer all depend on manual operation, making the process cumbersome and causing customers to wait... The process suffers from several drawbacks: excessively long waiting times and poor initial user experience; manual recording of information is prone to omissions and errors, and the information is often presented in isolated forms, making it difficult to synchronize with the sales manager in a timely and accurate manner. Sales managers are often completely unaware of the customer before answering the call, hindering effective preparation; regional matching relies on customer service staff's memory and manual searching, which is prone to errors and may result in customers being transferred to the wrong service; for non-franchise inquiries, there is a lack of a scientific seat allocation mechanism, which can lead to some customer service staff being overworked while others are idle, resulting in an uneven distribution of resources; the entire process cannot handle voice recognition in complex environments and lacks a mechanism for automatic information verification and closed-loop management. Summary of the Invention

[0003] To address the aforementioned problems, this application provides a customer service system. The customer service system provided by this application adopts the following technical solution: A customer service system comprising: a voice robot module, which responds to customer calls and integrates a voice recognition unit and a natural language processing unit; a geographic information system module, which acquires the geographical location information of the caller and stores a database of contact information for regional investment managers; an intelligent routing and transfer control module, which is communicatively connected to the voice robot module and the geographic information system module; and an information collaborative processing platform, which is communicatively connected to the voice robot module and the intelligent routing and transfer control module, and is used to automatically execute information recording and distribution processes during or after call transfer. As a preferred technical solution of this application, the voice robot module converts the customer's voice into text using the integrated voice recognition unit and analyzes the text using the natural language processing unit to identify the customer's caller intent and distinguish between franchise-related and non-franchise-related calls. As a preferred technical solution of this application, the intelligent routing and transfer control module is used to: when the voice robot module identifies a call as a franchise inquiry, trigger the geographic information system module to match the customer's geographical location and generate a first transfer instruction based on the matching result, directly transferring the call to the mobile phone of the corresponding regional sales manager; when the voice robot module identifies a call as a non-franchise inquiry, generate a second transfer instruction based on a preset intelligent routing strategy, transferring the call to the corresponding human customer service agent. As a preferred technical solution of this application, the information collaborative processing platform includes an information capture and extraction unit, an electronic form generation unit, and an information distribution and synchronization unit. The information capture and extraction unit is used to capture the call content in real time or near real time through the voice recognition unit of the voice robot module during the call and automatically extract key information such as customer name, contact information, and intended region based on a preset key information model. The electronic form generation unit is used to automatically fill the extracted key information into a standardized franchise inquiry electronic form. The information distribution and synchronization unit is used to push the generated electronic form to a shared database through a secure data interface and simultaneously send notifications to the matched sales manager and relevant collaborative departments through at least one of email, SMS, or enterprise applications. As a preferred technical solution of this application, the output end of the information collaborative processing platform is connected to a business development manager mobile terminal application. The business development manager mobile terminal application is installed on the business development manager's mobile phone. The business development manager mobile terminal application is used to receive and display the electronic forms and notification reminders in real time before the call is transferred or during the call, so that the business development manager can obtain key customer information before answering the call.As a preferred technical solution of this application, the intelligent routing and transfer control module integrates an intelligent routing strategy. Specifically, the intelligent routing strategy performs multi-dimensional calculations based on the skill sets of human customer service agents, their current idle status, and the efficiency of handling similar issues in the past, to achieve load balancing for non-franchise inquiry calls. As a preferred technical solution of this application, the information collaboration processing platform internally includes an information verification and completion mechanism. Specifically, after a call ends, the system automatically sends a confirmation SMS or email containing key information from the electronic form to the customer; if the customer's reply is incorrect, the system triggers an information revision task and notifies the relevant sales manager, achieving dynamic updating and calibration of the recorded information. As a preferred technical solution of this application, the voice robot module adopts an end-to-end deep learning model. This deep learning model integrates speech recognition and intent recognition into a unified model for joint optimization, used to improve the accuracy of intent recognition in complex call environments. As a preferred technical solution of this application, a method for handling franchise inquiries in a customer service system includes the following steps: a voice robot module answers the call and identifies the caller's intent through voice recognition and natural language processing technology; if it is a franchise inquiry, the geographic information system module is invoked to obtain the customer's location and match a franchise manager, while simultaneously initiating the information recording process of the information collaboration processing platform; the intelligent routing and transfer control module transfers the call to the matched franchise manager, and the information collaboration processing platform simultaneously pushes the generated inquiry information form to the franchise manager; if it is not a franchise inquiry, the call is transferred to the optimal human customer service agent according to the intelligent routing strategy.In summary, this application includes at least one of the following beneficial technical effects of the customer service system: 1. This application adopts an end-to-end deep learning model, integrating speech recognition and intent recognition into a single model for joint optimization. This avoids the problem of error accumulation and amplification in traditional serial models. Through attention mechanisms and contextual semantic understanding, it significantly improves the accuracy and robustness of intent recognition in complex scenarios such as noisy environments and diverse accents, ensuring the reliability of customer call intent judgment; 2. This application achieves three-level precise routing of "intent-location-agent". For franchise inquiries, it uses a high-precision geographic information system to achieve precise matching and direct transfer of "customer-region-recruitment manager", greatly shortening the transfer path. For non-franchise inquiries, it uses intelligent calculation based on multiple dimensions such as skills, availability, and historical efficiency. 1. This application employs a load balancing method to ensure optimal utilization of customer service resources and efficient resolution of customer issues; 2. Through an information collaboration processing platform, this application automatically captures and extracts key customer information during calls, generating standardized electronic forms, eliminating inefficient manual recording. More importantly, the platform can push form information to the sales manager's mobile terminal in real time while the call is transferred, achieving "zero-delay" information synchronization. This ensures that the sales manager has grasped the customer details before answering the call and provides an accurate data foundation for subsequent follow-up; 3. This application's seamless automation of the entire process greatly shortens customer waiting time and optimizes enterprise resource utilization through intelligent load balancing of agents, thereby improving customer satisfaction while also enhancing internal operational efficiency and management level. Figure 1 is a diagram of the customer service system architecture of this application; Figure 2 is a flowchart of the franchise consultation processing method of this application. The specific implementation method is further described in detail below with reference to Figures 1-2. Referring to Figures 1-2, a customer service system includes: a voice robot module, which is used to respond to customer calls. The voice robot module integrates a speech recognition unit and a natural language processing unit. The voice robot module converts the customer's speech into text through the integrated speech recognition unit and analyzes the text through the natural language processing unit to identify the customer's caller intent and distinguish between franchise-related and non-franchise-related calls. The voice robot module adopts an end-to-end deep learning model, which integrates speech recognition and intent recognition into a unified model for joint optimization to improve the accuracy of intent recognition in complex call environments. The voice robot module employs an end-to-end deep learning architecture, integrating speech recognition and intent recognition tasks into a unified model for joint optimization. This module uses a built-in high-sensitivity speech recognition unit to capture customer voice signals in real time and convert them into text information. Subsequently, the natural language processing unit performs multi-level analysis of the text based on deep semantic understanding technology, accurately distinguishing between franchise-related and non-franchise-related calls. Unlike traditional sequential processing flows, this module achieves direct mapping from acoustic features to intent labels through an end-to-end model, effectively avoiding the accumulation and amplification of speech recognition errors at the intent recognition stage. Especially in complex call scenarios, such as diverse customer accents, environmental noise interference, or intermittent speech, the model strengthens focus on key speech segments through an attention mechanism and uses contextual semantic association for intent correction, significantly improving recognition robustness in noisy environments. This module can also dynamically adjust according to changes in intent during the conversation, ensuring continuous and accurate capture of customer needs. The Geographic Information System (GIS) module is used to obtain the geographical location information of incoming callers and stores a database of contact information for regional investment managers. The Geographic Information System (GIS) module integrates multi-source geographic location data and services to construct a precise customer area mapping and a database of contact information for business development managers. Upon receiving an incoming call, the module first obtains the customer's geographic location information based on technologies such as number attribution resolution, base station positioning, or IP address positioning, and automatically matches it with preset regional division rules to determine the customer's service area. The system's built-in business development manager database records detailed contact information, responsibilities, and current service status of regional managers, ensuring that each region has a dedicated person in charge. To improve matching accuracy, the module also supports multi-level regional division and can dynamically adjust regional boundaries and manager assignments according to business needs. In addition, the module has an automatic data update mechanism; when business development manager information or regional divisions change, the system can update in real time to avoid transfer errors caused by information lag. Through high-precision geographic location matching and dynamic data maintenance, accurate targeted transfer of franchise inquiry calls is achieved.The intelligent routing and transfer control module communicates with both the voice robot module and the geographic information system module. The intelligent routing and transfer control module is used for the following: when the voice robot module identifies a call as a franchise inquiry, it triggers the geographic information system module to match the customer's geographical location and generates a first transfer instruction based on the matching result, directly transferring the call to the mobile phone of the corresponding regional sales manager; when the voice robot module identifies a non-franchise inquiry call, it generates a second transfer instruction based on a preset intelligent routing strategy, transferring the call to the appropriate human customer service agent. The intelligent routing and transfer control module integrates an intelligent routing strategy, which specifically performs multi-dimensional calculations based on the human customer service agent's skill set, current idle status, and historical efficiency in handling similar issues, to achieve load balancing for non-franchise inquiry calls. The intelligent routing and transfer control module executes differentiated routing strategies based on the intent recognition results of the voice robot module. When an incoming call is identified as a franchise inquiry, the module immediately triggers the geographic information system module to match the customer's location and generate the first transfer instruction, directly transferring the call to the mobile phone of the corresponding regional franchise manager, achieving a rapid response for localized service. For non-franchise inquiry calls, the module activates a multi-dimensional intelligent routing strategy. This strategy comprehensively considers the skill set expertise of human customer service agents, their current idle status, and historical efficiency data for handling similar issues. It calculates the suitability score of each agent through a weighted algorithm and selects the optimal agent for transfer. This strategy not only effectively avoids the problem of uneven resource allocation where some agents are overloaded while others are idle, but also improves problem-solving efficiency through precise matching of agent capabilities. The module also has a real-time monitoring function, which can dynamically adjust routing parameters according to system load to ensure the stability and efficiency of the transfer process.The information collaborative processing platform communicates with the voice robot module and the intelligent routing and transfer control module. The platform automatically executes information recording and distribution processes during or after call transfers. It includes an information capture and extraction unit, an electronic form generation unit, and an information distribution and synchronization unit. The information capture and extraction unit captures call content in real-time or near real-time through the voice recognition unit of the voice robot module and automatically extracts key information such as customer name, contact information, and intended region based on a preset key information model. The electronic form generation unit automatically generates the extracted key information. The system automatically populates standardized franchise consultation electronic forms. The information distribution and synchronization unit pushes the generated electronic forms to a shared database via a secure data interface and simultaneously sends notifications to the matched franchise managers and relevant collaborating departments via email, SMS, or at least one of the enterprise applications. The information collaboration processing platform is equipped with an information verification and completion mechanism. Specifically, after a call ends, the system automatically sends a confirmation SMS or email containing key information from the electronic form to the customer. If the customer's reply is incorrect, the system generates an information revision task and notifies the relevant franchise manager, achieving dynamic updates and calibration of the recorded information. The information collaboration processing platform automatically executes the entire process of information recording and distribution during or after a call transfer. The platform's information capture and extraction unit uses the voice recognition unit of the voice robot module to capture call content in real-time or near real-time and automatically extracts key information such as customer name, contact information, and intended region based on a preset key information model. The electronic form generation unit automatically fills the extracted information into a standardized franchise consultation electronic form, ensuring the uniformity and completeness of the information format. The information distribution and synchronization unit pushes the generated electronic form to the enterprise's shared database through a secure data interface and simultaneously sends notifications to the matched franchise manager and relevant collaborating departments via email, SMS, or enterprise applications. The platform's built-in information verification and completion mechanism automatically sends a confirmation SMS or email containing key information to the customer after the call ends. If the customer's reply is incorrect, the system automatically generates an information revision task and notifies the relevant franchise manager, achieving dynamic updates and calibration of the recorded information. This closed-loop information management process effectively improves the accuracy of information recording and collaborative efficiency. The information collaboration processing platform's output communication connection includes a business development manager's mobile terminal application, which is installed on the business development manager's mobile phone. This application is used to receive and display electronic forms and notifications in real time before or during a call, allowing the business development manager to obtain key customer information before answering the call.The mobile application for franchise managers is installed on their mobile phones and maintains real-time data synchronization with the information collaboration platform. When a franchise inquiry call is transferred, the application proactively pushes electronic forms and notifications to the customer before the call arrives or during the call. This allows franchise managers to fully understand the customer's key information, such as basic information, intended area, and historical inquiry records, before answering the call. The application interface adopts a clear information layout and visual design, highlighting key information to facilitate franchise managers' quick access to core content. In addition, the application supports online updates of customer information, adding follow-up notes, one-click call-back, and seamless integration with the enterprise CRM system to ensure complete recording of the service process and efficient follow-up. Through the real-time information synchronization and convenient operation functions of the mobile application, the service preparation time of franchise managers is shortened, and the professionalism and responsiveness of customer reception are improved. A method for handling franchise inquiries in a customer service system includes the following steps: a voice robot module answers the call and identifies the caller's intent through voice recognition and natural language processing technology; if it is a franchise inquiry, a geographic information system module is invoked to obtain the customer's location and match them with a franchise manager, while simultaneously initiating the information recording process of the information collaboration processing platform; an intelligent routing and transfer control module transfers the call to the matched franchise manager, and the information collaboration processing platform simultaneously pushes the generated inquiry information form to the franchise manager; if it is not a franchise inquiry, the call is transferred to the optimal human customer service agent according to the intelligent routing strategy. This application achieves highly efficient and automated processing of franchise inquiries through the coordinated operation of a voice robot module, a geographic information system module, an intelligent routing and transfer control module, and an information collaboration processing platform. The voice robot module first answers the call and uses an end-to-end deep learning model to recognize and analyze the customer's voice, accurately distinguishing between franchise inquiries and non-franchise inquiries. If it is determined to be a franchise inquiry, the geographic information system module is simultaneously triggered to match the customer's geographical location, identify the corresponding franchise manager, and initiate the information recording process of the information collaboration processing platform, automatically extracting and generating the customer's electronic form. The intelligent routing and transfer control module then generates a first transfer instruction, transferring the call to the matched franchise manager's mobile phone. Simultaneously, the information collaboration processing platform pushes the electronic form to the franchise manager's mobile terminal application. If it is determined to be a non-franchise inquiry, the intelligent routing and transfer control module calculates the optimal human agent based on a multi-dimensional intelligent routing strategy and generates a second transfer instruction to achieve load balancing. This achieves automation and intelligence in intent recognition, routing, and information collaboration, improving the processing efficiency of the customer service system and the customer experience.This application employs an end-to-end deep learning model, integrating speech recognition and intent recognition into a single model for joint optimization. This avoids the error accumulation and amplification problem inherent in traditional sequential models. Through attention mechanisms and contextual semantic understanding, it significantly improves the accuracy and robustness of intent recognition in complex scenarios such as noisy environments and diverse accents, ensuring the reliability of customer call intent judgment. This application also implements a three-level precise routing system of "intent-location-agent." For franchise inquiries, a high-precision geographic information system enables accurate matching and direct transfer of "customer-region-recruitment manager," greatly shortening the transfer path. For non-franchise inquiries, a multi-dimensional intelligent algorithm based on skills, availability, and historical efficiency is used for load balancing, ensuring customer... This application optimizes the utilization of service resources and efficiently resolves customer issues. Through an information collaborative processing platform, it automatically captures and extracts key customer information during calls, generating standardized electronic forms. This eliminates inefficient manual recording. More importantly, the platform can push form information to the sales manager's mobile terminal in real time during call transfer, achieving "zero-delay" information synchronization. This ensures that the sales manager has customer details before answering the call and provides an accurate data foundation for subsequent follow-up. The seamless automation of the entire process significantly shortens customer waiting time and optimizes enterprise resource utilization through intelligent load balancing of agents. This improves customer satisfaction while also enhancing internal operational efficiency and management level. The above are preferred embodiments of this application and are not intended to limit the scope of protection of this application. Therefore, all equivalent changes made to the structure, shape, and principle of this application should be included within the scope of protection of this application.

Claims

1. A customer service system, characterized in that: include: The system includes a voice robot module for responding to customer calls, which integrates a voice recognition unit and a natural language processing unit; and a geographic information system module for acquiring the geographical location information of the calling customer and storing a database of contact information for regional sales managers. An intelligent routing and transfer control module is communicatively connected to both the voice robot module and the geographic information system module. An information collaborative processing platform is communicatively connected to the voice robot module and the intelligent routing and transfer control module. The information collaborative processing platform is used to automatically execute information recording and distribution processes during or after telephone transfer.

2. The customer service system according to claim 1, characterized in that: The voice robot module converts customer speech into text through an integrated speech recognition unit and analyzes the text through a natural language processing unit to identify the customer's caller intent and distinguish between franchise-related and non-franchise-related calls.

3. A customer service system according to claim 1, characterized in that: The intelligent routing and transfer control module is used to: when the voice robot module identifies a call as a franchise inquiry, trigger the geographic information system module to match the customer's geographical location and generate a first transfer instruction based on the matching result, directly transferring the call to the mobile phone of the corresponding regional business development manager; when the voice robot module identifies a call as a non-franchise inquiry, generate a second transfer instruction based on a preset intelligent routing strategy, transferring the call to the corresponding human customer service agent.

4. A customer service system according to claim 1, characterized in that: The information collaborative processing platform includes an information capture and extraction unit, an electronic form generation unit, and an information distribution and synchronization unit. The information capture and extraction unit is used to capture the call content in real time or near real time through the voice recognition unit of the voice robot module during the call and automatically extract key information such as customer name, contact information, and intended area based on a preset key information model. The electronic form generation unit is used to automatically fill the extracted key information into a standardized franchise consultation electronic form. The information distribution and synchronization unit is used to push the generated electronic form to a shared database through a secure data interface and simultaneously send notifications and reminders to the matched franchise manager and relevant collaborative departments through at least one of the following methods: email, SMS, or enterprise application.

5. A customer service system according to claim 4, characterized in that: The output end of the information collaboration processing platform is connected to a mobile terminal application for investment managers. The mobile terminal application for investment managers is installed on the investment manager's mobile phone. The mobile terminal application for investment managers is used to receive and display the electronic forms and notification reminders in real time before the call is transferred or during the call, so that the investment manager can obtain key customer information before answering the call.

6. A customer service system according to claim 1, characterized in that: The intelligent routing and transfer control module integrates an intelligent routing strategy, which specifically involves multi-dimensional calculations based on the skill set of human customer service agents, their current idle status, and their historical efficiency in handling similar issues, in order to achieve load balancing for non-franchise inquiry calls.

7. A customer service system according to claim 4, characterized in that: The information collaborative processing platform is equipped with an information verification and completion mechanism. Specifically, after a call ends, the system automatically sends a confirmation SMS or email to the customer containing key information from the electronic form. If the customer's reply is incorrect, the system generates an information revision task and notifies the relevant sales manager, thereby achieving dynamic updating and calibration of the recorded information.

8. A customer service system according to claim 1, characterized in that: The voice robot module adopts an end-to-end deep learning model, which integrates speech recognition and intent recognition into a unified model for joint optimization, in order to improve the accuracy of intent recognition in complex call environments.

9. A method for handling franchise inquiries in a customer service system, wherein the customer service system is described in any one of claims 1-8, characterized in that: Includes the following steps: The voice robot module answers incoming calls, using speech recognition and natural language processing technology to identify the caller's intent. If it is a franchise consultation, the geographic information system module is called to obtain the customer's location and match the franchise manager, and the information recording process of the information collaboration processing platform is initiated at the same time. The intelligent routing and transfer control module transfers calls to the matching business development manager, while the information collaboration processing platform simultaneously pushes the generated consultation information form to the business development manager. For inquiries from non-franchisees, the call will be transferred to the optimal human customer service representative based on the intelligent routing strategy.

Images