Non-intrusive call center system of intelligent voice robot

Abstract

The invention relates to the technical field of call centers, and discloses an intelligent voice robot non-intrusive call center system. Comprising a non-intrusive docking module, an intelligent voice robot core module, a call scheduling module, an interaction management module, a data acquisition and analysis module, an authority control module, a fault-tolerant backup module, a terminal adaptation module, a knowledge base module and a visual operation and maintenance module. According to the invention, non-intrusive docking is realized, codes and architecture of an original business system of an enterprise do not need to be transformed, seamless data intercommunication with the original business system of the enterprise is realized through interface adaptation, data isolation and protocol conversion technologies, the system deployment difficulty and transformation cost are reduced, and the stability and data integrity of the original system are prevented from being damaged; meanwhile, a one-way data reading and encryption transmission mechanism is adopted, the data transmission safety is improved, and the method is suitable for various old and novel service systems and high in compatibility.

Description

Intelligent voice robot non-intrusive call center system Technical Field

[0001] This invention belongs to the field of call center technology, specifically a non-intrusive intelligent voice robot call center system. Background Technology

[0002] With the rapid development of artificial intelligence and speech recognition technologies, intelligent voice robots have been widely used in call center scenarios of various enterprises to replace human agents in handling a large number of repetitive and standardized call tasks, reducing labor costs and improving customer service efficiency. However, existing intelligent voice robot call center systems often employ intrusive integration methods when interfacing with existing enterprise business systems (such as CRM, ERP, and customer service systems). This requires modification and reconstruction of the existing system's code and interfaces, and may even necessitate replacing some modules to achieve data interoperability and business collaboration.

[0003] This intrusive integration method has several drawbacks: First, it is difficult to modify, requiring a significant investment of technical manpower and time, and for older business systems, it may even be impossible to complete the modification. Second, it has poor compatibility, as the interface protocols and data formats of different companies' original business systems vary greatly, making intrusive integration difficult to adapt to various systems and prone to problems such as abnormal data transmission and system conflicts. Third, it has low security, as modifying the original business system code and interfaces can compromise the stability of the original system, and bidirectional data writing can easily lead to risks such as leakage of customer privacy data and data corruption. Fourth, it is cumbersome to deploy, as intrusive integration requires downtime modification of the original system, affecting the normal operation of the enterprise and resulting in a long deployment cycle. Fifth, it has poor scalability, as the call center system needs to be modified simultaneously when the enterprise's original business system is upgraded or iterated, resulting in high maintenance costs.

[0004] Furthermore, existing intelligent voice robot call center systems suffer from poor interactive experience, unreasonable call scheduling, and inconvenient operation and maintenance, making it difficult to meet enterprises' needs for non-intrusive, highly compatible, highly secure, and highly efficient call center systems. Therefore, developing an intelligent voice robot call center system that can achieve non-intrusive integration, strong compatibility, efficient deployment, and accurate interaction has become an urgent technical challenge. Summary of the Invention

[0005] In view of the above situation and to overcome the shortcomings of the prior art, the present invention provides a non-intrusive intelligent voice robot call center system, which effectively solves the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a non-intrusive intelligent voice robot call center system, comprising a non-intrusive interface module, an intelligent voice robot core module, a call scheduling module, an interaction management module, a data acquisition and analysis module, an access control module, a fault-tolerant backup module, a terminal adaptation module, a knowledge base module, and a visualized operation and maintenance module; the non-intrusive interface module is connected to the intelligent voice robot core module and the enterprise's existing business system; the intelligent voice robot core module is connected to the call scheduling module, the interaction management module, and the knowledge base module; the call scheduling module is connected to the terminal adaptation module; the interaction management module is connected to the data acquisition and analysis module and the access control module; the fault-tolerant backup module is bidirectionally connected to each module; and the visualized operation and maintenance module is connected to the data acquisition and analysis module, the fault-tolerant backup module, and the access control module; all modules work together to achieve non-intrusive intelligent processing throughout the entire call lifecycle.

[0007] Preferably, the non-intrusive interface module includes an interface adaptation unit, a data isolation unit, and a protocol conversion unit. The interface adaptation unit is used to adapt to various open-source and private interfaces of the enterprise's original business system without modifying the interface configuration and code of the original system. The data isolation unit is used to achieve physical data isolation between the system and the enterprise's original business system, adopting a one-way data reading and encrypted transmission mechanism, only obtaining the basic customer information and business-related data required for call processing, without writing any data to the original business system. The protocol conversion unit is used to convert various data transmission protocols of the enterprise's original business system into the unified and compatible HTTP / HTTPS protocol of the system, realizing data interoperability.

[0008] Preferably, the core module of the intelligent voice robot includes a speech recognition unit, a speech synthesis unit, an intent recognition unit, and a multi-turn dialogue management unit. The speech recognition unit adopts a hybrid noise reduction and recognition model based on the Transformer architecture to convert customer speech signals into text information, with a recognition accuracy of no less than 98%, and supports speech recognition in dialects and noisy environments. The speech synthesis unit adopts an emotional speech synthesis algorithm to convert system response text into natural and fluent speech signals, supporting dynamic adjustment of speech rate and tone to match the customer's emotional state. The intent recognition unit adopts a deep learning fusion model, combined with keyword matching and contextual association analysis, to accurately identify the customer's call intent, with a recognition response time of no more than 100ms. The multi-turn dialogue management unit is used to maintain the dialogue context state, record dialogue history information and slot filling data, realize multi-turn coherent dialogue, and support dialogue breakpoint resumption.

[0009] Preferably, the call scheduling module includes a number allocation unit, a line monitoring unit, and an intelligent transfer unit. The number allocation unit uses a load balancing algorithm to automatically allocate the optimal call line and number based on the line's idle status and call priority, avoiding line congestion. The line monitoring unit monitors the operating status of all call lines in real time, including line connectivity, call quality, and bandwidth utilization, and immediately triggers an alarm when a line is abnormal. The intelligent transfer unit automatically transfers the call to an available human agent in the corresponding business area when the intelligent voice robot cannot handle the customer's intent or the customer explicitly requests a human agent, and simultaneously pushes call history, customer intent, and business-related data to the human agent's terminal to achieve seamless transfer.

[0010] Preferably, the interaction management module includes a customer emotion analysis unit, an interaction process customization unit, and an abnormal interaction handling unit. The customer emotion analysis unit analyzes the customer's emotional state in real time based on the customer's speech rate, tone, and text semantics, classifying it into four emotion types: calm, satisfied, anxious, and angry. The emotion analysis results are then synchronized to the speech synthesis unit and the intelligent transfer unit. The interaction process customization unit allows enterprises to visually customize call interaction processes for different business scenarios according to their own business needs, without requiring code development. The abnormal interaction handling unit handles various abnormal situations during the call process, including customer hang-up, speech recognition failure, and line interruption. When an abnormality occurs, it automatically records the abnormal information and executes a preset processing strategy. After the call is hung up, it automatically triggers a callback reminder (customer authorization required).

[0011] Preferably, the data acquisition and analysis module includes a full data acquisition unit, a multi-dimensional analysis unit, and a data visualization unit. The full data acquisition unit is used to collect various types of data throughout the entire call lifecycle, including call duration, connection rate, customer intent, interaction rounds, emotional state, transfer status, and operation and maintenance logs. The acquisition frequency is real-time, and the data storage duration is no less than one year. The multi-dimensional analysis unit is used to perform statistical analysis on the collected data, including business volume analysis, customer satisfaction analysis, robot processing efficiency analysis, and human agent workload analysis, generating multi-dimensional analysis reports. The data visualization unit is used to display the analysis results in chart form, supports data drill-down queries, and facilitates staff to quickly grasp the system's operating status and business development.

[0012] Preferably, the access control module includes a role definition unit, a permission allocation unit, and an operation auditing unit. The role definition unit is used to preset various user roles in the system, including super administrators, operation and maintenance personnel, business administrators, and human agents, and to clarify the operation scope of each role. The permission allocation unit is used to finely allocate system operation permissions according to user roles, to achieve minimal permission control, and to support dynamic adjustment and revocation of permissions. The operation auditing unit is used to record the system operation behavior of all users, including operation time, operation content, operation results, and terminal information. The audit log is tamper-proof and facilitates subsequent traceability and compliance checks.

[0013] Preferably, the fault-tolerant backup module includes an anomaly monitoring unit, a real-time backup unit, and a fault recovery unit. The anomaly monitoring unit monitors the operating status of each module in real time, and immediately triggers the fault tolerance mechanism and alarm notification when a module failure, data anomaly, or line interruption is detected. The real-time backup unit adopts a master-slave backup architecture to perform real-time synchronous backup of system configuration data, call data, and knowledge base data. The backup data is stored on a remote server to ensure that no data is lost. The fault recovery unit automatically switches to the backup system when a system failure occurs, quickly restoring system operation, with a fault recovery time of no more than 5 minutes.

[0014] Preferably, the terminal adaptation module is used to adapt to different types of terminal devices, including landline phones, smartphones, computers, and smart wearable devices, supporting call access and interactive operations of different terminals; the knowledge base module includes a knowledge input unit, a knowledge update unit, and an intelligent retrieval unit. The knowledge input unit supports knowledge input in three formats: text, voice, and images. The knowledge update unit supports manual and automatic updates and can be integrated with the enterprise's existing knowledge base for synchronous updates. The intelligent retrieval unit is used to quickly retrieve relevant knowledge in the knowledge base based on customer intent and query keywords, with a retrieval response time of no more than 50ms.

[0015] Preferably, the visualized operation and maintenance module includes a system monitoring unit, an operation and maintenance unit, and an alarm management unit; the system monitoring unit is used to monitor the overall operating status of the system in real time, including module operating status, line status, data transmission status, and resource usage; the operation and maintenance unit is used by staff to perform operation and maintenance operations such as system configuration, fault diagnosis, and data cleanup, and supports remote operation and maintenance; the alarm management unit is used to receive alarm information from each module, classify and display alarm content, alarm level, and alarm location, and support multi-channel push of alarm notifications, including SMS, email, and system messages.

[0016] Compared with existing technologies, the beneficial effects of this invention are: 1. This invention achieves non-intrusive integration, requiring no modification to the code and architecture of the enterprise's original business system. Through interface adaptation, data isolation, and protocol conversion technologies, it achieves seamless data interoperability with the enterprise's original business system, reducing system deployment difficulty and modification costs, avoiding damage to the stability and data integrity of the original system. Simultaneously, it employs a one-way data reading and encrypted transmission mechanism to enhance data transmission security, making it suitable for various old and new business systems with strong compatibility; 2. This invention provides an excellent intelligent voice interaction experience, employing a hybrid noise reduction speech recognition model based on the Transformer architecture and an emotional speech synthesis algorithm to improve speech recognition accuracy and naturalness. It supports recognition in dialects and noisy environments, and combined with customer emotion analysis and multi-turn dialogue management technology, it achieves accurate, coherent, and humanized voice interaction, enhancing the customer service experience; 3. This invention has a high degree of intelligent call scheduling, employing a load balancing algorithm to implement call lines. 4. This invention offers convenient and efficient operation and maintenance management. A visual operation and maintenance module enables real-time monitoring and remote operation and maintenance of the overall system status. Combined with a fault-tolerant backup module, it achieves anomaly monitoring, real-time backup, and rapid fault recovery, ensuring stable system operation and reducing manpower investment in operation and maintenance. Simultaneously, a data collection and analysis module generates multi-dimensional analysis reports, providing data support for enterprise business decisions. 5. This invention boasts high security and reliability. A permission control module enables refined permission management and operation auditing, ensuring system operation security and preventing unauthorized operations. A fault-tolerant backup module enables real-time data backup and rapid fault recovery, ensuring no data loss and no system interruption. Data isolation and encrypted transmission technologies protect customer privacy and meet compliance requirements.

[0017] 6. This invention is highly scalable and flexible, supporting the visual customization of interactive processes and the custom addition and modification of intents and knowledge. The system functions can be flexibly adjusted according to the business needs of enterprises, adapting to the call center needs of enterprises of different industries and sizes. At the same time, it supports multi-type adaptation of terminal devices, improving the applicability of the system. Attached Figure Description

[0018] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof.

[0019] In the accompanying drawings: Figure 1 is an architecture diagram of the non-intrusive intelligent voice robot call center system of the present invention; Figure 2 is a logic block diagram of the non-intrusive intelligent voice robot call center system of the present invention; Figure 3 is a call processing flowchart of the non-intrusive intelligent voice robot call center system of the present invention. Detailed Implementation

[0020] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0021] As shown in Figures 1-3, this invention relates to a non-intrusive call center system with an intelligent voice robot, comprising a non-intrusive interface module, an intelligent voice robot core module, a call dispatch module, an interaction management module, a data acquisition and analysis module, an access control module, a fault-tolerant backup module, a terminal adaptation module, a knowledge base module, and a visualized operation and maintenance module. The non-intrusive interface module connects to the intelligent voice robot core module and the enterprise's existing business system. The intelligent voice robot core module connects to the call dispatch module, the interaction management module, and the knowledge base module. The call dispatch module connects to the terminal adaptation module. The interaction management module connects to the data acquisition and analysis module and the access control module. The fault-tolerant backup module connects bidirectionally to each module. The visualized operation and maintenance module connects to the data acquisition and analysis module, the fault-tolerant backup module, and the access control module. All modules work together to achieve non-intrusive intelligent processing throughout the entire call lifecycle.

[0022] The non-intrusive integration module includes an interface adaptation unit, a data isolation unit, and a protocol conversion unit. The interface adaptation unit adapts to various open-source and private interfaces of the enterprise's existing business systems without requiring modification to the original system's interface configuration and code. Adapted interface types include RESTful, SOAP, JDBC, and WebSocket interfaces, and the unit automatically matches the adaptation strategy based on the interface type of the enterprise's existing business systems. The data isolation unit achieves physical data isolation between the system and the enterprise's existing business systems. It employs a one-way data reading and encrypted transmission mechanism, only acquiring basic customer information (including customer name, contact information, and customer level) and business-related data (including customer historical business records and business processing requirements) required for call processing. No data is written to the existing business systems. Encrypted transmission uses the AES-256 encryption algorithm to ensure data transmission security. The protocol conversion unit converts various data transmission protocols of the enterprise's existing business systems into the system's unified and compatible HTTP / HTTPS protocol, enabling data interoperability. Supported protocol types include TCP / IP, UDP, FTP, and SFTP protocols, with a protocol conversion latency of no more than 30ms.

[0023] The core modules of the intelligent voice robot include a speech recognition unit, a speech synthesis unit, an intent recognition unit, and a multi-turn dialogue management unit. The speech recognition unit employs a hybrid noise reduction and recognition model based on the Transformer architecture to convert customer speech signals into text information, achieving an accuracy rate of no less than 98%. It supports multiple dialects such as Mandarin, Cantonese, and Sichuanese, as well as speech recognition in noisy environments. The speech signal sampling rate is 8kHz-48kHz, supporting real-time recognition with a recognition latency of no more than 80ms. The speech synthesis unit uses an emotional speech synthesis algorithm to convert system response text into natural and fluent speech signals, supporting dynamic adjustment of speech rate (50 words / minute - 200 words / minute) and intonation. The system can automatically match the corresponding voice tone according to the customer's emotional state, with a speech synthesis delay of no more than 50ms. The intent recognition unit uses a deep learning fusion model, combined with keyword matching and contextual analysis, to accurately identify the customer's call intent, with a recognition response time of no more than 100ms. The types of intents that can be recognized include business inquiries, business processing, complaints and suggestions, fault reporting, and manual transfer. It supports the addition and modification of custom intents. The multi-turn dialogue management unit is used to maintain the dialogue context state, record dialogue history information and slot filling data, realize multi-turn coherent dialogue, and support dialogue breakpoint resumption. When a customer interrupts the dialogue and calls again, it can automatically restore the previous dialogue state without the customer having to repeat themselves.

[0024] The call dispatch module includes a number allocation unit, a line monitoring unit, and an intelligent transfer unit. The number allocation unit uses a load balancing algorithm (combining round-robin and minimum connection number algorithms) to automatically allocate the optimal call line and number based on line availability and call priority (divided into three levels: normal call, priority call, and emergency call), avoiding line congestion and increasing line utilization to over 90%. The line monitoring unit monitors the operating status of all call lines in real time, including line connectivity, call quality (signal-to-noise ratio not less than 30dB), and bandwidth utilization. When a line abnormality occurs (such as line interruption, call lag, or bandwidth utilization exceeding 80%), an alarm is immediately triggered. The intelligent transfer unit automatically transfers calls to available human agents in the corresponding business area when the intelligent voice robot cannot process customer intent (intent recognition accuracy is less than 85%) or when the customer explicitly requests a human agent. It also simultaneously pushes call history, customer intent, and business-related data to the human agent's terminal, achieving seamless transfer with a transfer delay of no more than 200ms and a transfer success rate of no less than 99%.

[0025] The interaction management module includes a customer sentiment analysis unit, an interaction flow customization unit, and an abnormal interaction handling unit. The customer sentiment analysis unit analyzes the customer's emotional state in real time based on speech rate, tone, and text semantics, categorizing it into four emotion types: calm, satisfied, anxious, and angry. The sentiment analysis accuracy is no less than 95%, and the results are synchronized to the speech synthesis unit and intelligent transfer unit. When a customer's emotion is anxious or angry, the speech synthesis unit automatically adjusts to a gentler tone, and the intelligent transfer unit prioritizes transferring the call to a human agent. The interaction flow customization unit allows businesses to visually customize calls for different business scenarios according to their own business needs. The interactive workflow requires no code development and supports drag-and-drop editing, saving, and activation. Customizable workflows include welcome messages, business consultations, business processing, and complaint handling. The abnormal interaction handling unit is used to handle various abnormal situations during the call, including customer hang-up, voice recognition failure, and line interruption. When an abnormality occurs, it automatically records the abnormal information (including abnormality type, abnormal time, and customer information) and executes the preset handling strategy. After hanging up, it automatically triggers a callback reminder (with customer authorization required). When voice recognition fails, it automatically triggers a repeat inquiry mechanism. When the line is interrupted, it automatically attempts to reconnect. If the connection fails, it records the failure and issues an alarm.

[0026] The data acquisition and analysis module includes a full data acquisition unit, a multi-dimensional analysis unit, and a data visualization unit. The full data acquisition unit collects various data throughout the entire call lifecycle, including call duration, connection rate, customer intent, interaction rounds, emotional state, transfer status, and operation logs. Data is collected in real-time, stored for at least one year using a distributed database, and supports rapid data querying and batch export. The multi-dimensional analysis unit performs statistical analysis on the collected data, including business volume analysis (call volume statistics by time period and business type), customer satisfaction analysis (combining emotional state and interaction feedback statistics), robot processing efficiency analysis (processing success rate and processing time statistics by intent type), and human agent workload analysis (transfer volume and processing time statistics by agent). It generates multi-dimensional analysis reports, which can be exported daily, weekly, monthly, and quarterly. The data visualization unit displays the analysis results in chart format (bar charts, line charts, pie charts, heatmaps), supports data drill-down queries, and facilitates staff's quick understanding of system operation status and business progress.

[0027] The access control module includes a role definition unit, a permission allocation unit, and an operation audit unit. The role definition unit pre-defines various user roles within the system, including super administrators, operations and maintenance personnel, business administrators, and human agents, clearly defining the operational scope of each role. Super administrators have full system operation permissions, operations and maintenance personnel only have system monitoring and troubleshooting permissions, business administrators have interactive process customization and knowledge base management permissions, and human agents only have call processing and customer information query permissions. The permission allocation unit finely allocates system operation permissions based on user roles, achieving minimal permission control and supporting dynamic adjustment and revocation of permissions. Permission allocation records are traceable. The operation audit unit records all user system operation behaviors, including operation time, operation content, operation results, and terminal information (IP address, device model). Audit logs are tamper-proof and stored for at least two years for easy traceability and compliance checks.

[0028] The fault-tolerant backup module includes an anomaly monitoring unit, a real-time backup unit, and a fault recovery unit. The anomaly monitoring unit monitors the operational status of each module in real time using a heartbeat detection mechanism, detecting once per second. When a module failure, data anomaly, or line interruption is detected, the fault-tolerant mechanism and alarm notification are immediately triggered. Alarm levels are categorized as general alarms, important alarms, and emergency alarms. The real-time backup unit employs a master-slave backup architecture, performing real-time synchronous backups of system configuration data, call data, and knowledge base data. Backup data is stored on a remote server, with real-time synchronization to ensure no data loss. It supports backup data verification and recovery testing. The fault recovery unit automatically switches to the backup system when a system failure occurs, quickly restoring system operation. Fault recovery time is no more than 5 minutes. After recovery, the data from the fault period is automatically synchronized to ensure data integrity.

[0029] The terminal adaptation module is used to adapt to different types of terminal devices, including landlines, smartphones, computers, and smart wearable devices. It supports call access and interactive operations on different terminals. For smartphones and computers, it also supports text interaction and voice recording and playback functions. The knowledge base module includes a knowledge input unit, a knowledge update unit, and an intelligent retrieval unit. The knowledge input unit supports knowledge input in multiple formats such as text, voice, and images, and supports batch input and single input. The input knowledge is automatically classified and archived. The knowledge update unit supports manual and automatic updates and can be integrated with the enterprise's existing knowledge base for synchronous updates. The automatic update frequency can be customized (daily or weekly). The intelligent retrieval unit is used to quickly retrieve relevant knowledge in the knowledge base based on customer intent and query keywords. The retrieval response time is no more than 50ms, and the retrieval results are sorted by relevance. It supports fuzzy and precise retrieval of knowledge.

[0030] The visualized operation and maintenance module includes a system monitoring unit, an operation and maintenance unit, and an alarm management unit. The system monitoring unit is used to monitor the overall system operation status in real time, including module operation status, line status, data transmission status, and resource usage (CPU utilization, memory utilization, and disk space utilization), and supports custom settings of monitoring indicators. The operation and maintenance unit is used by staff to perform operation and maintenance operations such as system configuration, fault diagnosis, and data cleanup. It supports remote operation and maintenance and can be performed through computer terminals and mobile terminals. The alarm management unit is used to receive alarm information from various modules, classify and display alarm content, alarm level, and alarm location, and supports multi-channel push of alarm notifications, including SMS, email, and system messages. It supports alarm confirmation, processing, and archiving, making it convenient for operation and maintenance personnel to track and handle alarms.

[0031] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0032] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A non-intrusive intelligent voice robot call center system, characterized in that, The system includes a non-intrusive interface module, an intelligent voice robot core module, a call scheduling module, an interaction management module, a data acquisition and analysis module, an access control module, a fault-tolerant backup module, a terminal adaptation module, a knowledge base module, and a visualized operation and maintenance module. The non-intrusive interface module connects to the intelligent voice robot core module and the enterprise's existing business system. The intelligent voice robot core module connects to the call scheduling module, the interaction management module, and the knowledge base module. The call scheduling module connects to the terminal adaptation module. The interaction management module connects to the data acquisition and analysis module and the access control module. The fault-tolerant backup module connects bidirectionally to each module. The visualized operation and maintenance module connects to the data acquisition and analysis module, the fault-tolerant backup module, and the access control module. All modules work together to achieve non-intrusive intelligent processing throughout the entire call lifecycle.

2. The intelligent voice robot non-intrusive call center system according to claim 1, characterized in that: The non-intrusive interface module includes an interface adaptation unit, a data isolation unit, and a protocol conversion unit. The interface adaptation unit is used to adapt to various open-source and private interfaces of the enterprise's original business system without modifying the interface configuration and code of the original system. The data isolation unit is used to achieve physical data isolation between the system and the enterprise's original business system. It adopts a one-way data reading and encrypted transmission mechanism, only obtaining the basic customer information and business-related data required for call processing, without writing any data to the original business system. The protocol conversion unit is used to convert various data transmission protocols of the enterprise's original business system into the unified and compatible HTTP / HTTPS protocol of the system to achieve data interoperability.

3. The intelligent voice robot non-intrusive call center system according to claim 1, characterized in that: The core modules of the intelligent voice robot include a speech recognition unit, a speech synthesis unit, an intent recognition unit, and a multi-turn dialogue management unit. The speech recognition unit employs a hybrid noise reduction and recognition model based on the Transformer architecture to convert customer speech signals into text information, achieving an accuracy rate of no less than 98% and supporting speech recognition in dialects and noisy environments. The speech synthesis unit uses an emotional speech synthesis algorithm to convert system response text into natural and fluent speech signals, supporting dynamic adjustment of speech rate and tone to match the customer's emotional state. The intent recognition unit uses a deep learning fusion model, combining keyword matching and contextual analysis, to accurately identify the customer's call intent, with a response time of no more than 100ms. The multi-turn dialogue management unit maintains the dialogue context state, records dialogue history information and slot filling data, enabling continuous multi-turn dialogue and supporting breakpoint continuation.

4. The intelligent voice robot non-intrusive call center system according to claim 1, characterized in that: The call dispatch module includes a number allocation unit, a line monitoring unit, and an intelligent transfer unit; the number allocation unit adopts a load balancing algorithm to automatically allocate the optimal call line and number based on the line idle status and call priority, thereby avoiding line congestion; The line monitoring unit is used to monitor the operating status of all call lines in real time, including line connectivity, call quality and bandwidth utilization. When a line is abnormal, an alarm is triggered immediately. The intelligent transfer unit is used to automatically transfer the call to an available human agent in the corresponding business area when the intelligent voice robot cannot process the customer's intent or the customer explicitly requests a human agent. It also pushes the call history, customer intent, and business-related data to the human agent's terminal simultaneously to achieve seamless transfer.

5. The intelligent voice robot non-intrusive call center system according to claim 1, characterized in that: The interaction management module includes a customer emotion analysis unit, an interaction process customization unit, and an abnormal interaction handling unit. The customer emotion analysis unit is used to analyze the customer's emotional state in real time by analyzing the customer's speech rate, tone, and text semantics. It is divided into four emotion types: calm, satisfied, anxious, and angry. The emotion analysis results are synchronized to the speech synthesis unit and the intelligent transfer unit. The interactive process customization unit is used to support enterprises to visually customize call interactive processes for different business scenarios according to their own business needs, without the need for code development; The abnormal interaction processing unit is used to handle various abnormal situations during the call process, including customer hanging up midway, voice recognition failure, and line interruption. When an abnormality occurs, it automatically records the abnormal information and executes the preset processing strategy. After hanging up, it automatically triggers a callback reminder.

6. The intelligent voice robot non-intrusive call center system according to claim 1, characterized in that: The data acquisition and analysis module includes a full data acquisition unit, a multi-dimensional analysis unit, and a data visualization unit. The full data acquisition unit is used to collect various types of data throughout the entire call lifecycle, including call duration, connection rate, customer intent, interaction rounds, emotional state, transfer status, and operation and maintenance logs. The acquisition frequency is real-time, and the data storage duration is no less than one year. The multi-dimensional analysis unit is used to perform statistical analysis on the collected data, including business volume analysis, customer satisfaction analysis, robot processing efficiency analysis, and human agent workload analysis, and generate multi-dimensional analysis reports; the data visualization unit is used to display the analysis results in the form of charts and graphs, supports data drill-down queries, and makes it easy for staff to quickly grasp the system operation status and business development.

7. The intelligent voice robot non-intrusive call center system according to claim 1, characterized in that: The access control module includes a role definition unit, a permission allocation unit, and an operation auditing unit. The role definition unit is used to preset various user roles in the system, including super administrators, maintenance personnel, business administrators, and human agents, and to clarify the operation scope of each role. The permission allocation unit is used to finely allocate system operation permissions according to user roles, to achieve minimal permission control, and to support dynamic adjustment and revocation of permissions. The operation auditing unit is used to record the system operation behavior of all users, including operation time, operation content, operation results, and terminal information. The audit log is tamper-proof and facilitates subsequent traceability and compliance checks.

8. The intelligent voice robot non-intrusive call center system according to claim 1, characterized in that: The fault-tolerant backup module includes an anomaly monitoring unit, a real-time backup unit, and a fault recovery unit. The anomaly monitoring unit is used to monitor the operating status of each module in real time. When a module failure, data anomaly, or line interruption is detected, the fault tolerance mechanism and alarm notification are immediately triggered. The real-time backup unit adopts a master-slave backup architecture to perform real-time synchronous backup of system configuration data, call data, and knowledge base data. The backup data is stored on a remote server to ensure that the data is not lost. The fault recovery unit is used to automatically switch to the backup system when the system fails, quickly restore system operation, and the fault recovery time is no more than 5 minutes.

9. The intelligent voice robot non-intrusive call center system according to claim 1, characterized in that: The terminal adaptation module is used to adapt to different types of terminal devices, including landline phones, smartphones, computers, and smart wearable devices, supporting call access and interactive operations on different terminals. The knowledge base module includes a knowledge input unit, a knowledge update unit, and an intelligent retrieval unit. The knowledge input unit supports knowledge input in three formats: text, voice, and images. The knowledge update unit supports manual and automatic updates and can be integrated with the enterprise's existing knowledge base for synchronous updates. The intelligent retrieval unit is used to quickly retrieve relevant knowledge in the knowledge base based on customer intent and query keywords, with a retrieval response time of no more than 50ms.

10. The intelligent voice robot non-intrusive call center system according to claim 1, characterized in that: The visualized operation and maintenance module includes a system monitoring unit, an operation and maintenance unit, and an alarm management unit. The system monitoring unit is used to monitor the overall system operation status in real time, including module operation status, line status, data transmission status, and resource usage. The operation and maintenance unit is used by staff to perform operation and maintenance operations such as system configuration, fault diagnosis, and data cleanup, and supports remote operation and maintenance. The alarm management unit is used to receive alarm information from each module, classify and display alarm content, alarm level, and alarm location, and supports multi-channel push of alarm notifications, including SMS, email, and system messages.

Images