External incoming call processing method and system, call processing method and system and electronic equipment

Abstract

The invention discloses an external incoming call processing method, an external incoming call processing method, an external incoming call processing system and electronic equipment. The external incoming call processing method comprises the following steps: matching a calling number of an external incoming call with a preset list database so as to classify the external incoming call into a white list number, a black list number or an unknown number; if the classification result is a white list, a ringing is initiated to the user terminal to prompt the user to get through; if the user is not connected within the specified time limit, a digital human assistant is started for answering; if the classification result is a blacklist, intercepting the external incoming call; if the number is classified as an unknown number, starting a digital human assistant to execute substitution, and receiving an incoming call intention fed back by the digital human assistant; judging whether the current call needs to be transferred to the user terminal or not; and if it is judged that switching is not needed, the digital human assistant performs voice reply based on a preset verbal skill library and the current conversation context. According to the external incoming call processing method, refined and intelligent management of the external incoming call can be realized.

Description

Technical Field

[0001] This invention relates to the field of communication technology, and in particular to an external call processing method, telephone processing method, system, and electronic device. Background Technology

[0002] With the widespread application of mobile communication technology and smart devices, telephones have become an indispensable means of communication in people's daily lives and work. However, the accompanying problems of telephone harassment, invalid information delivery, and users being unable to answer calls in certain situations have become increasingly prominent and widespread annoyances. Especially in business office settings, employees are often burdened by various sales calls and repetitive inquiries, which not only disrupts their normal work rhythm but also seriously affects their work efficiency.

[0003] Most current telephone answering solutions rely on pre-set rules and fixed script templates to assess the risk and categorize unknown calls. However, this approach has significant limitations: First, pre-set rules and scripts struggle to keep up with the diverse changes in actual call scenarios, lacking flexibility and prone to misjudgment when faced with new forms of harassment or complex call content; second, fixed response patterns result in a lack of natural flow in the conversation, failing to adapt to the actual progress of the call and the other party's emotions. Although some solutions incorporate virtual digital human technology, their response mechanisms are still primarily limited to switching pre-set scripts between different scenarios, lacking truly intelligent communication capabilities.

[0004] Therefore, existing technical solutions are still unable to fully meet the diverse needs of users in various scenarios where they are unable or unwilling to answer the phone, resulting in a poor actual user experience. Summary of the Invention

[0005] The technical problem to be solved by this invention is to address the aforementioned shortcomings of the prior art by proposing an external call processing method, telephone processing method, system, and electronic device. This external call processing method can improve the user's call processing efficiency, achieve refined and intelligent management of external calls, and thus enhance the user's communication experience.

[0006] In a first aspect, the present invention provides a method for handling external incoming calls based on a digital human assistant, the method comprising the following steps:

[0007] Receive incoming calls and extract the caller ID;

[0008] The caller ID number is matched against a pre-set database of numbers, and the incoming calls are categorized as whitelist numbers, blacklist numbers, or unknown numbers based on the matching results.

[0009] The decision to initiate a ringing signal to the user terminal is based on the classification results;

[0010] If the number is classified as a whitelist number, a ringing signal is sent to the user's terminal to prompt the user to answer. If the user answers, the external call is transferred to the user's terminal for a live call. If the user does not answer within the specified time limit, the digital human assistant is activated to answer the call and receives the caller's intent from the digital human assistant. The digital human assistant is a virtual intelligent service system based on artificial intelligence technology.

[0011] If the number is classified as a blacklist number, the incoming call will be blocked to terminate the call connection.

[0012] If the classification result is an unknown number, no ringing will be initiated to the user terminal. Instead, the digital human assistant will be activated to answer the call and receive the caller's intent as fed back by the digital human assistant.

[0013] Based on the caller intent fed back by the digital human assistant, it is determined whether the current call needs to be transferred to the user terminal. If it is determined that transfer is necessary, the external call is transferred to the user terminal for a human-assisted conversation. If it is determined that transfer is not necessary, the digital human assistant will provide a voice response based on a preset script library and the current call context until the call ends, thus completing the external call processing based on the digital human assistant.

[0014] Furthermore, the caller ID number is matched against a pre-set database, and based on the matching results, incoming calls are categorized as whitelisted numbers, blacklisted numbers, or unknown numbers. Specifically, this includes:

[0015] Standardize the calling number to generate a standardized number;

[0016] The standardized number is matched with the whitelist sub-library in the pre-set list database; if the match is successful, the external call is classified as a whitelist number.

[0017] If the standardized number does not match in the whitelist sub-database, it will be matched with the blacklist sub-database in the list database; if it matches in the blacklist sub-database, the incoming external call will be classified as a blacklist number.

[0018] If the standardized number fails to match in either the whitelist or the blacklist sub-database, the incoming call will be classified as an unknown number.

[0019] Furthermore, the digital human assistant provides voice responses based on a pre-set script library and the current call context, specifically including:

[0020] Configure the preset dialogue script library as a tree structure. Each node in the tree structure is set with at least one trigger keyword, the corresponding preset response content, and the response mode identifier. The response mode identifier is used to indicate whether to use recording mode or text-to-speech mode.

[0021] During the call, the other party's voice is converted into recognized text using speech recognition technology, and natural language processing technology is used to calculate the semantic similarity between the recognized text and the trigger keywords of each node.

[0022] When the highest semantic similarity value exceeds the first preset threshold, the corresponding node is selected as the response strategy, and the reply is executed according to the response mode identifier of the node; wherein, if the response mode identifier is recording mode, a pre-recorded human voice audio segment is played; if the response mode identifier is text-to-speech mode, the preset response content of the node is synthesized into speech output through the text-to-speech engine.

[0023] If the semantic similarity values ​​of all nodes do not exceed the first preset threshold, the large language model is invoked to generate dynamic reply text based on the recognized text and the current call context. The dynamic reply text is then synthesized into speech output through the speech engine to achieve a voice reply.

[0024] Furthermore, based on the caller intent fed back by the digital human assistant, it is determined whether the current call needs to be transferred to the user's terminal, specifically including:

[0025] The incoming call intent fed back by the digital human assistant is input into a pre-trained intent classification model; the intent classification model outputs the probability value that the incoming call intent belongs to a preset set of intents to be transferred.

[0026] If the probability value is greater than the second preset threshold, it is determined that the call needs to be transferred to the user terminal; otherwise, it is determined that no transfer is needed.

[0027] Furthermore, after the call ends, the method also includes:

[0028] The call log is pushed to the user's terminal, and the user is asked to mark any unknown numbers in the log.

[0029] Receive user marking of unknown numbers: If an unknown number is marked as an important contact or allowed to be answered, add the unknown number and its corresponding standardized format to the whitelist sub-library; if an unknown number is marked as a spam number or rejected, add the unknown number and its corresponding standardized format to the blacklist sub-library.

[0030] Furthermore, after the call ends, the method also includes scheduling in conjunction with the schedule management system;

[0031] The schedule includes:

[0032] When the digital assistant recognizes that the caller intends to make an appointment, check the schedule, or change the schedule, it calls the application programming interface of the schedule management system.

[0033] Use the application programming interface to query a user's available time, create new schedule entries, or modify existing schedule entries;

[0034] The results of the scheduled tasks are broadcast to the other party in the call using speech synthesis technology, thus completing the schedule arrangement.

[0035] Secondly, the present invention provides a telephone processing method based on a digital human assistant, the telephone processing method including external call processing and automatic outbound calling;

[0036] External call processing is performed according to the external call processing method based on the digital human assistant described in the first aspect;

[0037] Automated outbound calling, specifically including:

[0038] Monitor preset trigger conditions, including the arrival of a schedule reminder or the fulfillment of preset business node notification conditions;

[0039] When the triggering conditions are met, the digital human assistant is activated and an outbound call is initiated to the target number according to the preset outbound call task information;

[0040] After the outbound call is connected, the digital assistant will read out the preset notification or reminder content and interact with the called party via voice.

[0041] Furthermore, automated outbound calling includes a group calling mode;

[0042] The implementation of the group call mode specifically includes:

[0043] Receive a group call task instruction, which includes a list of target numbers and the corresponding broadcast content;

[0044] Control the digital human assistant to make outbound calls to numbers in the target number list sequentially or in parallel according to a preset strategy;

[0045] The call results for each number are recorded, and a statistical report is generated after the group call task is completed to realize the group call mode.

[0046] Thirdly, the present invention provides a telephone processing system based on a digital human assistant, the system comprising:

[0047] The incoming call receiving module is used to receive external calls and extract the caller ID of the external caller;

[0048] The number classification module, connected to the call receiving module, is used to match the calling number with a preset list database and classify external calls into whitelist numbers, blacklist numbers, or unknown numbers based on the matching results.

[0049] The call routing module, connected to the number classification module, determines whether to initiate a ringing signal to the user terminal based on the classification results. Specifically, when an incoming call is classified as a whitelisted number, a ringing signal is initiated to the user terminal to prompt the user to answer. If the user answers, the incoming call is transferred to the user terminal for a live call. If the user does not answer within a set time limit, the digital human interaction control module is triggered. When an incoming call is classified as a blacklisted number, the incoming call is blocked to terminate the call connection. When an incoming call is classified as an unknown number, no ringing signal is initiated to the user terminal; instead, the digital human assistant is activated to answer the call and receives the caller's intent from the digital human assistant.

[0050] The digital human interaction control module communicates with the call routing module and the external digital human assistant, which is a virtual intelligent service system based on artificial intelligence technology.

[0051] Among them, the digital human interaction control module is used to respond to the trigger of the call routing module, and is also used to start an interactive session with the digital human assistant to answer the call on behalf of the user; it is also used to obtain the call intention fed back by the digital human assistant during the interaction; and it is also used to determine whether the current call needs to be transferred to the user terminal based on the call intention.

[0052] If the digital human interaction control module determines that a call needs to be transferred, it triggers the call routing module to transfer the external call to the user terminal for a human-assisted conversation. If the digital human interaction control module determines that a call does not need to be transferred, it instructs the digital human assistant to provide a voice response based on a preset script library and the current call context until the call ends, thus completing the external call processing based on the digital human assistant.

[0053] Fourthly, the present invention provides an electronic device including a memory and a processor. The memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the external call processing method based on a digital human assistant according to the first aspect, or executes the telephone processing method based on a digital human assistant according to the second aspect.

[0054] This invention introduces an AI-based digital human assistant and combines it with an intelligent caller ID classification mechanism (whitelist, blacklist, unknown numbers) to differentiate incoming calls, significantly improving the efficiency of external call processing and user experience, while effectively preventing interference from harassing or fraudulent calls. Specific beneficial effects are as follows:

[0055] (1) Improve user connection efficiency and response timeliness: For whitelisted numbers, the digital human assistant will be automatically activated to answer the call when the user fails to answer in time, so as to avoid missing important calls; for unknown numbers, the digital human will identify the intent first and transfer the call to a human only when necessary, so as to reduce unnecessary disturbance to the user.

[0056] (2) Enhance communication security and privacy protection: Automatically block high-risk or known harassing numbers through the blacklist mechanism to effectively prevent malicious calls and ensure the security of the user's communication environment.

[0057] (3) Reduce human burden and operating costs: The digital human assistant can automatically answer unknown or non-urgent calls around the clock and perform intelligent interaction based on a preset script library and contextual understanding, greatly reducing the frequency of human customer service intervention. It is suitable for various scenarios such as enterprise front desk and personal assistant.

[0058] (4) Enhance intelligent interactive experience: The digital human assistant relies on natural language processing and speech synthesis technology to achieve human-like voice dialogue, making the call processing process more natural and smooth, and enhancing user satisfaction.

[0059] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0060] 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. The above and other features and advantages will become more apparent to those skilled in the art from the detailed description of exemplary embodiments with reference to the accompanying drawings, in which:

[0061] Figure 1 A flowchart of an external call processing method based on a digital human assistant provided in an embodiment of the present invention;

[0062] Figure 2 The following is an implementation architecture diagram of a digital human assistant provided in an embodiment of the present invention;

[0063] Figure 3 This is a schematic diagram of the human-computer dialogue process provided in an embodiment of the present invention;

[0064] Figure 4 This is a schematic diagram of a dialogue response provided in an embodiment of the present invention;

[0065] Figure 5 The intelligent call service process provided in the embodiments of the present invention;

[0066] Figure 6 A schematic diagram of a telephone processing system based on a digital human assistant provided in an embodiment of the present invention;

[0067] Figure 7 This is an architectural diagram of an electronic device provided in an embodiment of the present invention.

[0068] Reference numerals: 10, incoming call receiving module; 20, number classification module; 30, call routing module; 40, digital human interaction control module; 50, record storage module; 100, processor; 200, memory. Detailed Implementation

[0069] It is understood that the specific embodiments and accompanying drawings described herein are merely for explaining the invention and are not intended to limit the invention.

[0070] It is understood that, without conflict, the various embodiments and features in the embodiments of the present invention can be combined with each other.

[0071] It is understood that, for ease of description, only the parts related to the present invention are shown in the accompanying drawings, while the parts unrelated to the present invention are not shown in the drawings.

[0072] It is understood that each unit or module involved in the embodiments of the present invention may correspond to only one entity structure, or may be composed of multiple entity structures, or multiple units or modules may be integrated into one entity structure.

[0073] It is understood that, without conflict, the functions and steps marked in the flowcharts and block diagrams of this invention may occur in a different order than that marked in the accompanying drawings.

[0074] It is understood that the flowcharts and block diagrams of this invention illustrate the possible architecture, functions, and operations of systems, apparatuses, devices, and methods according to various embodiments of this invention. Each block in the flowchart or block diagram may represent a unit, module, program segment, or code, containing executable instructions for implementing the specified function. Furthermore, each block or combination of blocks in the block diagram and flowchart can be implemented using a hardware-based system to achieve the specified function, or using a combination of hardware and computer instructions.

[0075] It is understood that the units and modules involved in the embodiments of the present invention can be implemented by software or by hardware. For example, the units and modules can be located in a processor.

[0076] Example 1:

[0077] With the widespread adoption of communication technologies and smart terminals, users are receiving an increasing number of calls daily, often plagued by spam calls, telemarketing, and other unwanted calls. This not only wastes time but also poses information security risks. To address this, a digital human assistant-based external call handling method has emerged. This method can automatically identify, classify, and intelligently answer incoming calls, significantly improving the call experience and meeting the needs of modern communication management. This technology is widely used in enterprise customer service centers to improve call routing and processing efficiency; in the financial and insurance industry to protect customer information security and prevent fraud; for individual users to block spam and fraudulent calls and easily answer calls from delivery and takeout services; and in smart office environments to help managers efficiently manage business calls, reduce interference, and optimize workflows.

[0078] like Figure 1 As shown, the external call handling method based on a digital human assistant provided in this embodiment specifically includes the following steps:

[0079] Step 1: Receive incoming calls and extract the caller ID.

[0080] Step 2: Match the calling number with the preset list database, and classify the incoming call as a whitelist number, blacklist number, or unknown number based on the matching results.

[0081] Step 3: Determine whether to initiate a ringing signal to the user terminal based on the classification results.

[0082] If the number is classified as a whitelist number, a ringtone is initiated to the user's terminal to prompt the user to answer. If the user answers, the call is transferred to the user's terminal for a live call. If the user does not answer within a specified time limit, a digital human assistant is activated to answer the call and receive the caller's intent from the digital human assistant. The digital human assistant is a virtual intelligent service system based on artificial intelligence technology. If the number is classified as a blacklist number, the call is blocked to terminate the call connection. If the number is classified as an unknown number, no ringtone is initiated to the user's terminal; instead, the digital human assistant is activated to answer the call and receives the caller's intent from the digital human assistant.

[0083] As a specific implementation method, the calling number is matched with a pre-set list database, and the incoming calls are categorized into whitelist numbers, blacklist numbers, or unknown numbers based on the matching results. Specifically, this includes:

[0084] Standardize the calling number to generate a standardized number;

[0085] The standardized number is matched with the whitelist sub-library in the pre-set list database; if the match is successful, the external call is classified as a whitelist number.

[0086] If the standardized number does not match in the whitelist sub-database, it will be matched with the blacklist sub-database in the list database; if it matches in the blacklist sub-database, the incoming external call will be classified as a blacklist number.

[0087] If the standardized number fails to match in either the whitelist or the blacklist sub-database, the incoming call will be classified as an unknown number.

[0088] In practical applications, incoming calls are identified and processed through pre-set permission tags, enabling flexible and controllable call management. Specifically, when an incoming number is on the "whitelist," its corresponding permission tag is marked "Allow to answer," and the call is connected normally. Conversely, if an incoming number is on the "blacklist," its permission tag is marked "Disallow to answer," automatically triggering blocking or rejection. This mechanism ensures that important contacts are always accessible while effectively preventing interference from irrelevant or nuisance calls, making communication management clearer, more efficient, and more aligned with actual needs.

[0089] Step 4: Based on the caller intent fed back by the digital human assistant, determine whether the current call needs to be transferred to the user terminal; if it is determined that a transfer is needed, the external call will be transferred to the user terminal for a human-assisted conversation; if it is determined that no transfer is needed, the digital human assistant will provide a voice response based on a preset script library and the current call context until the call ends.

[0090] As a specific implementation method, the digital human assistant provides voice responses based on a preset script library and the current call context, specifically including:

[0091] Configure the preset dialogue script library as a tree structure. Each node in the tree structure is set with at least one trigger keyword, the corresponding preset response content, and the response mode identifier. The response mode identifier is used to indicate whether to use recording mode or text-to-speech mode.

[0092] During the call, the other party's voice is converted into recognized text using speech recognition technology, and natural language processing technology is used to calculate the semantic similarity between the recognized text and the trigger keywords of each node.

[0093] When the highest semantic similarity value exceeds the first preset threshold, the corresponding node is selected as the response strategy, and the reply is executed according to the response mode identifier of the node; wherein, if the response mode identifier is recording mode, a pre-recorded human voice audio segment is played; if the response mode identifier is text-to-speech mode, the preset response content of the node is synthesized into speech output through the text-to-speech engine.

[0094] If the semantic similarity values ​​of all nodes do not exceed the first preset threshold, the large language model is invoked to generate dynamic reply text based on the recognized text and the current call context. The dynamic reply text is then synthesized into speech output through the speech engine to achieve a voice reply.

[0095] As a specific implementation method, based on the caller intent fed back by the digital human assistant, it is determined whether the current call needs to be transferred to the user terminal, specifically including:

[0096] The incoming call intent fed back by the digital human assistant is input into a pre-trained intent classification model; the intent classification model outputs the probability value that the incoming call intent belongs to a preset set of intents to be transferred.

[0097] If the probability value is greater than the second preset threshold, it is determined that the call needs to be transferred to the user terminal; otherwise, it is determined that no transfer is needed.

[0098] As a preferred approach, step 4 is followed by step 5.

[0099] Step 5: After the call ends, save the recording of this call and the corresponding text record.

[0100] As one specific implementation method, after the call ends, the method further includes:

[0101] The call log is pushed to the user's terminal, and the user is asked to mark any unknown numbers in the log.

[0102] Receive user marking of unknown numbers: If an unknown number is marked as an important contact or allowed to be answered, add the unknown number and its corresponding standardized format to the whitelist sub-library; if an unknown number is marked as a spam number or rejected, add the unknown number and its corresponding standardized format to the blacklist sub-library.

[0103] As another specific implementation, after the call ends, the method also includes a schedule linked to the schedule management system;

[0104] The schedule includes:

[0105] When the digital assistant recognizes that the caller intends to make an appointment, check the schedule, or change the schedule, it calls the application programming interface of the schedule management system.

[0106] Use the application programming interface to query a user's available time, create new schedule entries, or modify existing schedule entries;

[0107] The results of the scheduled tasks are broadcast to the other party in the call using speech synthesis technology, thus completing the schedule arrangement.

[0108] This embodiment implements several intelligent functions, comprehensively improving the efficiency of telephone management and life services. First, it possesses intelligent identification and blocking capabilities for harassing calls. Based on big data identification, it can directly block known harassing numbers. For numbers that are missed, an AI robot automatically answers the call, analyzes the call content to identify it as a harassing call, and then blocks it again, updating the relevant numbers to the harassing phone number database for continuous optimization. Second, it supports intelligent call answering. When the phone owner is unable to answer, an intelligent call assistant device and system can answer on their behalf, saving the entire call log for later retrieval and automatically converting the content into text for more efficient and convenient information retrieval. Third, it provides a preset group calling function, automatically dialing pre-defined groups according to pre-set scripts, significantly improving the efficiency and consistency of group calling operations. Fourth, it also has a schedule reminder function, acting as a smart secretary for the phone owner, assisting in scheduling and providing intelligent reminders at designated times to avoid forgetting important matters. Fifth, it achieves intelligent relationship maintenance, automatically calling key contacts in the address book on important holidays to send care or greetings, helping to maintain good interpersonal relationships. Sixth, it supports intelligent alarms and information notifications. By connecting with third-party application software interfaces, it can realize automatic alarms and message pushes in multiple scenarios. For example, it can automatically alarm when a traffic accident occurs by linking with vehicle devices, issue an alarm when health diagnosis software detects health abnormalities, provide flight status notifications in sync with flight software, and push promotional information by connecting with e-commerce platforms. This provides users with timely and thoughtful intelligent reminders and services in many aspects of life, travel, health, and consumption.

[0109] The architecture of this embodiment is as follows: Figure 2 As shown, it specifically includes the following 10 aspects:

[0110] 1. Intent Recognition Module:

[0111] The intent recognition module primarily identifies caller intent through two dimensions. First, it collects a massive database of spam phone numbers; once a match is found, it's marked as a spam call, allowing the user to choose to reject the call or decline with a short message. Second, it trains a private call pattern database, divided into several sub-databases such as spam call database, family and friend call database, and work call database, each corresponding to a specific knowledge base. Based on the collected call records, a cosine similarity algorithm is used to match the calls against these databases; higher similarity results indicate the corresponding type. Subsequently, based on user-manual labeling and corrections, machine learning is used to continuously optimize the call knowledge base, while simultaneously leveraging the capabilities of a large underlying model for intent judgment and outputting turn-based dialogue content.

[0112] 2. Call Management Module:

[0113] The call management module stores the audio and text records of each call for easy review. To save phone space, different compression algorithms are used to compress the audio and text files separately, achieving better recognition rates and compression ratios. Furthermore, this module indexes and segments the phone numbers and text records to improve retrieval efficiency and accuracy, ensuring users can quickly find the call records they need.

[0114] 3. Data storage module:

[0115] The data storage module is responsible for storing various information data required for system operation, including a blacklist of nuisance calls, call file data, call intent identification data, schedule information, and external access information. This module employs an efficient data management mechanism to ensure secure storage and rapid access to all types of data, providing data support for the stable operation of the entire system.

[0116] 4. Human-computer dialogue module:

[0117] The human-computer dialogue module enables the robot to converse with a real person based on a scripted dialogue. This invention primarily utilizes two modes for dialogue: recording mode and TTS mode. Recording mode uses pre-recorded audio, resulting in a more realistic, indistinguishable sound from a human voice. TTS mode automatically converts pre-recorded text into audio, offering greater flexibility and enabling more complex scenarios. The dialogue script design employs a tree structure, customizing specific keywords and responses for each inbound or outbound call scenario. After the speech recognition module translates the text, it calculates the similarity to the keywords and identifies the corresponding response.

[0118] 5. Voice intelligence module:

[0119] The intelligent voice module uses big data and deep learning technologies to convert the voice of different customers into text, achieving a one-to-one conversion from voice to text. The intelligent call module generates and sends the intelligent voice communication with the caller to the communication module. The human operator call function integrates with the system's built-in communication module; after confirming that a human operator is needed, this module transfers the call to achieve end-to-end voice communication, ensuring smooth and clear calls.

[0120] 6. Artificial Intelligence Module:

[0121] The artificial intelligence module comprises three core components: machine learning, a knowledge base, and a knowledge graph. Through deep learning technology, combined with knowledge bases, knowledge graphs, and large-scale modeling, it enables more intelligent and seamless human-computer dialogue, meeting the needs of various upper-level application scenarios. This module can continuously learn and optimize, improving the system's intelligence level and user experience.

[0122] 7. Large Model Module:

[0123] The large model module provides underlying capabilities for large models and enables the invocation of these capabilities through interfaces. As the core AI engine of the entire system, the large model provides powerful semantic understanding and generation capabilities for modules such as intent recognition, human-computer dialogue, and voice intelligence, ensuring that the system can handle complex voice interaction scenarios.

[0124] 8. Communication Module:

[0125] The communication module is primarily responsible for real-time monitoring of incoming calls. Based on call information, it enables the mobile device to automatically take over the call, conduct human-machine dialogue, or ring the phone and transfer the call to a human operator. Simultaneously, this module also accepts outbound call requests from upper-layer applications, maintains call status, and switches between human operators and human operators / bots according to upper-layer instructions, ensuring communication stability and flexibility.

[0126] 9. Application Service Module:

[0127] The application services module supports various services such as intelligent identification and blocking of nuisance calls, intelligent call answering, pre-set mass calling, schedule reminders, intelligent relationship maintenance, and intelligent alarms and information notifications. As the core business layer of the system, this module integrates the capabilities of various technical modules to provide users with a complete intelligent telephone management solution.

[0128] 10. External Interface Layer:

[0129] The external interface layer is responsible for connecting with third-party applications to expand the system's application scenarios. For example, it can connect to in-vehicle devices for automatic alerts in case of traffic accidents; connect to health diagnostic software for automatic alerts in case of health problems; connect to flight tracking software for flight status notifications; and connect to e-commerce platforms for promotional notifications, etc. Through open interfaces, the system can seamlessly integrate with various external services, providing a richer intelligent service experience.

[0130] The business process of this embodiment is described in detail below:

[0131] 1) When an external call comes in, it first enters the intent recognition module for processing, and then follows different business logic based on the number.

[0132] 2) If it is a blacklisted number, reject the call directly.

[0133] 3) If it is a whitelisted number, a ringtone will be triggered to notify the user to answer the call.

[0134] 4) If the user connects, a live conversation will take place.

[0135] 5) If the user does not answer, rejects, or the line is busy, the robot assistant will answer the call, conduct a human-computer dialogue, and save the dialogue record for the user to view later.

[0136] 6) If the number is unknown, the robot assistant will answer the call and engage in human-machine dialogue.

[0137] 7) Based on the content of the conversation, determine the intention and whether to transfer to a human operator.

[0138] 8) If the robot determines that it needs to be transferred to a human operator, then the call will be transferred to a human operator.

[0139] 9) If the user refuses to answer, misses the call, or the line is busy, the robot assistant will answer the call.

[0140] 10) End the call. All call recordings and text messages can be viewed.

[0141] Human-computer dialogue process as follows Figure 3 As shown, the specific human-computer dialogue process is as follows:

[0142] (1): Voice input and text conversion.

[0143] After the human-computer dialogue begins, the system first performs sentence segmentation, dividing the user's continuous voice input into independent sentence units, and then converting the voice signal into text format using speech recognition technology. This step is fundamental to the entire dialogue process, ensuring that the system can accurately understand the user's language expression.

[0144] (2): Semantic understanding and intent recognition.

[0145] The system invokes the artificial intelligence module to perform deep semantic analysis and intent recognition on the converted text content. Through natural language processing technology, the system parses key information in the user's speech, determines the user's true intent and needs, and generates corresponding response strategies based on the recognition results. This step is the core of human-computer dialogue, determining the accuracy and relevance of subsequent responses.

[0146] (3): Speech synthesis processing.

[0147] Based on the system-generated response, the speech synthesis module is activated to convert the text-based response into a natural and fluent speech signal. The system generates speech output that meets the needs of the scenario based on preset speech feature parameters, ensuring that the machine's response sounds natural and is easy to understand.

[0148] (4): Voice broadcast output.

[0149] The synthesized voice signal is transmitted to the user's device via the communication module for playback. The user can hear the system's generated response voice, enabling human-machine voice interaction. This step completes the closed loop from machine processing to user perception, ensuring that the user receives feedback information from the system.

[0150] Step 5: Save the conversation history.

[0151] The system automatically saves the complete record of the current conversation, including the original audio recording and the corresponding text transcription. These records are stored in the data storage module, used for subsequent conversation quality analysis and optimization, and also provide users with query and review functionality.

[0152] Step 6: Looping dialogue mechanism.

[0153] After completing one round of dialogue, the system automatically enters the preparation state for the next round of dialogue, and the process returns to the first step, waiting for the user's next voice input. This loop mechanism ensures the continuity and sustainability of human-computer dialogue, supports multi-round complex dialogue scenarios, until the user actively ends the dialogue or specific exit conditions are met.

[0154] Dialogue response process as follows Figure 4 As shown, the specific steps include the following:

[0155] (1) Receive call text content.

[0156] The dialogue response process begins with receiving the call text content; the system acquires the text information converted from the user's speech through speech recognition. This step, as the input to the process, is responsible for collecting the user's verbal expressions, providing raw data for subsequent intent understanding. The system performs preliminary formatting and preprocessing on the received text to ensure the integrity and readability of the text content.

[0157] (2) Understanding and judging intent.

[0158] The system invokes the artificial intelligence module to perform deep semantic analysis on the received text content, identifying the user's true intent through natural language processing technology. This step includes sub-processes such as semantic parsing, entity extraction, and intent classification, enabling the system to accurately determine the core needs and purposes the user wants to express. The accuracy of intent understanding directly determines the quality and relevance of subsequent responses.

[0159] (3) Knowledge generation in the knowledge base.

[0160] Based on the identified user intent, the system accesses a knowledge base to generate and retrieve corresponding knowledge. The knowledge base contains pre-built structured information and domain knowledge, allowing the system to match relevant knowledge entries according to the intent type and generate response content that meets the user's needs. This step ensures the accuracy and professionalism of the response.

[0161] (4) Output response content

[0162] Finally, the generated knowledge content is transformed into natural and fluent response text and delivered to the user via the communication module. The response content is personalized based on the dialogue context and user characteristics to ensure the answer is targeted and user-friendly. The output response can be in text form or converted into speech through speech synthesis.

[0163] This embodiment proposes an external call handling method based on a digital human assistant. Through automatic identification, classification, and intelligent response, it significantly improves call management efficiency and can be widely applied in scenarios such as enterprise customer service, financial insurance, personal anti-harassment, and smart office. The system process includes receiving incoming calls, number matching and classification, intelligent response, and transfer judgment, and supports extended functions such as recording and text record storage, user tagging, and schedule management. In its implementation, speech recognition, natural language processing, knowledge bases, and large-scale model technologies are used for semantic understanding and intelligent dialogue. These technologies, combined with multi-layered modules, make telephone interaction more intelligent, efficient, and user-friendly. This solution not only effectively blocks harassing calls and enhances user experience but also provides diverse services such as scheduling and information reminders, meeting the multiple needs of modern intelligent communication.

[0164] Example 2:

[0165] This embodiment provides a telephone processing method based on a digital human assistant, which includes external call processing and automatic outbound calling;

[0166] External call processing is performed according to the external call processing method based on digital human assistant described in Example 1;

[0167] Automated outbound calling, specifically including:

[0168] Monitor preset trigger conditions, including the arrival of a schedule reminder or the fulfillment of preset business node notification conditions;

[0169] When the triggering conditions are met, the digital human assistant is activated and an outbound call is initiated to the target number according to the preset outbound call task information;

[0170] After the outbound call is connected, the digital assistant will read out the preset notification or reminder content and interact with the called party via voice.

[0171] As a specific implementation method, automatic outbound calling includes a group calling mode;

[0172] The implementation of the group call mode specifically includes:

[0173] Receive a group call task instruction, which includes a list of target numbers and the corresponding broadcast content;

[0174] Control the digital human assistant to make outbound calls to numbers in the target number list sequentially or in parallel according to a preset strategy;

[0175] The call results for each number are recorded, and a statistical report is generated after the group call task is completed to realize the group call mode.

[0176] The automated outbound calling in this embodiment can also be called intelligent calling service. The intelligent calling service process is a complete automated outbound calling processing system. Through steps such as task creation, call execution, intelligent dialogue, and manual transfer, it achieves efficient outbound calling service management. Its specific process is as follows: Figure 5 As shown, the specific steps include the following:

[0177] Step 1: Create a smart outbound call task:

[0178] The process begins with creating an intelligent outbound calling task. The system requires setting key parameters including the outbound call script template, the target outbound number list, and the outbound call time window. The outbound call script design uses a tree structure, including trigger keywords, preset response content, and response mode identifiers to ensure the relevance and effectiveness of the dialogue. Simultaneously, the system configures outbound call timing rules based on business needs to avoid disturbing customers at inappropriate times.

[0179] Step 2: Initiate the outbound calling task:

[0180] Once the task is created, the system initiates the outbound call task and enters the task execution preparation phase. At this time, the system checks the completeness of the task configuration, verifies the validity of the outbound call number, and allocates the corresponding communication resources and AI processing capabilities to fully prepare for subsequent call execution.

[0181] Step 3: Perform outbound call tasks one by one:

[0182] The system executes outbound calls one by one according to a preset time window and number list. Each call is automatically initiated through the communication module to establish a communication connection with the target customer. During the call, the system monitors the connection status in real time to ensure successful call establishment.

[0183] Step 4: Conduct human-computer interaction:

[0184] Once the call is connected, the system activates the human-computer dialogue module to engage in intelligent interaction with the customer. During the conversation, the system uses speech recognition technology to convert the customer's voice into text, employs natural language processing technology to understand the customer's intent, and generates corresponding responses based on a pre-set script library or a large model. The system supports both recorded and TTS (Text-to-Speech) response modes to ensure a natural and fluent dialogue.

[0185] Step 5: Determine if it is necessary to transfer to a human operator.

[0186] During the conversation, the system analyzes the content and customer reactions in real time, using an intent classification model to determine whether to transfer the customer to a human agent. The system considers multiple factors, including customer emotions, problem complexity, and business type. When it determines that the customer's needs exceed the AI's processing capabilities or the customer explicitly requests human assistance, it triggers the human agent transfer mechanism.

[0187] Step 6: Branching:

[0188] Based on the judgment result, the process is divided into two branches:

[0189] No need to transfer to a human: If the system determines that a transfer to a human is unnecessary, the digital assistant will continue the human-machine dialogue until all business objectives are achieved or the customer voluntarily ends the call. During the dialogue, the system will continuously record the interaction content and adjust its script strategy based on the conversation's progress.

[0190] If a transfer to a human agent is required, the system will immediately transfer the call to a human agent and simultaneously transmit the previous conversation record and customer information to ensure that the human agent can seamlessly continue the conversation and provide more professional service.

[0191] Step 7: End of call and next task:

[0192] Whether the AI ​​completes the conversation or a human agent ends the service, the system records the entire call process, including audio and text recordings. After the call ends, the system automatically checks the task list for any unfinished outbound calls. If so, it returns to step three to continue executing the next call until all tasks are completed.

[0193] Example 3:

[0194] like Figure 6 As shown, the present invention provides a telephone processing system based on a digital human assistant, the system comprising:

[0195] The incoming call receiving module 10 is used to receive external incoming calls and extract the caller ID of the external incoming call;

[0196] The number classification module 20 is connected to the incoming call receiving module 10 and is used to match the calling number with a preset list database and classify external incoming calls into whitelist numbers, blacklist numbers or unknown numbers based on the matching results.

[0197] The call routing module 30, connected to the number classification module 20, is used to determine whether to initiate a ringing to the user terminal based on the classification result. Specifically, when an incoming call is classified as a whitelisted number, a ringing is initiated to the user terminal to prompt the user to answer. If the user answers, the incoming call is transferred to the user terminal for a live call. If the user does not answer within a set time limit, the digital human interaction control module is triggered. When an incoming call is classified as a blacklisted number, the incoming call is blocked to terminate the call connection. When an incoming call is classified as an unknown number, no ringing is initiated to the user terminal, and the digital human assistant is activated to perform call answering and receive the caller's intent from the digital human assistant.

[0198] The digital human interaction control module 40 is communicatively connected to the call routing module 30 and an external digital human assistant, wherein the digital human assistant is a virtual intelligent service system based on artificial intelligence technology.

[0199] The digital human interaction control module 40 is used to respond to the triggering of the call routing module, and is also used to start an interaction session with the digital human assistant to answer the call; it is also used to obtain the call intention fed back by the digital human assistant during the interaction; and it is also used to determine whether the current call needs to be transferred to the user terminal based on the call intention.

[0200] If the digital human interaction control module 40 determines that a call needs to be transferred, it triggers the call routing module to transfer the external call to the user terminal for a human-assisted call. If the digital human interaction control module 40 determines that a call does not need to be transferred, it instructs the digital human assistant to provide a voice response based on a preset script library and the current call context until the call ends, thus completing the external call processing based on the digital human assistant.

[0201] The system in this embodiment is capable of executing the method in Embodiment 1.

[0202] Example 4:

[0203] like Figure 7 As shown, this embodiment provides an electronic device, which includes a memory 200 and a processor 100. The memory 200 stores a computer program. When the processor 100 runs the computer program stored in the memory 200, the processor 100 executes the external call processing method based on a digital human assistant as described in Embodiment 1, or executes the telephone processing method based on a digital human assistant as described in Embodiment 2.

[0204] It is understood that the above embodiments are merely exemplary implementations used to illustrate the principles of the present invention, and the present invention is not limited thereto. For those skilled in the art, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also considered to be within the scope of protection of the present invention.

Claims

1. A method for handling incoming external calls based on a digital human assistant, characterized in that, The method includes the following steps: Receive incoming external calls and extract the caller ID of the incoming external calls; The calling number is matched with a preset list database, and the external call is classified as a whitelist number, a blacklist number, or an unknown number based on the matching result. The decision to initiate a ringing signal to the user terminal is based on the classification results; If the classification result is a whitelist number, a ringing signal is sent to the user terminal to prompt the user to answer; if the user answers, the external call is transferred to the user terminal for a live call; if the user does not answer within the specified time limit, a digital human assistant is activated to answer the call and receives the caller's intent from the digital human assistant; wherein, the digital human assistant is a virtual intelligent service system based on artificial intelligence technology. If the classification result is a blacklist number, the external call will be blocked to terminate the call connection; If the classification result is an unknown number, no ringing is initiated to the user terminal, the digital human assistant is activated to answer the call, and the caller's intent is received from the digital human assistant. Based on the caller intent fed back by the digital human assistant, it is determined whether the current call needs to be transferred to the user terminal. If it is determined that transfer is necessary, the external call is transferred to the user terminal for a human-assisted conversation. If it is determined that transfer is not necessary, the digital human assistant provides a voice response based on a preset script library and the current call context until the call ends, thus completing the external call processing based on the digital human assistant.

2. The external call processing method based on a digital human assistant according to claim 1, characterized in that, The step of matching the calling number with a pre-set list database and classifying the incoming external call as a whitelist number, blacklist number, or unknown number based on the matching result specifically includes: The calling number is processed to standardize its format, generating a standardized number; The standardized number is matched with the whitelist sub-library in the pre-set list database; if the match is successful, the external call is classified as a whitelist number. If the standardized number does not match successfully in the whitelist sub-database, then the standardized number is matched with the blacklist sub-database in the list database; if it matches successfully in the blacklist sub-database, then the external call is classified as a blacklist number. If the standardized number fails to match in either the whitelist or the blacklist sub-database, the incoming external call will be classified as an unknown number.

3. The external call processing method based on a digital human assistant according to claim 1, characterized in that, The digital human assistant provides voice responses based on a preset dialogue script library and the current call context, specifically including: The preset dialogue script library is configured as a tree structure, and each node in the tree structure is set with at least one trigger keyword, a corresponding preset response content, and a response mode identifier; wherein, the response mode identifier is used to indicate whether to use recording mode or text-to-speech mode. During the call, the other party's voice is converted into recognized text using speech recognition technology, and natural language processing technology is used to calculate the semantic similarity between the recognized text and the trigger keywords of each node. When the highest semantic similarity value exceeds the first preset threshold, the corresponding node is selected as the response strategy, and the reply is executed according to the response mode identifier of the node; wherein, if the response mode identifier is recording mode, a pre-recorded human voice audio segment is played; if the response mode identifier is text-to-speech mode, the preset response content of the node is synthesized into speech output through the text-to-speech engine. When the semantic similarity values ​​of all nodes do not exceed the first preset threshold, the large language model is invoked to generate dynamic reply text based on the recognized text and the current call context. The dynamic reply text is then synthesized into speech output through the speech engine to achieve a voice reply.

4. The external call processing method based on a digital human assistant according to claim 1, characterized in that, The step of determining whether to transfer the current call to the user terminal based on the caller intent fed back by the digital human assistant specifically includes: The incoming call intent fed back by the digital human assistant is input into a pre-trained intent classification model; the intent classification model outputs the probability value that the incoming call intent belongs to a preset set of intents to be transferred. If the probability value is greater than the second preset threshold, it is determined that the call needs to be transferred to the user terminal; otherwise, it is determined that no transfer is needed.

5. The external call processing method based on a digital human assistant according to claim 1, characterized in that, After the call ends, the method further includes: The call log is pushed to the user's terminal, and the user is asked to mark any unknown numbers in the log. Receive user marking operations for the unknown number: if the unknown number is marked as an important contact or allowed to be answered, then add the unknown number and its corresponding standardized format to the whitelist sub-library; if the unknown number is marked as a harassing number or rejected, then add the unknown number and its corresponding standardized format to the blacklist sub-library.

6. The external call processing method based on a digital human assistant according to claim 1, characterized in that, After the call ends, the method also includes scheduling in conjunction with the schedule management system; The schedule specifically includes: When the digital assistant recognizes that the caller's intention is to make an appointment, check the schedule, or change the schedule, it calls the application programming interface of the schedule management system. Users can use the application programming interface to query their available time, create new schedule entries, or modify existing schedule entries. The results of the scheduled tasks are broadcast to the other party in the call using speech synthesis technology, thus completing the schedule arrangement.

7. A telephone processing method based on a digital human assistant, characterized in that, The method includes external call processing and automatic outbound calling; The external call processing is performed according to any one of claims 1 to 6; The automatic outbound calling specifically includes: Monitor preset trigger conditions, including the arrival of a schedule reminder time or the fulfillment of a preset business node notification condition; When the triggering condition is met, the digital human assistant is activated and initiates an outbound call to the target number according to the preset outbound call task information; After the outbound call is connected, the digital human assistant will broadcast the preset notification or reminder content and interact with the called party via voice.

8. The telephone processing method based on a digital human assistant according to claim 7, characterized in that, The automatic outbound calling includes a group calling mode; The implementation of the group call mode specifically includes: Receive a group call task instruction, the group call task instruction containing a target number list and corresponding broadcast content; The digital human assistant is controlled to make outbound calls to numbers in the target number list sequentially or in parallel according to a preset strategy; The call results for each number are recorded, and a statistical report is generated after the group call task is completed to realize the group call mode.

9. A telephone processing system based on a digital human assistant, characterized in that, The system includes: The incoming call receiving module is used to receive external incoming calls and extract the caller ID of the external incoming call; The number classification module, connected to the incoming call receiving module, is used to match the calling number with a preset list database, and classify the external incoming call into a whitelist number, a blacklist number, or an unknown number based on the matching result; The call routing module, connected to the number classification module, is used to determine whether to initiate a ringing signal to the user terminal based on the classification result. Specifically, when an incoming call is classified as a whitelisted number, a ringing signal is initiated to the user terminal to prompt the user to answer. If the user answers, the incoming call is transferred to the user terminal for a live call. If the user does not answer within a set time limit, the digital human interaction control module is triggered. When an incoming call is classified as a blacklisted number, the incoming call is blocked to terminate the call connection. When an incoming call is classified as an unknown number, no ringing signal is initiated to the user terminal; instead, the digital human assistant is activated to handle the call and receives the caller's intent from the digital human assistant. The digital human interaction control module is communicatively connected to the call routing module and an external digital human assistant, wherein the digital human assistant is a virtual intelligent service system based on artificial intelligence technology. The digital human interaction control module is used to respond to the triggering of the call routing module, and is also used to initiate an interaction session with the digital human assistant to answer the call on behalf of the user; it is also used to obtain the caller intent fed back by the digital human assistant during the interaction; and it is also used to determine whether the current call needs to be transferred to the user terminal based on the caller intent. If it is determined that a transfer is needed, the call routing module is triggered to transfer the external call to the user terminal for a human operator; if it is determined that no transfer is needed, the digital human assistant is instructed to provide a voice response based on a preset script library and the current call context until the call ends, thus completing the external call processing based on the digital human assistant.

10. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program. When the processor runs the computer program stored in the memory, the processor executes the external call processing method based on a digital human assistant according to any one of claims 1 to 6, or executes the telephone processing method based on a digital human assistant according to any one of claims 7 to 8.

Images