A progressive cooperative outbound call method and device, electronic equipment and storage medium

Abstract

The application discloses a progressive cooperative outbound call method and device, electronic equipment and storage medium, and relates to the technical field of intelligent outbound call. The method comprises the following steps: acquiring an outbound call task and user information of a first to-be-called user corresponding to the outbound call task; based on the outbound call task and the user information of the first to-be-called user, performing a first round of TTS outbound call on the first to-be-called user, and collecting a corresponding first outbound call result; determining a second to-be-called user from the first to-be-called user according to the corresponding first outbound call result; performing a second round of robot outbound call on the second to-be-called user, and collecting a corresponding second outbound call result; determining a third to-be-called user from the second to-be-called user according to the corresponding second outbound call result; and performing a third round of manual outbound call processing on the third to-be-called user by an artificial agent. According to the application, different users are processed in different outbound call stages according to actual outbound call results, so that the outbound call efficiency, artificial resource utilization rate and outbound call task completion rate are improved.

Description

Technical Field

[0001] This application relates to the field of intelligent outbound calling technology, and in particular to a progressive collaborative outbound calling method, apparatus, electronic device and storage medium. Background Technology

[0002] Outbound calling systems are widely used in scenarios such as finance, telecommunications, e-commerce, government services, customer follow-up, business notifications, and marketing outreach to send telephone notifications to users, remind them of business, confirm information, collect intentions, or conduct manual follow-ups.

[0003] In current technology, common outbound calling methods are usually executed using a single outbound calling method. Specifically, all outbound calls are made by human agents, all outbound calls are made by outbound calling robots, or all outbound calls are made through TTS voice broadcast.

[0004] Therefore, it can be seen that the current technology of using a single outbound calling method to perform outbound calls has low outbound calling efficiency, low utilization of human resources, and low outbound calling task completion rate. Summary of the Invention

[0005] Based on the above problems, this application provides a progressive collaborative outbound calling method, device, electronic device and storage medium, which performs layered processing according to the actual outbound calling results of different users, thereby improving outbound calling efficiency, human resource utilization and outbound calling task completion rate.

[0006] The embodiments of this application disclose the following technical solutions: In a first aspect, embodiments of this application provide a progressive collaborative outbound calling method, including: Obtain the outbound call task and the user information of the first user to be called corresponding to the outbound call task; Based on the outbound call task and the user information of the first user to be called, a first round of text-to-speech (TTS) outbound calls is performed on the first user to be called, and the first outbound call result corresponding to each first user to be called is collected; Based on the first outbound call result corresponding to each of the first outbound users, a second outbound user that meets the preset machine outbound call conditions is determined from the first outbound users; wherein, the machine outbound call conditions include: the first outbound call result indicates that the first round of TTS outbound calls has not completed the outbound call task; The robot performs a second round of outbound calls to the second user to be called, and collects the results of the second outbound calls for each user. Based on the second outbound call result corresponding to each second user to be called, a third user to be called who meets the preset manual outbound call conditions is determined from the second users to be called; wherein, the manual outbound call conditions include: the second outbound call result indicates that the second round of robot outbound calls has not completed the outbound call task; The user information of the third user to be called, the corresponding first outbound call result, the corresponding second outbound call result, and the call record are pushed to the human agent so that the human agent can perform the third round of human outbound call processing for the third user to be called.

[0007] In one possible implementation, the step of performing a first round of text-to-speech (TTS) outbound calls to the first user to be called based on the outbound call task and the user information of the first user to be called, and collecting the first outbound call results corresponding to each first user to be called, includes: Based on the outbound call task, obtain the outbound call script template corresponding to the outbound call task; Based on the outbound call script template and the user information of the first user to be called, generate an outbound call text corresponding to each of the first users to be called. Based on the user information of the first user to be called, a first round of TTS outbound calls is made to the first user to be called; After the first round of TTS outbound calls is connected, the corresponding TTS voice converted from the outbound text is played to the corresponding first callee. Collect the first outbound call result of the first user to be called during the first round of TTS outbound call process.

[0008] In one possible implementation, the first outbound call result includes: the call connection status of the first round of TTS outbound calls, the TTS voice playback status, and the first user feedback information; wherein, the call connection status of the first round of TTS outbound calls includes: connected or not connected; the TTS voice playback status includes: playback completed, playback interrupted, or not played.

[0009] In one possible implementation, the step of performing a second round of robot outbound calls to the second user to be called via an outbound calling robot, and collecting the second outbound call results corresponding to each second user to be called, includes: Based on the outbound call task, the user information of the second user to be called, and the first outbound call result corresponding to the second user to be called, the outbound call script flow of the outbound call robot is determined; Based on the user information of the second user to be called, the outbound calling robot makes a second round of robot outbound calls to the second user to be called. After the second round of robot outbound calls is connected, the outbound call robot interacts with the second user to be called via voice according to the outbound call script. Collect the second outbound call result during the voice interaction between the second user to be called and the outbound call robot.

[0010] In one possible implementation, the second outbound call result includes: the call connection status of the second round of robot outbound calls, user speech recognition text, and user intent recognition result; wherein, the call connection status of the second round of robot outbound calls includes: connected or not connected; the user speech recognition text is the text obtained by speech recognition of the speech data during the voice interaction between the second user to be called and the outbound call robot; the user intent recognition result is the result of intent recognition of the speech data.

[0011] In one possible implementation, pushing the user information of the third user to be called, the corresponding first outbound call result, the corresponding second outbound call result, and the call record to the human agent includes: Based on the user information of the third user to be called, the first outbound call result and the second outbound call result corresponding to the third user to be called, a work order to be processed is generated. The pending work order, the first round of outbound call records and the second round of outbound call records corresponding to the third user to be called are pushed to the human agent; wherein, the first round of outbound call records and the second round of outbound call records corresponding to the third user to be called together constitute the call record corresponding to the third user to be called.

[0012] In one possible implementation, pushing the pending work order, the first round of outbound call records and the second round of outbound call records corresponding to the third user to the human agent includes: According to at least one of the agent skill group, agent idle status and user level tag included in the user information of the third caller, assign a corresponding human agent to the third caller. The pending work order and the first and second rounds of outbound calls corresponding to the third user to be called are pushed to the assigned human agent.

[0013] Secondly, embodiments of this application provide a progressive collaborative outbound calling device, comprising: The information acquisition module is used to acquire outbound call tasks and user information of the first user to be called corresponding to the outbound call tasks; The first round outbound call module is used to perform a first round of text-to-speech (TTS) outbound calls to the first user to be called based on the outbound call task and the user information of the first user to be called, and to collect the first outbound call results corresponding to each first user to be called. The first filtering module is used to determine, based on the first outbound call result corresponding to each of the first outbound users, a second outbound user that meets the preset machine outbound call conditions from the first outbound users; wherein, the machine outbound call conditions include: the first outbound call result indicates that the first round of TTS outbound calls has not completed the outbound call task; The second-round outbound call module is used to perform a second-round robot outbound call to the second user to be called through an outbound call robot, and to collect the second outbound call results corresponding to each second user to be called; The second filtering module is used to determine a third user who meets the preset manual outbound calling conditions from the second users according to the second outbound calling result corresponding to each second user to be called; wherein, the manual outbound calling conditions include: the second outbound calling result indicates that the second round of robot outbound calling has not completed the outbound calling task; The third-round outbound call module is used to push the user information of the third user to be called, the corresponding first outbound call result, the corresponding second outbound call result and the call record to the human agent, so that the human agent can perform the third round of human outbound call processing for the third user to be called.

[0014] Thirdly, embodiments of this application provide an electronic device, including a processor and a memory; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the progressive collaborative outbound calling method as described in the first aspect.

[0015] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program or instructions, which, when executed, implements the progressive collaborative outbound calling method as described in the first aspect.

[0016] Beneficial Effects: This application first performs a first round of TTS outbound calls to the first pending users, and then filters out the second pending users who require further robot processing based on the first outbound call results for each first pending user; then, the outbound call robot performs a second round of robot outbound calls to the second pending users, and filters out the third pending users who require human intervention based on the second outbound call results for each second pending user; finally, the user information of the third pending users, the first outbound call results, the second outbound call results, and the call records are pushed to human agents, enabling human agents to perform a third round of human outbound call processing based on the previous outbound call records. Therefore, standardized, low-complexity outbound call tasks can be prioritized for completion via TTS outbound calls, outbound call tasks requiring interaction can be handled by outbound call robots, and outbound call tasks that robots cannot complete can be handled by human agents, thereby improving outbound call efficiency and human resource utilization, reducing the pressure on human agents, and increasing the overall completion rate of outbound call tasks. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating a progressive collaborative method provided in an embodiment of this application; Figure 2 This is a schematic diagram of the process of executing the first round of TTS outbound calls provided in an embodiment of this application; Figure 3 This is a schematic diagram of the process for executing the second round of robot outbound calls provided in an embodiment of this application; Figure 4 This is a schematic diagram of the structure of a progressive collaborative device provided in an embodiment of this application; Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0019] As described above, current outbound calling methods typically employ a single outbound calling approach, such as all outbound calls being made by human agents, all outbound calls being made by outbound calling robots, or all outbound calls being made via TTS voice broadcast.

[0020] While all outbound calls are handled by human agents, which can address simple payment reminders and handle complex communication scenarios, human agent resources are limited, resulting in high outbound call costs and low utilization of human resources. At the same time, human agents are limited by time and energy, making it impossible to make mass outbound calls 24 / 7. This lengthens the customer coverage cycle, ultimately leading to the ineffective consumption of human resources, low efficiency in mass outreach, failure to reach a large number of overdue customers in a timely manner, and a decline in the completion rate of outbound call tasks.

[0021] All outbound calls rely on one-way TTS voice broadcasting, lacking two-way dialogue and interaction capabilities. It cannot proactively identify customers' willingness to call back, and cannot effectively handle user inquiries, objections, or complex feedback, resulting in low outbound call efficiency and low outbound call task completion rate.

[0022] While NLP robots can perform some degree of voice interaction, they consume high computational resources for simple notifications to customers, resulting in wasted computing power. Furthermore, when intent recognition is inaccurate, user objections are complex, or users request human assistance, it is still difficult to complete the full loop, leading to low outbound call efficiency and task completion rate.

[0023] In summary, the current technology that uses a single outbound calling method results in low outbound calling efficiency, low utilization of human resources, and low outbound calling task completion rate.

[0024] This application provides a progressive collaborative outbound calling method, comprising: acquiring an outbound calling task and user information of a first user to be called corresponding to the outbound calling task; performing a first round of text-to-speech (TTS) outbound calls to the first user to be called based on the outbound calling task and the user information of the first user to be called, and collecting the first outbound call result for each first user to be called; determining a second user to be called from the first users to be called who meets preset machine outbound calling conditions based on the first outbound call result for each first user to be called; performing a second round of robot outbound calls to the second user to be called through an outbound calling robot, and collecting the second outbound call result for each second user to be called; determining a third user to be called from the second users to be called who meets preset human outbound calling conditions based on the second outbound call result for each second user to be called; and pushing the user information of the third user to be called, the corresponding first outbound call result, the corresponding second outbound call result, and the call record to a human agent, so that the human agent can perform a third round of human outbound calling processing for the third user to be called. This application performs layered processing based on the actual outbound call results of different users at different outbound calling stages, thereby improving outbound calling efficiency, human resource utilization, and outbound calling task completion rate.

[0025] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0026] The collection and processing of relevant data (including but not limited to experimental data, test data, simulation data, user data, etc.) involved in this application shall strictly comply with the requirements of national laws and regulations when applied in the following embodiments, obtain the informed consent or separate consent of the subject obtaining the data information, and carry out data use and processing within the scope of laws and regulations and the authorization of the subject.

[0027] Method Implementation Examples: The following is combined Figures 1-3 This application provides a detailed description of a progressive collaborative outbound calling method based on its embodiments.

[0028] like Figure 1 As shown in the embodiments of this application, a progressive collaborative outbound calling method includes the following steps: S101. Obtain the outbound call task and the user information of the first user to be called corresponding to the outbound call task.

[0029] Outbound calling tasks refer to tasks that require reaching users through outbound calls / telephone calls. For example, outbound calling tasks may include business notification tasks, bill reminder tasks, event notification tasks, customer follow-up tasks, satisfaction survey tasks, business confirmation tasks, marketing outreach tasks, risk warning tasks, etc.

[0030] Specifically, outbound calling tasks may include at least one of the following: task representation, task type, task name, task priority, task start time, task end time, outbound calling time window, outbound calling script template, task completion conditions, machine outbound calling conditions, manual outbound calling conditions, maximum number of outbound calls, outbound calling number pool, outbound calling strategy, etc.

[0031] The first user to be called refers to the user to be called at the beginning of the outbound call task. The first user to be called can be one user or multiple users.

[0032] Among them, the user information of the first user to be called refers to the user's contact information and business attribute data required to call the first user to be called.

[0033] Basic call information refers to outbound contact numbers such as the user's mobile phone number and landline number. Business attribute information refers to business information used for hierarchical outbound call determination, such as: policy number, premium due date, and whether it is in a grace period.

[0034] Furthermore, the user information of the first user to be called may also include: user identifier, user name, user's region, user's service type, user's historical outbound call records, user's historical interaction results, user level tag, user risk tag, user intention tag, user's preferred call time, etc.

[0035] In one possible implementation, outbound call tasks and the user information of the first user to be called corresponding to the outbound call task can be obtained from task databases, customer management systems, business systems, etc.

[0036] For example, suppose a certain business notification task (i.e. outbound call task) corresponds to 1,000 users to be called. Then, these 1,000 users to be called are the first users to be called corresponding to the business notification task, and the user information of these 1,000 first users to be called is obtained.

[0037] S102. Based on the outbound call task and the user information of the first user to be called, perform the first round of text-to-speech (TTS) outbound calls to the first user to be called, and collect the first outbound call results corresponding to each first user to be called.

[0038] Text-to-speech (TTS) is a speech synthesis technology. It can convert computer-readable text into playable natural human speech through algorithmic models, without requiring a real person to record audio in advance; it can generate speech output in real time simply by inputting text.

[0039] Among them, TTS outbound calling refers to an automated batch outbound calling technology based on text-to-speech technology, which belongs to one-way intelligent voice outreach methods.

[0040] The first outbound call result is used to determine whether the corresponding first user to be called needs to enter the second round of robot outbound calls.

[0041] Specifically, for each first user to be called, the corresponding first outbound call result is collected. For example, if the number of first outbound users corresponding to the outbound call task is 100, then 100 first outbound call results are collected, and the first outbound call result and the first outbound user are matched one by one.

[0042] To make it easier to understand, the following will be combined with... Figure 2 This document details how the first round of TTS outbound calls is performed in the embodiments of this application.

[0043] S201. Based on the outbound call task, obtain the outbound call script template corresponding to the outbound call task.

[0044] The outbound call script template is a pre-configured text template. Specifically, the outbound call script template can include fixed text content or variable fields. Variable fields can be used to populate personalized content such as user name, service name, date, amount, region, product name, and appointment time.

[0045] For example, the outbound call script template could be: "Dear (user name), the service (service name) you have signed up for will expire on (date). Please handle it in time. If you are aware of this, please press 1. If you need further information, please press 2."

[0046] S202. Based on the outbound call script template and the user information of the first user to be called, generate the outbound call text corresponding to each first user to be called.

[0047] Specifically, based on the user information of the first user to be called, the corresponding user information is filled into the variable fields of the outbound call script template to generate the outbound call text corresponding to the first user to be called.

[0048] For example, assuming the user information of the first user to be called includes: user name "Zhang San", service name "package renewal service", and expiration date "June 1, 2026", then the generated outbound call text corresponding to the first user to be called will be "Dear Zhang San, your package renewal service will expire on June 1, 2026. Please handle it in time. If you are aware of this, please press 1. If you need further information, please press 2."

[0049] S203. Based on the user information of the first user to be called, make the first round of TTS outbound calls to the first user to be called.

[0050] Specifically, the first round of TTS outbound calls is initiated based on the contact information of the first user to be called. The user's contact information may include mobile phone number, landline number, virtual number, enterprise communication number, etc.

[0051] In one possible implementation, the specific call time for initiating the first round of TTS outbound calls to the first user to be called is determined based on the outbound call time window of the outbound call task, the call frequency control strategy, the number reachability, and the user's preferred call time information, including user information. Based on the user information of the first user to be called, the first round of TTS outbound calls is made to the first user to be called at the specific call time.

[0052] S204. After the first round of TTS outbound calls is connected, play the corresponding TTS voice converted from the outbound text to the corresponding first callee.

[0053] Specifically, the outbound text corresponding to the first user to be called is converted to TTS to obtain the corresponding TTS voice. In the embodiments of this application, the TTS conversion can be completed in advance before initiating the first round of TTS outbound calls to the first user to be called, or the TTS conversion can be completed in real time after the first round of TTS outbound calls is connected. This application does not make any specific limitations.

[0054] S205. Collect the first outbound call result of the first user to be called in the first round of TTS outbound call process.

[0055] The first outbound call result includes: the call connection status of the first round of TTS outbound calls, the TTS voice playback status, and the first user feedback information.

[0056] The call connection status of the first round of TTS outbound calls includes: connected or not connected. Specifically, a call connection status of not connected in the first round of TTS outbound calls can include: no answer, busy line, rejected, phone switched off, phone disconnected, number not in service, network error, call failure, etc.

[0057] The TTS voice playback status includes: playback complete, playback interrupted, or not playing. Specifically, playback complete means that the TTS voice has finished playing; playback interrupted means that the user hung up during playback, the network was interrupted, or a system malfunction caused the TTS voice to not finish playing; not playing means that the user did not connect or was unable to start playing the TTS voice after connecting.

[0058] The first user feedback information may include at least one of the following: keypad feedback, voice feedback, confirmation feedback, rejection feedback, transfer to human assistance, or no feedback. For example, pressing "1" indicates that the user understands, pressing "2" indicates that the user needs further information, and pressing "0" indicates that the user will transfer to human assistance.

[0059] In one possible implementation, the first outbound call result may also include: the first round of outbound call time, the first round of call duration, the first round of call recording, the first round of call text record, etc.

[0060] The above combination Figure 2 This application provides a detailed description of how the first round of TTS outbound calls is performed in its embodiments. The following section will continue with further details. Figure 1 This application provides a detailed description of a progressive collaborative outbound calling method based on its embodiments.

[0061] S103. Based on the first outbound call result corresponding to each first callable user, determine the second callable user from the first callable users who meet the preset machine outbound call conditions.

[0062] The machine outbound calling conditions are used to determine which users in the first round of calls need to enter the second round of robot outbound calling. The machine outbound calling conditions include: the first outbound calling result indicates that the first round of TTS outbound calling did not complete the outbound calling task.

[0063] For example, the first outbound call result, indicating that the first round of TTS outbound calls did not complete the outbound call task, may include the following situations: the call connection status of the first round of TTS outbound calls is not connected; the TTS voice playback status is interrupted or not played; the TTS voice playback is completed, but no valid first user feedback information is collected; the first user feedback information does not match the completion conditions corresponding to the outbound call task; the first user feedback information indicates that the user needs further communication, the user has business consultation needs, or the user requests to proceed to the next step of processing, etc.

[0064] For example, for a business notification task, the completion condition for an outbound call task can be "the user answers and confirms receipt". If a first-waiting user does not answer the first round of TTS outbound calls, or does not listen to the entire TTS voice message after answering, or does not confirm receipt, then the first-waiting user meets the machine outbound call condition and can be identified as the second-waiting user.

[0065] For example, for a marketing intent collection task, the completion condition for the outbound call task can be "the user explicitly indicates no interest or that intent collection has been completed." If the user responds "needs to know more details" in the TTS outbound call, the first round of TTS outbound calls cannot complete a full communication, and the user can be identified as the second user to be called, and enter the second round of robot outbound calls.

[0066] In one possible implementation, the machine outbound call condition can be further combined with a preset outbound call termination condition. For example, when the first outbound call result indicates that the first round of TTS outbound calls has not been completed and the first pending user does not meet the preset outbound call termination condition, the first pending user is identified as the second pending user.

[0067] The preset outbound call termination conditions can include invalid numbers, disconnected numbers, blacklisted numbers, explicit refusal to continue contact, and reaching the maximum number of outbound calls. This method can prevent users who are not suitable for continued outbound calling from being pushed into the robot outbound call stage.

[0068] S104. The second round of robot outbound calls is executed to the second user to be called through the outbound call robot, and the second outbound call results corresponding to each second user to be called are collected.

[0069] The second round of robot outbound calls refers to outbound calls made using outbound calling robots.

[0070] Among them, the outbound call robot has the capabilities of automatic dialing, voice playback, voice recognition, semantic understanding, intent recognition, script flow control, and response generation.

[0071] To make it easier to understand, the following will be combined with... Figure 3 This application provides a detailed description of how the second round of robot outbound calls is performed in its embodiments.

[0072] S301. Based on the outbound call task, the user information of the second user to be called, and the first outbound call result corresponding to the second user to be called, determine the outbound call script flow of the outbound call robot.

[0073] The outbound call script process may include: opening remarks, identity verification, business introduction, problem confirmation, user intent judgment, exception handling, and closing remarks.

[0074] Specifically, the basic call script is determined based on the outbound call task, and then adjusted by combining the user information of the second user to be called and the result of the first outbound call to obtain the corresponding outbound call script.

[0075] For example: if the first outbound call result indicates that the user did not connect in the first round of TTS outbound calls, the second round of robot outbound calls can use standard reminder scripts; if the first outbound call result indicates that the user selected "need to know more" in the first round of TTS outbound calls, the second round of robot outbound calls can directly enter the business explanation or Q&A scripts; if the first outbound call result indicates that the user requested to be transferred to a human, the second round of robot outbound calls can be used to further confirm the user's needs or directly trigger the conditions for a human outbound call.

[0076] S302. Based on the user information of the second user to be called, the second round of robot outbound calls is made to the second user to be called through the outbound call robot.

[0077] Specifically, based on the contact information of the second user to be called, a second round of robot outbound calls is initiated through the outbound calling robot.

[0078] Furthermore, the calling order of the second round of robot outbound calls can be determined by combining factors such as user level, the result of the first outbound call, and the outbound call time window. The second round of robot outbound calls can then be made to the second user to be called based on the calling order of the second round of robot outbound calls.

[0079] S303. After the second round of robot outbound calls is connected, the outbound robot interacts with the second user to be called via voice according to the outbound call script.

[0080] Specifically, after the second round of robot outbound calls is connected, the outbound robot can play robot voice messages to the second user to be called according to the outbound call script, and receive voice data feedback from the second user to be called.

[0081] Specifically, the outbound call robot can jump to the corresponding script node based on the user's feedback. For example, when a user says "I want to know the specific details," the outbound call robot can enter the business description node; when a user says "I have already processed it," the outbound call robot can enter the result confirmation node; and when a user says "I want human customer service," the outbound call robot can record the request to transfer to a human.

[0082] S304. Collect the second outbound call results during the voice interaction between the second user to be called and the outbound call robot.

[0083] The second outbound call results include: the call connection status of the second round of robot outbound calls, the user's voice recognition text, and the user's intent recognition results.

[0084] The call connection status of the second round of robot outbound calls can include connected or not connected. Furthermore, not connected can include situations such as no one answering, busy line, rejected call, phone off, phone out of service, invalid number, and call failure.

[0085] User speech recognition text is text obtained by performing speech recognition on the speech data during the voice interaction between the second user to be called and the outbound call robot. For example, speech recognition can be achieved using an Automatic Speech Recognition (ASR) model.

[0086] User intent recognition results are obtained by performing intent recognition on speech data. Specifically, speech recognition is first performed on the speech data to obtain user speech recognition text, and then natural language understanding processing is performed on the user speech recognition text to obtain the user intent recognition result.

[0087] For example, the user intent recognition results may include: acknowledgment, consent to processing, refusal to process, need for consultation, request for manual assistance, objection, complaint, cannot be recognized, no valid intent, etc.

[0088] In one possible implementation, the second outbound call result may also include: intent recognition confidence level, robot dialogue rounds, slot information, user emotion tags, user objection tags, user complaint tags, user business intention tags, robot response results, second round call recordings, second round call text records, and other information.

[0089] The above combination Figure 3 This application provides a detailed description of how the second round of robot outbound calls is performed in its embodiments. The following section will continue in conjunction with... Figure 1 This application provides a detailed description of a progressive collaborative outbound calling method based on its embodiments.

[0090] S105. Based on the second outbound call result corresponding to each second callable user, determine the third callable user from the second callable users who meet the preset manual outbound call conditions.

[0091] The manual outbound calling criteria are used to determine which second-round callees need to enter the third round of manual outbound calling processing. Specifically, the manual outbound calling criteria include: the second outbound calling result indicates that the second round of robot outbound calling did not complete the outbound calling task.

[0092] For example, the second outbound call result indicating that the second round of robot outbound calls failed to complete the outbound call task may include: the call connection status of the second round of robot outbound calls is not connected; the outbound call robot failed to identify the user's valid intent; the user's speech recognition text is empty or cannot be recognized; the user's intent recognition result is unrecognizable or has no valid intent; the user explicitly requests human service; the user raises business inquiries, objections, complaints, or complex questions; the outbound call robot fails to obtain the information required to complete the outbound call task within the preset dialogue rounds; the second round of robot outbound calls is abnormally interrupted; the second outbound call result does not match the completion conditions corresponding to the outbound call task, etc.

[0093] For example, during the robot's outbound calling process, if a user says "I want to be transferred to a human operator," "You can't explain this clearly," or "I want to file a complaint," it is determined that the second round of robot outbound calling has not completed the outbound calling task. The second user waiting to be called meets the conditions for human outbound calling, and the user is identified as the third user waiting to be called.

[0094] For example, if the outbound call robot fails to recognize the user's intent after multiple inquiries, or if the user's response does not match the task completion conditions, it can be determined that the second round of robot outbound calls has not completed the outbound call task, and the corresponding user will be identified as the third user to be called.

[0095] S106. Push the user information of the third user to be called, the corresponding first outbound call result, the corresponding second outbound call result and the call record to the human agent so that the human agent can perform the third round of human outbound call processing for the third user to be called.

[0096] Among them, the third pending user is a user who has not completed the outbound calling task after the first round of TTS outbound calling and the second round of robot outbound calling and needs manual intervention.

[0097] In one possible implementation, a pending work order is generated based on the user information of the third user to be called, the first outbound call result, and the second outbound call result corresponding to the third user to be called. The pending work order, the first round of outbound call records, and the second round of outbound call records corresponding to the third user to be called are then pushed to a human agent. The human agent then performs the third round of manual outbound call processing for the third user to be called.

[0098] Among them, the first round of outbound call records and the second round of outbound call records corresponding to the third user to be called together constitute the call record corresponding to the third user to be called.

[0099] The first round of outbound call records may include: the call time of the first round of TTS outbound calls, the reason for the call not being connected, the content of the TTS voice playback, the first round of call recordings, and the first round of call text records, etc.

[0100] The second round of outbound call records may include: the call time of the second round of robot outbound calls, the robot's script, the number of robot dialogue rounds, the robot's response results, the user's request for human assistance, the recording of the second round of calls, and the text record of the second round of calls.

[0101] The pending work order can include information such as work order number, outbound call task identifier, user identifier of the third user to be called, user contact information, user level label, first outbound call result, second outbound call result, reason for task incompleteness, recommended processing method, recommended human agent script, and work order priority. Specifically, the pending work order is used to instruct the human agent on the outbound call method and purpose for the third user to be called.

[0102] Specifically, after receiving the work order and call records, the human agent can view the historical interactions of the third user in the first round of TTS outbound calls and the second round of robot outbound calls, thereby avoiding repeatedly asking the user for information that has already been provided and improving the efficiency of human outbound call processing.

[0103] In one possible implementation, a corresponding human agent is assigned to the third user waiting to be called based on at least one of the agent skill group, agent idle status, and user level tags included in the user information of the third user waiting to be called; the pending work order, the first round of outbound call records and the second round of outbound call records corresponding to the third user waiting to be called are pushed to the assigned human agent.

[0104] The call center skills groups include: business consultation skills group, complaint handling skills group, marketing conversion skills group, accounting skills group, and after-sales follow-up skills group.

[0105] The "Agent Idle Status" indicates whether a human agent is available to receive new outbound call work orders.

[0106] Among them, the user level tag is used to represent the user's value level, service level, priority level, etc.

[0107] For example, when the second outbound call result of a third user indicates that the user intends to complain, the system can prioritize assigning the pending work order to an idle human agent in the complaint handling skill group; when the third user has a high-level tag, the system can prioritize assigning it to a senior agent or a dedicated agent.

[0108] Furthermore, after completing the third round of outbound calls, the human agent can collect the results of the third round of calls and update the outbound call task status and user status based on these results. The third round of outbound call results may include the connection status of the third round of calls, the human agent's processing conclusion, user confirmation information, business processing results, human agent remarks, and the third round of call records.

[0109] This application provides a progressive collaborative outbound calling method, comprising: acquiring an outbound calling task and user information of a first pending user corresponding to the outbound calling task; performing a first round of text-to-speech (TTS) outbound calling on the first pending user based on the outbound calling task and the user information of the first pending user, and collecting the first outbound calling result corresponding to each first pending user; determining a second pending user from the first pending users who meets preset machine outbound calling conditions based on the first outbound calling result corresponding to each first pending user; performing a second round of robot outbound calling on the second pending user through an outbound calling robot, and collecting the second outbound calling result corresponding to each second pending user; determining a third pending user from the second pending users who meets preset human outbound calling conditions based on the second outbound calling result corresponding to each second pending user; and pushing the user information of the third pending user, the corresponding first outbound calling result, the corresponding second outbound calling result, and the call record to a human agent so that the human agent can perform a third round of human outbound calling processing for the third pending user.

[0110] In this embodiment, a first round of TTS outbound calls is first performed on the first user to be called. Based on the results of the first outbound calls for each user, a second user to be called who did not complete the first round of TTS outbound calls is selected. This allows standardized, notification, or simple confirmation outbound calls to be completed first through TTS outbound calls, which are less costly and more efficient, thus preventing all users from directly entering robot or human outbound calls. Furthermore, a second round of robot outbound calls is performed on the second user to be called through an outbound calling robot. Based on the results of the second outbound calls for each user, a third user to be called who did not complete the second round of robot outbound calls is selected. This allows the voice interaction capabilities of the outbound calling robot to handle users who require further communication but do not yet need to be directly transferred to a human. Furthermore, only the third user who has not completed the outbound call task after the first round of TTS outbound calls and the second round of robot outbound calls is pushed to the human agent. At the same time, the corresponding first outbound call result, second outbound call result and call record are pushed to the human agent, so that the human agent can make targeted processing based on the user feedback, outbound call status and interaction record in the previous outbound call process, reducing duplicate communication and invalid outbound calls.

[0111] Therefore, in this embodiment of the application, the actual outbound call results of different users at different outbound call stages are processed in a hierarchical manner, so that simple tasks are completed by TTS outbound calls, interactive tasks are handled by outbound call robots, and complex or non-closed-loop tasks are handled by human agents, thereby improving outbound call efficiency, human resource utilization and outbound call task completion rate.

[0112] Device Example: The following is combined Figure 4 This application provides a detailed description of a progressive collaborative outbound calling device based on its embodiments.

[0113] like Figure 4 As shown in the embodiment of this application, a progressive collaborative outbound calling device includes the following modules: Information acquisition module 401 is used to acquire outbound call tasks and user information of the first user to be called corresponding to the outbound call task; The first-round outbound call module 402 is used to perform the first round of text-to-speech (TTS) outbound calls to the first user to be called based on the outbound call task and the user information of the first user to be called, and to collect the first outbound call results corresponding to each first user to be called. The first filtering module 403 is used to determine the second user to be called from the first users to be called that meets the preset machine outbound calling conditions based on the first outbound calling result corresponding to each first user to be called; wherein, the machine outbound calling conditions include: the first outbound calling result indicates that the first round of TTS outbound calling has not completed the outbound calling task; The second-round outbound call module 404 is used to perform a second-round robot outbound call to a second user to be called through an outbound call robot, and to collect the second outbound call results corresponding to each second user to be called; The second screening module 405 is used to determine the third user to be called from the second users to be called that meets the preset manual outbound calling conditions based on the second outbound calling result corresponding to each second user to be called; wherein, the manual outbound calling conditions include: the second outbound calling result indicates that the second round of robot outbound calling has not completed the outbound calling task; The third-round outbound call module 406 is used to push the user information of the third user to be called, the corresponding first outbound call result, the corresponding second outbound call result and the call record to the human agent, so that the human agent can perform the third round of human outbound call processing for the third user to be called.

[0114] In one possible implementation, the first-round outbound call module 402 is specifically used for: Based on the outbound call task, obtain the outbound call script template corresponding to the outbound call task; Based on the outbound call script template and the user information of the first user to be called, generate the outbound call text corresponding to each first user to be called. Based on the user information of the first user to be called, the first round of TTS outbound calls is made to the first user to be called; After the first round of TTS outbound calls is connected, the corresponding TTS voice converted from the outbound text is played to the first callee. Collect the first outbound call result of the first user to be called in the first round of TTS outbound call process.

[0115] In one possible implementation, the second-round outbound call module 404 is specifically used for: Based on the outbound call task, the user information of the second user to be called, and the first outbound call result corresponding to the second user to be called, determine the outbound call script flow of the outbound call robot; Based on the user information of the second user to be called, a second round of robot outbound calls is made to the second user to be called through an outbound calling robot; After the second round of robot outbound calls is connected, the outbound robot will conduct voice interaction with the second callee according to the outbound call script. Collect the results of the second outbound call during the voice interaction between the second user to be called and the outbound call robot.

[0116] In one possible implementation, the third-round outbound call module 406 is specifically used to: generate a pending work order based on the user information of the third pending user, the first outbound call result and the second outbound call result corresponding to the third pending user; and push the pending work order, the first-round outbound call record and the second-round outbound call record corresponding to the third pending user to the human agent; wherein the first-round outbound call record and the second outbound call record corresponding to the third pending user together constitute the call record corresponding to the third pending user.

[0117] In one possible implementation, the third round outbound call module 406 is specifically used to: assign a corresponding human agent to the third user to be called according to at least one of the agent skill group, agent idle status, and user level tags included in the user information of the third user to be called; and push the pending work order, the first round outbound call record and the second round outbound call record corresponding to the third user to the assigned human agent.

[0118] It should be noted that the progressive collaborative outbound calling device provided in this application embodiment has the same beneficial effects as the progressive collaborative outbound calling method provided in the above method embodiment, and therefore will not be described again.

[0119] In one possible implementation, see Figure 5 This figure is a schematic diagram of an electronic device provided in an embodiment of this application.

[0120] The electronic device may include a memory 501 and a processor 502. For example... Figure 5 As shown, the memory can be random access memory (RAM), flash memory, read-only memory (ROM), EPROM, non-volatile read-only memory (Electronic Programmable ROM), registers, hard disks, removable disks, etc.

[0121] The memory 501 can store computer-executed instructions. When the computer-executed instructions stored in the memory 501 are executed by the processor 502, the processor 502 can be used to execute a progressive, collaborative outbound call method. The memory 501 can also store data.

[0122] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to the embodiments of this application is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that integrates one or more available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape) or a semiconductor medium (e.g., solid-state disk (SSD)).

[0123] This application also provides a computer-readable storage medium storing a computer program or instructions that, when executed, implement a progressive, collaborative outbound call method. Examples include random access memory (RAM), flash memory, read-only memory (ROM), EPROM, non-volatile read-only memory (EPROM), registers, hard disks, removable disks, or any other storage medium in the art.

[0124] In the embodiments of this application, the terms "first" and "second" (if they exist) are used only as name identifiers and do not represent the order of first and second.

[0125] It should be noted that the various embodiments in this specification are described in a progressive manner, with each embodiment focusing on the differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. Regarding the methods disclosed in the embodiments, since they correspond to the product embodiments disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the description of the product embodiments.

[0126] The above description of the disclosed embodiments enables those skilled in the art to make or use this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A progressive collaborative outbound calling method, characterized in that, include: Obtain the outbound call task and the user information of the first user to be called corresponding to the outbound call task; Based on the outbound call task and the user information of the first user to be called, a first round of text-to-speech (TTS) outbound calls is performed on the first user to be called, and the first outbound call result corresponding to each first user to be called is collected; Based on the first outbound call result corresponding to each of the first outbound users, a second outbound user that meets the preset machine outbound call conditions is determined from the first outbound users; wherein, the machine outbound call conditions include: the first outbound call result indicates that the first round of TTS outbound calls has not completed the outbound call task; The robot performs a second round of outbound calls to the second user to be called, and collects the results of the second outbound calls for each user. Based on the second outbound call result corresponding to each second user to be called, a third user to be called who meets the preset manual outbound call conditions is determined from the second users to be called; wherein, the manual outbound call conditions include: the second outbound call result indicates that the second round of robot outbound calls has not completed the outbound call task; The user information of the third user to be called, the corresponding first outbound call result, the corresponding second outbound call result, and the call record are pushed to the human agent so that the human agent can perform the third round of human outbound call processing for the third user to be called.

2. The method according to claim 1, characterized in that, Based on the outbound call task and the user information of the first user to be called, the first round of text-to-speech (TTS) outbound calls is performed on the first user to be called, and the first outbound call results corresponding to each first user to be called are collected, including: Based on the outbound call task, obtain the outbound call script template corresponding to the outbound call task; Based on the outbound call script template and the user information of the first user to be called, generate an outbound call text corresponding to each of the first users to be called. Based on the user information of the first user to be called, a first round of TTS outbound calls is made to the first user to be called; After the first round of TTS outbound calls is connected, the corresponding TTS voice converted from the outbound text is played to the corresponding first callee. Collect the first outbound call result of the first user to be called during the first round of TTS outbound call process.

3. The method according to claim 2, characterized in that, The first outbound call result includes: the call connection status of the first round of TTS outbound calls, the TTS voice playback status, and the first user feedback information; wherein, the call connection status of the first round of TTS outbound calls includes: connected or not connected; the TTS voice playback status includes: playback completed, playback interrupted, or not played.

4. The method according to claim 1, characterized in that, The step of performing a second round of robot outbound calls to the second user to be called via an outbound calling robot, and collecting the second outbound call results for each second user to be called, includes: Based on the outbound call task, the user information of the second user to be called, and the first outbound call result corresponding to the second user to be called, the outbound call script flow of the outbound call robot is determined; Based on the user information of the second user to be called, the outbound calling robot makes a second round of robot outbound calls to the second user to be called. After the second round of robot outbound calls is connected, the outbound call robot interacts with the second user to be called via voice according to the outbound call script. Collect the second outbound call result during the voice interaction between the second user to be called and the outbound call robot.

5. The method according to claim 4, characterized in that, The second outbound call result includes: the call connection status of the second round of robot outbound calls, the user's voice recognition text, and the user's intent recognition result; wherein, the call connection status of the second round of robot outbound calls includes: connected or not connected; the user's voice recognition text is the text obtained by voice recognition of the voice data during the voice interaction between the second user to be called and the outbound call robot; the user's intent recognition result is the result of intent recognition of the voice data.

6. The method according to claim 1, characterized in that, The step of pushing the user information of the third user to be called, the corresponding first outbound call result, the corresponding second outbound call result, and the call record to the human agent includes: Based on the user information of the third user to be called, the first outbound call result and the second outbound call result corresponding to the third user to be called, a work order to be processed is generated. The pending work order, the first round of outbound call records and the second round of outbound call records corresponding to the third user to be called are pushed to the human agent; wherein, the first round of outbound call records and the second round of outbound call records corresponding to the third user to be called together constitute the call record corresponding to the third user to be called.

7. The method according to claim 6, characterized in that, The step of pushing the pending work order, the first round of outbound call records and the second round of outbound call records corresponding to the third user to the human agent includes: According to at least one of the agent skill group, agent idle status and user level tag included in the user information of the third caller, assign a corresponding human agent to the third caller. The pending work order and the first and second rounds of outbound calls corresponding to the third user to be called are pushed to the assigned human agent.

8. A progressive collaborative outbound calling device, characterized in that, include: The information acquisition module is used to acquire outbound call tasks and user information of the first user to be called corresponding to the outbound call tasks; The first round outbound call module is used to perform a first round of text-to-speech (TTS) outbound calls to the first user to be called based on the outbound call task and the user information of the first user to be called, and to collect the first outbound call results corresponding to each first user to be called. The first filtering module is used to determine, based on the first outbound call result corresponding to each of the first outbound users, a second outbound user that meets the preset machine outbound call conditions from the first outbound users; wherein, the machine outbound call conditions include: the first outbound call result indicates that the first round of TTS outbound calls has not completed the outbound call task; The second-round outbound call module is used to perform a second-round robot outbound call to the second user to be called through an outbound call robot, and to collect the second outbound call results corresponding to each second user to be called; The second filtering module is used to determine a third user who meets the preset manual outbound calling conditions from the second users according to the second outbound calling result corresponding to each second user to be called; wherein, the manual outbound calling conditions include: the second outbound calling result indicates that the second round of robot outbound calling has not completed the outbound calling task; The third-round outbound call module is used to push the user information of the third user to be called, the corresponding first outbound call result, the corresponding second outbound call result and the call record to the human agent, so that the human agent can perform the third round of human outbound call processing for the third user to be called.

9. An electronic device, characterized in that, The device includes: a processor and a memory; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the progressive collaborative outbound calling method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program or instructions that, when executed, implement the progressive collaborative outbound calling method as described in any one of claims 1-7.

Images