Voice outbound method for questionnaire return visit, voice outbound device and storage medium

Abstract

The invention discloses a voice outbound method for questionnaire return visit, voice outbound equipment and a storage medium. The method comprises the following steps: acquiring questionnaire configuration data set by a user on a first configuration page corresponding to any type of work order; generating a return visit questionnaire corresponding to the work order according to the questionnaire configuration data, and generating an outbound return visit task corresponding to each work order according to the return visit questionnaire when the state of the work order is a complete state; calling a return visit object of the call-out return visit task, and converting each questionnaire question in a return visit questionnaire corresponding to the call-out return visit task into a voice question; and collecting reply data of the return visit object for each voice question in real time so as to execute the jump logic of each questionnaire question until any end question of the return visit questionnaire is jumped to, thereby completing the call-out return visit task. According to the scheme, the call-out return visit task of a large number of work orders can be automatically completed, and the return visit efficiency, the accuracy of questionnaire results and the flexibility of questionnaire questioning are improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent voice technology, and specifically to a voice outbound calling method, voice outbound calling device and storage medium for questionnaire follow-up. Background Art

[0002] With the rapid development of artificial intelligence technology, especially the significant progress in the field of natural language processing, the application of AI for human-computer interaction is becoming more and more widespread. Among them, automated customer service systems, such as intelligent customer service robots, have become an important tool for many companies to improve service efficiency and service quality. Questionnaire feedback, as an important part of customer relationship management, is of great significance for collecting user feedback and improving service quality. Traditional questionnaire feedback usually relies on manual operation, which is not only costly but also inefficient, and it is difficult to meet the needs of large-scale data collection. Summary of the invention

[0003] The purpose of the embodiments of the present application is to provide a voice outbound calling method, a voice outbound calling device and a storage medium for questionnaire return interviews, so as to solve the technical problem of low efficiency of traditional questionnaire return interviews in the prior art.

[0004] In order to achieve the above-mentioned purpose, the first aspect of the present application provides a voice outbound calling method for questionnaire return interview, the method comprising:

[0005] Obtain the questionnaire configuration data set by the user on the first configuration page corresponding to any type of work order;

[0006] Generate a callback questionnaire corresponding to the work order according to the questionnaire configuration data, the callback questionnaire including a plurality of questionnaire questions with jump logic;

[0007] When the status of the work order is completed, an outbound call back task corresponding to each work order is generated according to the back-call questionnaire;

[0008] Call the object of the outbound call back task, and convert each questionnaire question in the questionnaire corresponding to the outbound call back task into a voice question;

[0009] Collect the response data of the interviewees to each voice question in real time;

[0010] The jump logic of each questionnaire question is executed according to the response data of each voice question until jumping to any end question of the follow-up questionnaire to complete the outbound call follow-up task.

[0011] In an embodiment of the present application, the questionnaire configuration data includes a prompt word and a question type for each questionnaire question, and multiple preset question answers. The jump logic of each questionnaire question is executed according to the reply data of each voice question, including: the large language model analyzes the reply data of each voice question according to the prompt word and the question type of each questionnaire question to output the answer to the voice question; determines whether the answer to each voice question output by the large language model matches any one of the multiple preset question answers corresponding to the voice question; when the answer to each voice question matches any one of the corresponding multiple preset question answers, executes the jump logic of the questionnaire question corresponding to the voice question according to the answer to each voice question.

[0012] In an embodiment of the present application, the large language model analyzes the response data of each voice question according to the prompt words and the question type of each questionnaire question to output the answer to the question of the voice question, including: when the question type is a multiple-choice question, the large language model semantically matches the response data of each voice question with multiple preset question answers corresponding to the voice question according to the prompt words of each questionnaire question, so as to output a target question answer that matches the response data of the voice question, and uses the target question answer as the answer to the question of the voice question; when the question type is a fill-in-the-blank question, the large language model determines the keywords in the response data of each voice question according to the prompt words of each questionnaire question, and generates the answer to the question of the voice question according to the keywords; when the question type is a question-and-answer question, the large language model determines whether the response data of each voice question matches the voice question according to the prompt words of each questionnaire question, so as to obtain a matching result, and uses the matching result as the answer to the question of the voice question.

[0013] In an embodiment of the present application, generating an answer to the voice question based on the keyword includes: generating a key-value pair based on the keyword and a blank part of each questionnaire title, so as to use the key-value pair as the answer to the voice question.

[0014] In an embodiment of the present application, the questionnaire configuration data also includes whether to repeat the question for each questionnaire question, the threshold of the number of questions, and the threshold of the waiting time. The method also includes: when the answer to each voice question does not match any one of the multiple preset questions, and it is determined that the questionnaire question corresponding to the voice question is to be repeated, the voice question is repeatedly played within the threshold of the number of questions; when the waiting time for collecting the answer to each voice question is greater than the threshold of the waiting time, jump to any end question of the follow-up questionnaire corresponding to the voice question to complete the outbound call follow-up task.

[0015] In an embodiment of the present application, the questionnaire configuration data includes a question type, jump logic, prompt words, multiple end questions, a query template, and inserted variables corresponding to each questionnaire question, wherein the inserted variables are generated according to the work order information of each work order.

[0016] In an embodiment of the present application, calling the follow-up object includes: obtaining the task configuration data set by the user on the second configuration page corresponding to the outbound call follow-up task, the task configuration data including the execution time period, outbound call rules and task priority of the outbound call follow-up task; and initiating a follow-up operation on the follow-up object according to the task configuration data.

[0017] In an embodiment of the present application, the questionnaire configuration data also includes a background call interface set for a target questionnaire question among multiple questionnaire questions, and a target question answer that triggers the background call interface. The method also includes: when it is determined that the response data of the follow-up object to the target questionnaire question is the answer to the target question, triggering the background call interface to execute the corresponding background service operation.

[0018] A second aspect of the present application provides a voice outbound calling device, including:

[0019] a memory configured to store instructions;

[0020] The processor is configured to call instructions from the memory and implement the above-mentioned voice outbound calling method for questionnaire return interview when executing the instructions.

[0021] A third aspect of the present application provides a machine-readable storage medium having instructions stored thereon, the instructions being used to enable a machine to execute the above-mentioned voice outbound calling method for questionnaire follow-up.

[0022] Through the above scheme, according to the configuration page provided by the system, after configuring the callback questionnaire on the configuration page, the callback task is automatically triggered based on the conditions. For example, if the service type is a work order for equipment maintenance and the service order status is completed, the callback task for the equipment maintenance status of the corresponding customer of the work order is triggered, and the outbound callback task is performed through the voice outbound call robot. The system can automatically complete the outbound callback tasks for a large number of work orders, saving labor costs and improving callback efficiency. In addition, the system understands the customer's reply and fills it in the corresponding answer to the questionnaire, and can jump to the subsequent corresponding questions in real time based on the answer to the current question to continue asking, ensuring the accuracy of the questionnaire results and the flexibility of the questionnaire questions.

[0023] Other features and advantages of the embodiments of the present application will be described in detail in the subsequent specific implementation section. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the following specific implementations, they are used to explain the embodiments of the present application, but do not constitute a limitation on the embodiments of the present application. In the accompanying drawings:

[0025] Figure 1 A schematic diagram of a flow chart of a voice outbound calling method for questionnaire return interview according to an embodiment of the present application is shown;

[0026] Figure 2 A schematic diagram of a flow chart of a voice outbound calling method for questionnaire return interview according to a specific embodiment of the present application is shown;

[0027] Figure 3 A structural block diagram of a voice outbound calling device for questionnaire return interview according to an embodiment of the present application is schematically shown;

[0028] Figure 4 The schematic diagram shows a structure of a computer device according to an embodiment of the present application. DETAILED DESCRIPTION

[0029] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the specific implementation methods described herein are only used to illustrate and explain the embodiments of the present application, and are not used to limit the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0030] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present application, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or suggesting their relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0031] Figure 1 The flowchart of a voice outbound calling method for questionnaire return interview according to an embodiment of the present application is schematically shown. Figure 1 As shown, an embodiment of the present application provides a voice outbound calling method for questionnaire follow-up, which may include the following steps.

[0032] S102, obtaining questionnaire configuration data set by the user on a first configuration page corresponding to any type of work order;

[0033] S104, generating a callback questionnaire corresponding to the work order according to the questionnaire configuration data, the callback questionnaire including a plurality of questionnaire questions with jump logic;

[0034] S106, when the work order is in the completed state, generating an outbound call back task corresponding to each work order according to the back-call questionnaire;

[0035] S108, calling the object of the outbound call back task, and converting each questionnaire question in the back questionnaire corresponding to the outbound call back task into a voice question;

[0036] S110, collecting the response data of the interviewee to each voice question in real time;

[0037] S112, executing the jump logic of each questionnaire question according to the reply data of each voice question, until jumping to any end question of the return questionnaire, so as to complete the outbound call return task.

[0038] It can be understood that work orders include many different types. For example, a work order with a service type of equipment maintenance and a service order status of completed service requires a return visit to the equipment maintenance status of the corresponding customer of the work order. Due to the variety of work order types, different return questionnaires need to be designed for different types of work orders, so a configuration page is designed for configuring multiple sets of return questionnaires. Specifically, the first configuration page is used to configure the return questionnaire. On the first configuration page, the user can set the corresponding questionnaire configuration data for different types of work orders, and generate multiple sets of return questionnaires according to the questionnaire configuration data. For example, the questionnaire configuration data includes but is not limited to the prompt words, question types, preset questions and answers, jump logic, etc. for each questionnaire title. When the status of the work order is completed, an outbound call return task corresponding to the work order can be automatically generated according to the return questionnaire. The outbound call return task refers to calling the return object according to the questionnaire title of the return questionnaire through a voice outbound call robot, and collecting the reply data of the return object. The return object usually refers to the service object of the work order. In the process of calling the return object of the outbound call return task, each questionnaire title in the return questionnaire can be converted into a voice question. Specifically, the TTS service (Text-to-Speech) can be called to convert the robot speech configured in the question or the text of the question itself into voice and play it to the customer. At the same time, the voice outbound call robot can collect the response data of the return visit object to each voice question in real time, and use the ASR service (Automatic Speech Recognition) to convert it into a text reply. For example, if the robot is still reading the question, the customer has made a reply, and the robot will immediately stop reading the question and enter the processing link of the customer's reply. In addition, the jump logic of each questionnaire question can be executed according to the reply data of each voice question. For example, if the current question is question 2, you can set option A to jump to question 4, and option B to jump to question 5, etc. In this way, it will jump to any end question of the return visit questionnaire to complete the outbound call return visit task. Since different end questions are ultimately caused by the jump question selection, a set of return visit questionnaires can set multiple end questions, and each end question can be configured as a closing sentence. After reading the question, the phone can be automatically hung up to end the current telephone return visit.

[0039] Through the above scheme, according to the configuration page provided by the system, after configuring the callback questionnaire on the configuration page, the callback task is automatically triggered based on the conditions. For example, if the service type is a work order for equipment maintenance and the service order status is completed, the callback task for the equipment maintenance status of the corresponding customer of the work order is triggered, and the outbound callback task is performed through the voice outbound call robot. The system can automatically complete the outbound callback tasks for a large number of work orders, saving labor costs and improving callback efficiency. In addition, the system understands the customer's reply and fills it in the corresponding answer to the questionnaire, and can jump to the subsequent corresponding questions in real time based on the answer to the current question to continue asking, ensuring the accuracy of the questionnaire results and the flexibility of the questionnaire questions.

[0040] In an embodiment of the present application, the questionnaire configuration data includes a prompt word and a question type for each questionnaire question, and multiple preset question answers. The jump logic of each questionnaire question is executed according to the reply data of each voice question, including: the large language model analyzes the reply data of each voice question according to the prompt word and the question type of each questionnaire question to output the answer to the voice question; determines whether the answer to each voice question output by the large language model matches any one of the multiple preset question answers corresponding to the voice question; when the answer to each voice question matches any one of the corresponding multiple preset question answers, executes the jump logic of the questionnaire question corresponding to the voice question according to the answer to each voice question.

[0041] It can be understood that the types of questions in the questionnaire include multiple-choice questions, fill-in-the-blank questions, and essay questions. Among them, multiple-choice questions include single-choice questions and multiple-choice questions. The preset answers to the questions include the answers preset by the technicians for users to choose from or the answers that the users may answer in advance. The jump logic of the questions can be preset by presetting the answers to the questions. For single-choice questions and multiple-choice questions, it supports the configuration to jump to the subsequent specified questions after selecting a certain option. For example, if the current question is question 2, it can be set to jump to question 4 by selecting option A, and jump to question 5 by selecting option B. The prompt words are set for the large language model to analyze the user's response data. For single-choice questions, multiple-choice questions, fill-in-the-blank questions, and essay questions, the system will provide a general prompt word for each type of question type to guide the large model to obtain the answer to the question of this type according to the customer's reply. At the same time, it supports users to configure their targeted prompt words for each questionnaire question to meet the special needs of the corresponding questions. Therefore, the large language model analyzes the response data of each voice question according to the prompt words and question type of each questionnaire question to output the answer to the question asked by the voice question. Determine whether the answer to each voice question output by the large language model matches any one of the multiple preset answers corresponding to the voice question. If the answer to each voice question matches any one of the corresponding multiple preset answers, execute the jump logic of the questionnaire question corresponding to the voice question according to the answer to each voice question.

[0042] In an embodiment of the present application, the large language model analyzes the response data of each voice question according to the prompt words and the question type of each questionnaire question to output the answer to the question of the voice question, including: when the question type is a multiple-choice question, the large language model semantically matches the response data of each voice question with multiple preset question answers corresponding to the voice question according to the prompt words of each questionnaire question, so as to output a target question answer that matches the response data of the voice question, and uses the target question answer as the answer to the question of the voice question; when the question type is a fill-in-the-blank question, the large language model determines the keywords in the response data of each voice question according to the prompt words of each questionnaire question, and generates the answer to the question of the voice question according to the keywords; when the question type is a question-and-answer question, the large language model determines whether the response data of each voice question matches the voice question according to the prompt words of each questionnaire question, so as to obtain a matching result, and uses the matching result as the answer to the question of the voice question.

[0043] In an embodiment of the present application, generating an answer to the voice question based on the keyword includes: generating a key-value pair based on the keyword and a blank part of each questionnaire title, so as to use the key-value pair as the answer to the voice question.

[0044] refer to Figure 2, the outbound voice robot can dial the return customer phone number corresponding to the questionnaire and call the return object. In this process, the TTS service is called in real time to convert the text of the return question in the questionnaire into voice playback, and the ASR service is called in real time to identify the customer's reply and convert it into corresponding text information. At this time, it is necessary to determine the type of the questionnaire question currently being asked. Specifically, after obtaining the customer's reply, different processing logic is used according to the current question type. If the current question is a multiple-choice question (including single-choice questions and multiple-choice questions), the prompt word is used to guide the large language model to obtain the answer options for the current question based on the customer's reply. An example of using general prompt words is as follows: General prompt words: You are a call center agent. Please receive the multiple-choice question title, options and customer's text reply, analyze the customer's reply, extract key information, match the corresponding option or judge that the reply is invalid. Please answer the matching option identifier (such as "A", "B", "C", "D") or "unrecognizable". Input information: Title: Dear customer, I would like to visit and ask whether the engineer has recently installed your device? A. Already served B. Not served C. Not sure. Customer reply: Installed. Large language model output results: A.

[0045] The following is an example of using exclusive prompt words: Topic exclusive prompt words: You are a call center agent, and you will receive the customer's objective rating of this service. The rating options are as follows: A.10; B.9; C.8; D.7; E.6; F.5; G.4; H.3; I.2; J.1; K. Unwilling to rate. In addition to the direct reply score, the customer's "very good", "very satisfied", "satisfied", "pretty good" is 10 points, "okay", "not bad", "okay" is 9 points, "unsatisfied", "poor service", "not good service" is 8 points, and "lowest score" is 1 point. Please follow the above logic and match the corresponding options according to the customer's text reply or judge that the reply is invalid. Please output the matching option identifier (such as "A", "B", "C", "D") or "unrecognizable". Input information: too bad. The large language model output result: C.

[0046] If the current question is a fill-in-the-blank question, use prompt words to guide the large language model to obtain the answer to the blank in the question from the customer's reply. An example of using exclusive prompt words is as follows: Exclusive prompt words for a certain question: You are a call center agent. Please receive the fill-in-the-blank question title and the customer's text reply, analyze the customer's reply, extract key information, and output each blank and its answer in JSON format as a key-value pair. Question: Do you want to buy equipment in ____ province ____ city ____ district? If the customer only replies about the urban area, please fill in the information of the entire province based on common sense. Input information: Customer reply: Yuelu District, Changsha. The large model output result: {"province":"Hunan","city":"Changsha","district":"Yuelu"}.

[0047] If the current question is a question-and-answer question, use prompt words to guide the large model to determine whether the customer's text reply is an answer to the question. If so, directly use the customer's text reply as the answer to the question. If not, the answer to the question is invalid. Examples of using general prompt words are as follows: General prompt words: You are a call center agent. Please receive the question-and-answer question and the customer's text reply, analyze the customer's reply, and determine whether the customer's text reply is an answer to the question. Please answer "yes" or "no". Input information: Title: Dear customer, do you have any suggestions for our service? Customer reply: Are you a robot? Model output result: No.

[0048] If the big model recognizes the answer to the current question, it will make the following flow judgment based on the question type of the current question: if the current question is a multiple-choice question, it will jump to the next question based on the jump logic of the selected answer option; if the current question is a fill-in-the-blank question or a question-and-answer question, it will directly flow to the next question in sequence.

[0049] In an embodiment of the present application, the questionnaire configuration data also includes whether to repeat the question for each questionnaire question, the threshold of the number of questions, and the threshold of the waiting time. The method also includes: when the answer to each voice question does not match any one of the multiple preset questions, and it is determined that the questionnaire question corresponding to the voice question is to be repeated, the voice question is repeatedly played within the threshold of the number of questions; when the waiting time for collecting the answer to each voice question is greater than the threshold of the waiting time, jump to any end question of the follow-up questionnaire corresponding to the voice question to complete the outbound call follow-up task.

[0050] It can be understood that the inquiry number threshold refers to the maximum number of repeated inquiries set for each questionnaire question. Figure 2 If the big model determines that the answer cannot be recognized or the answer is invalid, it will be processed according to the strategy configured by the voice outbound call robot. If it is necessary to repeat the question and the maximum number of repeated questions for the question has not been reached, repeat the above steps according to the script configured for the question and ask the customer the question again. If it is not necessary to repeat the question or the maximum number of repeated questions for the question has been reached, mark the question as "unrecognizable" and jump to the corresponding closing remarks. At this time, you can call the TTS service to convert the closing remarks text into voice playback, and then automatically hang up the phone to end the outbound call callback task.

[0051] In an embodiment of the present application, the questionnaire configuration data includes a question type, jump logic, prompt words, multiple end questions, a query template, and inserted variables corresponding to each questionnaire question, wherein the inserted variables are generated according to the work order information of each work order.

[0052] Specifically, robot script configuration: you can configure the script used by the robot when asking the customer about the topic. If the robot supports repeated asking, you can also configure the script used for the first, second, and other askings differently. If this content is not configured, the system can default to asking the customer the topic content as the script. Topic variable configuration: you can insert variables in the topic or script content. When calling back, the system will dynamically fill in the variable content according to the work order information corresponding to the call back task. For example: "Dear #customer name#, hello! Has the #fault description# you reported at #reporting time# been resolved?" Among them, #customer name#, #reporting time#, and #fault description# will be dynamically filled in according to the corresponding field information in the repair work order corresponding to the outbound call back task.

[0053] Special question attribute configuration: Supports the marking configuration of the "first question" and "last question" actually asked to customers in the questionnaire. Before the first question and after the last question, you can also configure questions that are not directly asked to customers, but can be answered based on the background logic. There can usually be only one first question, and there can be multiple last questions (depending on the choice of skipping questions, which ultimately leads to different last questions), and the last question can be configured as a concluding sentence, and the call will automatically hang up after reading the question to end the call. For example, before the first question, you can also add the question: "Question: Was the call successfully connected? (Single choice), options: A. Successfully connected; B. No one answered; C. Rejected; D. Powered off; E. Out of service; F. No number available". The system can automatically get the answer to the question in the background based on the call connection status, so that users can know the status of the outbound call.

[0054] In an embodiment of the present application, calling the follow-up object includes: obtaining the task configuration data set by the user on the second configuration page corresponding to the outbound call follow-up task, the task configuration data including the execution time period, outbound call rules and task priority of the outbound call follow-up task; and initiating a follow-up operation on the follow-up object according to the task configuration data.

[0055] It can be understood that the system supports different ways to generate outbound call return tasks, which can be automatically triggered based on conditions. For example, if the service type is a work order for equipment maintenance, when the service order status is completed, the generation of a return task for the equipment maintenance situation of the corresponding customer of the work order is triggered. It can also be generated by batch import, such as batch importing a large number of customer phone numbers and returning them in turn. Among them, the execution time period, outbound call rules and task priority of the voice outbound call robot to perform the outbound call return task can be configured. For example, only call from 9:00 to 23:00 on weekdays. The outbound call rules include strategies when the call fails, such as whether to repeat the call, the maximum number of repetitions, and the time interval for repeated calls. The voice outbound call robot can call the return object phone in the task in turn according to the task priority. If the call can be dialed, the customer will be automatically returned according to the following steps. For example, the current question of the initial state system is the question marked as "First Question" in the return questionnaire. If there is a question before "First Question", the background automatically generates the corresponding question answer based on logic.

[0056] In an embodiment of the present application, the questionnaire configuration data also includes a background call interface set for a target questionnaire question among multiple questionnaire questions, and a target question answer that triggers the background call interface. The method also includes: when it is determined that the response data of the follow-up object to the target questionnaire question is the answer to the target question, triggering the background call interface to execute the corresponding background service operation.

[0057] The background operation configuration can configure a certain question or a certain option of a question, and then the system background needs to execute the operation API. For example, the question asks "Have our staff contacted you about the equipment repair problem you just reported?" If the customer answers "No contact", the system needs to automatically call the interface and generate a telephone reminder for the corresponding service engineer.

[0058] Through the above technical solutions, a large number of outbound call-back tasks for work orders can be automatically completed, saving labor costs and improving the efficiency of the call-back. The semantic understanding technology based on the large language model can accurately and real-time understand customer responses without a lot of manual annotation and training. It can jump to the corresponding subsequent questions in real time based on the answer to the current question to continue the inquiry, ensuring the accuracy of the questionnaire results and the flexibility of the questionnaire questions. The voice outbound call robot can flexibly respond to the customer's statement based on the large model, and supports customers to interrupt the reading of questions. It is highly interactive and can enhance the enthusiasm of customers to participate in the call-back, thereby improving the efficiency of the call-back results. The call-back questionnaire supports multiple question types such as single-choice questions, multiple-choice questions, fill-in-the-blank questions, and question-and-answer questions. It can accurately obtain answers to multiple question types from user responses based on the large voice model to meet the needs of questionnaire call-backs in different scenarios. At the same time, it supports the configuration of the call-back questionnaire and the call-back task configuration. No background coding is required. Through page configuration, it can be quickly adjusted according to actual needs to adapt to different business scenarios.

[0059] Figure 1 FIG. 1 is a flow chart of a voice outbound calling method for questionnaire return interview in one embodiment. It should be understood that although Figure 1 The steps in the flowchart are shown in sequence as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified in this document, there is no strict order restriction for the execution of these steps, and these steps can be executed in other orders. Moreover, Figure 1 At least part of the steps may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least part of the sub-steps or stages of other steps.

[0060] Figure 3 The structure block diagram of a voice outbound calling device according to an embodiment of the present application is schematically shown. Figure 3 As shown, the embodiment of the present application provides a voice outbound call device, which may include:

[0061] a memory configured to store instructions;

[0062] The processor is configured to call instructions from the memory and implement the above-mentioned voice outbound calling method for questionnaire return interview when executing the instructions.

[0063] Specifically, in the embodiment of the present application, the processor may be configured to:

[0064] Obtain the questionnaire configuration data set by the user on the first configuration page corresponding to any type of work order;

[0065] Generate a callback questionnaire corresponding to the work order according to the questionnaire configuration data, the callback questionnaire including a plurality of questionnaire questions with jump logic;

[0066] When the status of the work order is completed, an outbound call back task corresponding to each work order is generated according to the back-call questionnaire;

[0067] Call the object of the outbound call back task, and convert each questionnaire question in the questionnaire corresponding to the outbound call back task into a voice question;

[0068] Collect the response data of the interviewees to each voice question in real time;

[0069] The jump logic of each questionnaire question is executed according to the response data of each voice question until jumping to any end question of the follow-up questionnaire to complete the outbound call follow-up task.

[0070] In an embodiment of the present application, the processor may also be configured to:

[0071] The questionnaire configuration data includes prompt words and question types for each questionnaire question, as well as multiple preset question answers. The jump logic of each questionnaire question is executed according to the reply data of each voice question, including: the large language model analyzes the reply data of each voice question according to the prompt words and question type of each questionnaire question to output the answer to the voice question; determines whether the answer to each voice question output by the large language model matches any one of the multiple preset question answers corresponding to the voice question; when the answer to each voice question matches any one of the corresponding multiple preset question answers, the jump logic of the questionnaire question corresponding to the voice question is executed according to the answer to each voice question.

[0072] In an embodiment of the present application, the processor may also be configured to:

[0073] The large language model analyzes the response data of each voice question according to the prompt words and question type of each questionnaire question to output the answer to the question of the voice question, including: when the question type is a multiple-choice question, the large language model semantically matches the response data of each voice question with multiple preset question answers corresponding to the voice question according to the prompt words of each questionnaire question to output the target question answer that matches the response data of the voice question, and uses the target question answer as the answer to the question of the voice question; when the question type is a fill-in-the-blank question, the large language model determines the keywords in the response data of each voice question according to the prompt words of each questionnaire question, and generates the answer to the question of the voice question according to the keywords; when the question type is a question-and-answer question, the large language model determines whether the response data of each voice question matches the voice question according to the prompt words of each questionnaire question to obtain a matching result, and uses the matching result as the answer to the question of the voice question.

[0074] In an embodiment of the present application, the processor may also be configured to:

[0075] Generating the answer to the voice question based on the keywords includes: generating a key-value pair based on the keywords and the blank part of each questionnaire title, so as to use the key-value pair as the answer to the voice question.

[0076] In an embodiment of the present application, the processor may also be configured to:

[0077] The questionnaire configuration data also includes whether to repeat the question for each questionnaire question, the threshold of the number of questions, and the threshold of the waiting time. The method also includes: when the answer to each voice question does not match any of the preset questions among the multiple preset answers, and it is determined that the questionnaire question corresponding to the voice question is to be repeated, the voice question is repeatedly played within the threshold of the number of questions; when the waiting time for collecting the answer to each voice question is greater than the threshold of the waiting time, jump to any end question of the follow-up questionnaire corresponding to the voice question to complete the outbound call follow-up task.

[0078] In an embodiment of the present application, the processor may also be configured to:

[0079] The questionnaire configuration data includes the question type, jump logic, prompt words, multiple end questions, inquiry templates and insertion variables corresponding to each questionnaire question, wherein the insertion variables are generated according to the work order information of each work order.

[0080] In an embodiment of the present application, the processor may also be configured to:

[0081] Calling the callback object includes: obtaining the task configuration data set by the user on the second configuration page corresponding to the outbound call callback task, the task configuration data including the execution time period, outbound call rules and task priority of the outbound call callback task; initiating a callback operation on the callback object according to the task configuration data.

[0082] In an embodiment of the present application, the processor may also be configured to:

[0083] The questionnaire configuration data also includes a background call interface set for a target questionnaire topic among multiple questionnaire topics, and a target question answer that triggers the background call interface. The method also includes: when it is determined that the response data of the follow-up object to the target questionnaire topic is the target question answer, triggering the background call interface to execute the corresponding background service operation.

[0084] An embodiment of the present application also provides a machine-readable storage medium, on which instructions are stored, and the instructions are used to enable a machine to execute the above-mentioned voice outbound calling method for questionnaire follow-up.

[0085] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 4As shown. The computer device includes a processor A01, a network interface A02, a memory (not shown in the figure) and a database (not shown in the figure) connected through a system bus. Among them, the processor A01 of the computer device is used to provide computing and control capabilities. The memory of the computer device includes an internal memory A03 and a non-volatile storage medium A04. The non-volatile storage medium A04 stores an operating system B01, a computer program B02 and a database (not shown in the figure). The internal memory A03 provides an environment for the operation of the operating system B01 and the computer program B02 in the non-volatile storage medium A04. The database of the computer device is used to store voice outbound call data for questionnaire return visits. The network interface A02 of the computer device is used to communicate with an external terminal through a network connection. When the computer program B02 is executed by the processor A01, a voice outbound call method for questionnaire return visits is implemented.

[0086] Those skilled in the art will understand that Figure 4 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0087] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

[0088] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0089] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0090] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0091] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0092] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0093] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0094] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0095] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A voice outbound calling method for questionnaire return interview, characterized in that: The method comprises: Obtain the questionnaire configuration data set by the user on the first configuration page corresponding to any type of work order; generating a callback questionnaire corresponding to the work order according to the questionnaire configuration data, wherein the callback questionnaire includes a plurality of questionnaire questions with jump logic; When the work order is in a completed state, generating an outbound call back task corresponding to each work order according to the back-call questionnaire; Calling the object of the outbound call back task, and converting each questionnaire question in the back questionnaire corresponding to the outbound call back task into a voice question; Collecting the response data of the interviewee to each voice question in real time; The jump logic of each questionnaire question is executed according to the reply data of each voice question until jumping to any end question of the return questionnaire to complete the outbound call return task.

2. The voice outbound calling method for questionnaire return interview according to claim 1, characterized in that: The questionnaire configuration data includes prompt words and question types for each questionnaire question, as well as multiple preset question answers. The jump logic for each questionnaire question is executed according to the reply data of each voice question, including: The large language model analyzes the response data of each voice question according to the prompt words and question type of each questionnaire question to output the answer to the question asked by the voice; Determine whether the answer to each voice question output by the large language model matches any one of a plurality of preset question answers corresponding to the voice question; When the answer to each voice question matches any one of the corresponding multiple preset question answers, the jump logic of the questionnaire question corresponding to the voice question is executed according to the answer to each voice question.

3. The voice outbound calling method for questionnaire return interview according to claim 2 is characterized in that: The large language model analyzes the response data of each voice question according to the prompt word and question type of each questionnaire question to output the answer to the voice question including: In the case where the question type is a multiple-choice question, the large language model semantically matches the response data of each voice question with a plurality of preset question answers corresponding to the voice question according to the prompt words of each questionnaire question, so as to output a target question answer matching the response data of the voice question, and use the target question answer as the question answer of the voice question; In the case where the question type is a fill-in-the-blank question, the large language model determines the keywords in the reply data of each voice question according to the prompt words of each questionnaire question, and generates the answer to the voice question according to the keywords; In the case where the question type is a question-and-answer question, the large language model determines whether the response data of each voice question matches the voice question based on the prompt words of each questionnaire question to obtain a matching result, and uses the matching result as the answer to the voice question.

4. The voice outbound calling method for questionnaire return interview according to claim 3 is characterized in that: The generating of the answer to the voice question according to the keyword comprises: A key-value pair is generated according to the keyword and the blank part of each questionnaire question, so as to use the key-value pair as the answer to the voice question.

5. The voice outbound calling method for questionnaire return interview according to claim 2, characterized in that: The questionnaire configuration data also includes whether to repeat the inquiry for each questionnaire question, the inquiry number threshold, and the waiting time threshold. The method also includes: If the answer to each voice question does not match any of the preset questions among the multiple preset answers, and it is determined to perform repeated questioning for the questionnaire question corresponding to the voice question, repeatedly play the voice question within the threshold of the number of questions; When the waiting time for collecting answers to each voice question is longer than the waiting time threshold, the system jumps to any ending question of the follow-up questionnaire corresponding to the voice question to complete the outbound call follow-up task.

6. The voice outbound calling method for questionnaire return interview according to claim 1, characterized in that: The questionnaire configuration data includes a question type, jump logic, prompt words, multiple end questions, inquiry templates, and insertion variables corresponding to each questionnaire question, wherein the insertion variables are generated according to the work order information of each work order.

7. The voice outbound calling method for questionnaire return interview according to claim 1, characterized in that: The calling the return visit object comprises: Acquire task configuration data set by the user on a second configuration page corresponding to the outbound call return task, wherein the task configuration data includes an execution period, an outbound call rule, and a task priority of the outbound call return task; A return visit operation is initiated on the return visit object according to the task configuration data.

8. The voice outbound calling method for questionnaire return interview according to claim 1, characterized in that: The questionnaire configuration data also includes a background calling interface set for a target questionnaire topic among the multiple questionnaire topics, and a target question answer that triggers the background calling interface. The method also includes: When it is determined that the response data of the return interview object to the target questionnaire question is the answer to the target question, the background calling interface is triggered to execute the corresponding background service operation.

9. A voice outbound calling device, characterized in that: include: a memory configured to store instructions; A processor is configured to call the instructions from the memory and implement the voice outbound calling method for questionnaire return interview according to any one of claims 1 to 7 when executing the instructions.

10. A machine-readable storage medium, characterized in that: The machine-readable storage medium stores instructions for causing the machine to execute the voice outbound calling method for questionnaire follow-up according to any one of claims 1 to 8.

Images