Man-machine question and answer method and device, computer equipment and storage medium

Abstract

The invention relates to a man-machine question and answer method and device, computer equipment and a storage medium. The method comprises the steps of receiving a first statement input by a user; obtaining a standard processing workflow and order information according to the first statement; generating a first verbal skill of the first statement according to the standard processing workflow and the order information; receiving a second statement input by the user for the first verbal skill; determining a current emotion corresponding to the user according to the second statement; if the current emotion meets the target emotion, acquiring a current session identifier; sending the current session identifier to a device where an artificial customer service is located, obtaining current session content through the device according to the current session identifier, and generating an artificial reply talk of a second statement according to the session content; and displaying the manual reply verbal skill. By adopting the method, the man-machine question-answering efficiency can be improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and particularly to a human-machine question-and-answer method, device, computer device, and storage medium. Background Art

[0002] Online car-hailing is a complex system. Limited by the current technical level and laws and regulations, human drivers are currently the main force for online car-hailing drivers. Human drivers may encounter various unexpected situations when performing pick-up, drop-off, and other tasks. Therefore, an efficient customer service system is necessary.

[0003] The current mainstream online car-hailing customer service solution is that when a driver encounters a problem, they first click the "Help" button on the application to enter the page where the robot customer service is located. The robot customer service generally displays some recommended preset answers. When the driver solves the problem or finds the answer in the robot customer service, they generally exit the page where the robot customer service is located. The driver will choose to click the "Transfer to Human" button on the page to enter the page where the human customer service is located. The robot customer service and the human customer service are designed with different pages respectively. In this way, when the robot customer service cannot solve the problem, the driver still needs to exit the current page and re-enter the page where the human customer service is located, which is complicated to operate and results in low efficiency. Summary of the Invention

[0004] Based on this, in view of the above technical problems, it is necessary to provide a human-machine question-and-answer method, device, computer device, and storage medium that can improve the efficiency of human-machine question and answer.

[0005] A human-machine question-and-answer method, the method includes:

[0006] Receiving a first statement input by a user;

[0007] Obtaining a standard processing workflow and order information according to the first statement;

[0008] Generating a first response for the first statement according to the standard processing workflow and order information;

[0009] Receiving a second statement input by the user for the first response;

[0010] Determining the current emotion of the user according to the second statement;

[0011] If the current emotion meets the target emotion, obtaining the current session identifier;

[0012] Sending the current session identifier to the device where the human customer service is located, and the device obtaining the current session content according to the current session identifier, and generating an artificial response for the second statement according to the session content;

[0013] Displaying the artificial response.

[0014] In one embodiment, obtaining a standard processing workflow and order information according to the first statement includes:

[0015] Performing key information extraction on the first statement to obtain statement keywords;

[0016] Obtaining a standard processing workflow according to the statement keywords;

[0017] Obtaining a user identifier corresponding to the user;

[0018] Obtaining an order identifier and an order status according to the user identifier.

[0019] In one embodiment, generating a first script for the first statement according to the standard processing workflow and order information includes:

[0020] Sorting out the standard processing workflow to generate a process handling script;

[0021] Fusing the order information and the process handling script to generate a first script for the first statement.

[0022] In one embodiment, determining the current emotion corresponding to the user according to the second statement includes:

[0023] Performing intention recognition on the second statement to obtain statement intention;

[0024] Performing tone recognition on the second statement to obtain the user's tone;

[0025] Determining the current emotion corresponding to the user according to the user's tone and statement intention.

[0026] In one embodiment, before determining the current emotion corresponding to the user according to the second statement, it further includes:

[0027] Performing keyword extraction on the second statement to obtain second statement keywords;

[0028] Searching for a target workflow from the standard processing workflow according to the second statement keywords;

[0029] Generating a second script corresponding to the second statement according to the target workflow;

[0030] Displaying the second script.

[0031] In one embodiment, sending the current session identifier to the device where the human customer service is located, and the device obtains the current session content according to the current session identifier, and generates an artificial reply script for the second statement, including:

[0032] The current session identifier is sent to the device where the manual customer service is located, so that the device enters the page where the current session is located according to the current session identifier and displays the current session content corresponding to the current session on the page;

[0033] Receive manual reply scripts. Manual reply scripts are generated by the device's manual customer service based on the current conversation content and the reply editing operation.

[0034] In one embodiment, after receiving the first sentence input by the user, the method further includes:

[0035] Obtain a recommended question and answer list, the recommended question and answer list including at least one recommended sentence and a corresponding recommended answering script;

[0036] Determine a target recommended sentence from the recommended question and answer list according to the first sentence;

[0037] Get the target recommendation reply words corresponding to the target recommendation sentence;

[0038] Determine the target recommended reply speech as the first speech corresponding to the first sentence;

[0039] If the target recommended sentence cannot be determined from the recommended question and answer list according to the first sentence, the step of obtaining the standard processing workflow and order information according to the first sentence is entered.

[0040] A human-machine question-answering device, comprising:

[0041] A first receiving module, used for receiving a first sentence input by a user;

[0042] An acquisition module, used for acquiring standard processing workflow and order information according to the first statement;

[0043] A generation module, used for generating a first speech of a first sentence according to a standard processing workflow and order information;

[0044] A second receiving module is used to receive a second sentence input by the user for the first speech technique;

[0045] A processing module, used to determine the current emotion corresponding to the user according to the second sentence;

[0046] A detection module, used to obtain a current session identifier if the current emotion satisfies the target emotion;

[0047] A sending module is used to send the current session identifier to the device where the manual customer service is located, obtain the current session content through the device according to the current session identifier, and generate a manual reply script of the second sentence according to the session content;

[0048] Display module, used to display manual response scripts.

[0049] A computer device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein when the processor executes the computer program, the following steps are implemented:

[0050] Receiving a first statement input by a user;

[0051] Obtaining a standard processing workflow and order information according to the first statement;

[0052] Generating a first speech script for the first statement according to the standard processing workflow and order information;

[0053] Receiving a second statement input by the user for the first speech script;

[0054] Determining the current emotion corresponding to the user according to the second statement;

[0055] If the current emotion meets the target emotion, obtaining the current session identifier;

[0056] Sending the current session identifier to the device where the artificial customer service is located, obtaining the current session content by the device according to the current session identifier, and generating an artificial reply speech script for the second statement according to the session content;

[0057] Displaying the artificial reply speech script.

[0058] A computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:

[0059] Receiving a first statement input by a user;

[0060] Obtaining a standard processing workflow and order information according to the first statement;

[0061] Generating a first speech script for the first statement according to the standard processing workflow and order information;

[0062] Receiving a second statement input by the user for the first speech script;

[0063] Determining the current emotion corresponding to the user according to the second statement;

[0064] If the current emotion meets the target emotion, obtaining the current session identifier;

[0065] Sending the current session identifier to the device where the artificial customer service is located, obtaining the current session content by the device according to the current session identifier, and generating an artificial reply speech script for the second statement according to the session content;

[0066] Displaying the artificial reply speech script.

[0067] The above-mentioned human-computer question-and-answer method, device, computer equipment and storage medium receive a first sentence input by a user; obtain a standard processing workflow and order information based on the first sentence; generate a first speech for the first sentence based on the standard processing workflow and order information; receive a second sentence input by the user for the first speech; determine the current emotion corresponding to the user based on the second sentence; if the current emotion meets the target emotion, obtain the current session identifier; send the current session identifier to the device where the manual customer service is located, obtain the current session content through the device based on the current session identifier, generate a manual reply speech for the second sentence based on the session content, and display the manual reply speech.

[0068] Therefore, the human-machine automatic question and answer and the manual customer service question and answer are designed on the same page. The first speech corresponding to the first sentence input by the user is obtained through the human-machine automatic question and answer. When the first speech cannot solve the user's difficulty and it is felt that the user is angry and anxious when responding to the second sentence input by the first speech, the human-machine customer service can be automatically switched to appease and solve the user's difficulty through manual customer service. In this way, the efficiency of human-machine answering can be improved, and the same set of pages can be designed for simple operation. BRIEF DESCRIPTION OF THE DRAWINGS

[0069] Figure 1 is an application environment diagram of a human-computer question answering method in an embodiment;

[0070] Figure 2 Schematic diagram of a flow chart of a human-computer question-answering method in one embodiment;

[0071] Figure 3 A schematic diagram of a process flow of obtaining steps of a standard processing workflow in one embodiment;

[0072] Figure 4 A schematic diagram of a process of generating a first speech technique in an embodiment;

[0073] Figure 5 A schematic diagram of a flow chart of a current emotion determination step in one embodiment;

[0074] Figure 6 is a flow chart of the second statement processing step in one embodiment;

[0075] Figure 7 A schematic diagram of a flow chart of a step of sending a current session identifier in an embodiment;

[0076] Figure 8 Schematic diagram of a flow chart of a human-computer question-answering method in one embodiment;

[0077] Figure 9 is a structural block diagram of a human-machine question-answering device in one embodiment;

[0078] Figure 10Internal structure diagram of a computer device in an embodiment;

[0079] Figure 11 Internal structure diagram of a computer device in an embodiment. Detailed implementation manners

[0080] In order to make the objectives, technical solutions and advantages of the present application clearer and more understandable, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0081] The human-machine question-answering method provided by the present application can be applied to, for example Figure 1 the application environment shown. Among them, the terminal 102 and the device 106 communicate with the server 104 through the network respectively. Among them, the terminal 102 and the device 106 can be, but are not limited to, various personal computers, laptop computers, smart phones, tablet computers and portable wearable devices, and the server 104 can be implemented by an independent server or a server cluster composed of multiple servers.

[0082] Specifically, the terminal 102 receives the first statement input by the user and sends it to the server 104. The server 104 obtains the standard processing workflow and order information according to the first statement, generates the first speech script of the first statement according to the standard processing workflow and order information, and sends it to the terminal 102 for display.

[0083] Further, the terminal 102 receives the second statement input by the user for the first speech script and sends it to the server 104. The server 104 determines the current emotion corresponding to the user according to the second statement. If the current emotion meets the target emotion, the current session identifier is obtained, and the current session identifier is sent to the device 106 where the artificial customer service is located. The device 106 obtains the current session content according to the current session identifier, generates the artificial reply speech script of the second statement according to the session content, and sends it to the terminal 102 through the server 104. Finally, the terminal 102 displays the artificial reply speech script.

[0084] In one embodiment, as Figure 2 shown, a human-machine question-answering method is provided. Taking the method applied to the Figure 1 server as an example for illustration, it includes the following steps:

[0085] Step 202: Receive the first statement input by the user.

[0086] Step 204: Obtain the standard processing workflow and order information according to the first statement.

[0087] Among them, the first statement here is the statement input by the user, and the user inputs the first statement on the human-machine Q&A page. For example, the first statement is: Help me cancel the order.

[0088] Among them, the standard processing workflow SOP is the capitalization of the first letters of the three words Standard Operating Procedure, which refers to describing the standard operation steps and requirements of an event in a unified format for guiding and standardizing daily work.

[0089] Among them, the order information is the information related to the user's corresponding order, including but not limited to order identification, order status, order details, etc.

[0090] Specifically, the user inputs the first statement on the human-machine Q&A page and the first statement is displayed on the human-machine Q&A page. After receiving the first statement, the standard processing workflow and order information matching the first statement are obtained. Specifically, the matching standard processing workflow can be obtained from the standard processing workflow set according to the first statement, and this standard processing workflow is the work processing flow operation and standard related to the first statement. For the order information, the user identification corresponding to the input first statement can be obtained, and the order information can be obtained according to the user identification.

[0091] Step 206, generate the first speech for the first statement according to the standard processing workflow and order information.

[0092] Among them, the first speech is the reply speech corresponding to the first statement, which is generated according to the standard processing workflow and order information, and includes the process processing operation standards, content in the standard processing workflow, and order information.

[0093] Specifically, after obtaining the standard processing workflow, perform text content parsing on the standard processing workflow, extract the key information in the standard processing workflow, and fuse the key information with the order information to generate the first speech corresponding to the first statement.

[0094] For example, the standard processing workflow is:

[0095] If the user wants to cancel the order, first judge whether the driver currently has an order, inform the user of the risk of responsibility judgment, prompt the user of the risk of responsibility judgment, and if the user still hopes that the customer service cancels, ask for the reason for cancellation and help the user cancel.

[0096] Among them, the order information is:

[0097] Order identification: 233333, order status: in progress.

[0098] Among them, the key information extracted from the standard processing workflow is:

[0099] We have detected an order for you. You can cancel it automatically, but the cancellation may be punished by the system.

[0100] Furthermore, the key information is integrated with the order information to generate the first statement corresponding to the first dialogue: We have found that you have an ongoing order 233333. You can cancel it automatically, but the cancellation of the order may be judged by the system.

[0101] Step 208, receiving a second statement input by the user in response to the first speech script.

[0102] Step 210: Determine the current emotion corresponding to the user according to the second sentence.

[0103] Among them, the second sentence here is the sentence input by the user in response to the first speech. After the first speech is generated, it is displayed on the human-computer question and answer page. If the user finds that the first speech cannot solve the user's current problem, he will enter the second sentence. Conversely, if the user finds that the first speech can solve the user's current problem, the human-computer question and answer will be stopped.

[0104] The current emotion is the emotion carried by the second sentence input by the user, and the emotion of the user when inputting the second sentence can be determined by identifying the content, tone, and intention of the second sentence. The emotion here is the psychological state generated by the user inputting the second sentence, such as excitement, anger, sadness, etc.

[0105] Step 212: If the current emotion satisfies the target emotion, the current session identifier is obtained.

[0106] Among them, the target emotion can be determined according to actual business needs, actual product needs or actual application scenarios, and is the emotion used to automatically transfer to manual customer service. For example, the target emotion is anger. When it is detected that the user's current emotion when entering the second sentence is anger, it means that the manual automatic question and answer can no longer meet the user's requirements and manual intervention is required, then the current session identifier is obtained.

[0107] The current session identifier is a unique identifier for the current session. The current session includes at least one round of dialogue between the user and the machine, including a first sentence and a first speech corresponding to the first sentence. Different sessions have different corresponding session identifiers.

[0108] Step 214, the current session identifier is sent to the device where the manual customer service is located, the device obtains the current session content according to the current session identifier, and generates a manual reply script for the second sentence according to the session content.

[0109] Step 216, display the manual response script.

[0110] When it is detected that the user's current emotion meets the target emotion, it means that the human-machine automatic question and answer can no longer solve the user's problem, and manual intervention is required to switch from human-machine automatic question and answer to manual customer service question and answer. Specifically, the current session identifier is sent to the device where the manual customer service is located, and the current session corresponding to the current session identifier is displayed on the device where the manual customer service is located. After understanding all the contents of the current session, the manual reply script editing operation is performed, and the manual reply script is obtained according to the manual reply script editing operation, and the manual reply script is returned and displayed.

[0111] For example, the manual reply line is: Hello, I have understood your situation, but according to the platform regulations, I really cannot cancel the order. I’m sorry and thank you for your understanding.

[0112] In the above-mentioned human-computer question-and-answer method, a first sentence input by a user is received; a standard processing workflow and order information are obtained according to the first sentence; a first speech of the first sentence is generated according to the standard processing workflow and order information; a second sentence input by the user for the first speech is received; the current emotion corresponding to the user is determined according to the second sentence; if the current emotion meets the target emotion, the current session identifier is obtained; the current session identifier is sent to the device where the manual customer service is located, the current session content is obtained through the device according to the current session identifier, the manual reply speech of the second sentence is generated according to the session content, and the manual reply speech is displayed.

[0113] Therefore, the human-machine automatic question and answer and the manual customer service question and answer are designed on the same page. The first speech corresponding to the first sentence input by the user is obtained through the human-machine automatic question and answer. When the first speech cannot solve the user's difficulty and it is felt that the user is angry and anxious when responding to the second sentence input by the first speech, the human-machine customer service can be automatically switched to appease and solve the user's difficulty through manual customer service. In this way, the efficiency of human-machine answering can be improved, and the same set of pages can be designed for simple operation.

[0114] In one embodiment, Figure 3 As shown, the standard processing workflow and order information are obtained according to the first statement, including:

[0115] Step 302: extract key information from the first sentence to obtain sentence keywords.

[0116] Step 304, obtaining a standard processing workflow according to the sentence keywords.

[0117] Step 306: Obtain the user ID corresponding to the user.

[0118] Step 308: Obtain the order ID and order status according to the user ID.

[0119] Among them, obtaining the standard processing workflow and order information according to the first sentence can specifically include extracting keywords in the first sentence, obtaining sentence keywords, obtaining a standard processing workflow set, and the standard processing workflow set includes at least one standard processing workflow, and each standard processing workflow is associated with a processing work type. Therefore, after obtaining the sentence keywords, the sentence processing work type corresponding to the first sentence is determined according to the sentence keywords, and a matching standard processing workflow is found from the standard processing workflow set according to the sentence processing work type and the processing work type.

[0120] Further, obtain the user identifier corresponding to the user input in the first statement, such as the user ID, and then search for the order identifier corresponding to the user identifier based on the user identifier, and then obtain the matching order based on the order identifier, and obtain the order status of the order, which includes order completion, order in progress, and non-existent order.

[0121] In one embodiment, Figure 4 As shown, the first script for generating the first sentence according to the standard processing workflow and order information includes:

[0122] Step 402, organize the standard processing workflow and generate process processing scripts.

[0123] Step 404, the order information and the process processing script are integrated to generate the first script of the first sentence.

[0124] Among them, since the standard processing workflow is a description of the standard operating steps and requirements of a certain event in a unified format to guide and standardize daily work, it is necessary to sort out the standard processing workflow into corresponding process processing words, and the process processing words can reflect the standard processing workflow. Furthermore, the order information is integrated with the process processing words to generate the first words corresponding to the first sentence.

[0125] For example, the standard processing workflow is:

[0126] If the user wants to cancel an order, the customer service will first determine whether the driver currently has any orders, inform the user that there is a risk of being held liable, and remind the user that there is a risk of being held liable. If the user still wants the customer service to cancel the order, the customer service will ask the reason for the cancellation and help the user cancel the order.

[0127] The order information is:

[0128] Order ID: 233333, Order status: in progress.

[0129] The generated process processing words are:

[0130] We have detected an order for you. You can cancel it automatically, but the cancellation may be punished by the system.

[0131] Further, the process handling words are integrated with the order information to generate the first words corresponding to the first statement as follows:

[0132] We have found that you have an ongoing order 233333. You can cancel it automatically, but canceling the order may result in being held responsible by the system.

[0133] In one embodiment, as Figure 5 shown, determining the current emotion corresponding to the user according to the second statement includes:

[0134] Step 502: Perform intent recognition on the second statement to obtain the statement intent.

[0135] Step 504: Perform tone recognition on the second statement to obtain the user's tone.

[0136] Step 506: Determine the current emotion corresponding to the user according to the user's tone and the statement intent.

[0137] Among them, the intent is the purpose that the user hopes to achieve. After obtaining the second statement, parse the second statement to obtain the statement intent corresponding to the second statement. Through the statement intent, the user's purpose can be judged.

[0138] Further, tone recognition can also be performed on the second statement to recognize the tone when the user inputs the second statement to obtain the user's tone. Finally, determine the emotion when the user inputs the second statement, that is, the current emotion, through the user's tone and the statement intent.

[0139] For example, the second statement input by the user is:

[0140] I've already said it. Why am I being held responsible? I've already said that the tire is punctured. Hurry up and transfer me to an operator.

[0141] Among them, it can be determined that the current emotion of the user at this time is: anger.

[0142] In one embodiment, as Figure 6 shown, before determining the current emotion corresponding to the user according to the second statement, it further includes:

[0143] Step 602: Extract keywords from the second statement to obtain the second statement keywords.

[0144] Step 604: Search for the target work process from the standard processing work process according to the second statement keywords.

[0145] Step 606: Generate the second words corresponding to the second statement according to the target work process.

[0146] Step 608: Display the second words.

[0147] Among them, after receiving the second statement input by the user for the first conversation statement, if the second conversation statement corresponding to the second statement can be generated according to the standard processing workflow, the man-machine automatic question and answer can continue; if not, manual intervention is required.

[0148] Specifically, after receiving the second statement input by the user, the text content of the second statement is split to obtain multiple words, and keywords are extracted from the multiple words to obtain the second statement keywords.

[0149] Furthermore, the target workflow is searched from the standard processing workflow through the second statement keywords. Here, the target workflow is the workflow matched by the second statement keywords. The target workflow is sorted out to generate the second conversation statement corresponding to the second statement, and the second conversation statement is displayed to the user.

[0150] For example, the first statement input by the user is: Help me cancel the order.

[0151] The standard processing workflow is:

[0152] If the user wants to cancel the order, first judge whether the driver currently has an order, inform the user of the risk of responsibility judgment, prompt the user of the risk of responsibility judgment. If the user still hopes that the customer service cancels, ask the reason for cancellation and help the user cancel.

[0153] Among them, the first conversation statement corresponding to the first statement is generated as:

[0154] We have found that you have an ongoing order 233333. You can cancel it automatically, but canceling the order may be judged responsible by the system.

[0155] The second statement input by the user for the first conversation statement is: I got a flat tire.

[0156] The generated second conversation statement is: I see. Your tire is flat and you can't pick up passengers. Do you want me to help you cancel it? Note that there is still a risk of responsibility judgment if the customer service cancels. Please keep the evidence well and appeal in time if you receive a responsibility judgment notice.

[0157] In one embodiment, as Figure 7 shown, the current session identifier is sent to the device where the artificial customer service is located. The device obtains the current session content according to the current session identifier, and generates the artificial reply conversation statement of the second statement according to the session content, including:

[0158] Step 702, send the current session identifier to the device where the artificial customer service is located, so that the device enters the page where the current session is located according to the current session identifier, and displays the current session content corresponding to the current session on the page.

[0159] Step 704, receiving manual reply words, the manual reply words are generated by the manual customer service of the device according to the reply editing operation based on the current conversation content.

[0160] Among them, when the emotion of the user when entering the second sentence has met the target emotion, for example, when the emotion of the user when entering the second sentence is anger, it means that the human-computer automatic question and answer can no longer satisfy the user and cannot solve the user's difficulties. Therefore, a manual customer service response is required.

[0161] Specifically, the current session identifier is sent to the device where the manual customer service is located. The current session identifier is used to identify the current session. The current session corresponding to the current session identifier is displayed on the device where the manual customer service is located. After understanding all the contents of the current session, the manual reply speech editing operation is performed, and the manual reply speech is obtained according to the manual reply speech editing operation, and the manual reply speech is returned and displayed.

[0162] For example, the manual reply line is: Hello, I have understood your situation, but according to the platform regulations, I really cannot cancel the order. I’m sorry and thank you for your understanding.

[0163] In one embodiment, Figure 8 As shown, after receiving the first sentence input by the user, it also includes:

[0164] Step 802, obtaining a recommended question and answer list, the recommended question and answer list including at least one recommended sentence and a corresponding recommended answering script.

[0165] Step 804: determine a target recommended sentence from the recommended question and answer list according to the first sentence.

[0166] Step 806, obtaining the target recommendation reply words corresponding to the target recommendation sentence.

[0167] Step 808, determining the target recommended reply speech as the first speech corresponding to the first sentence.

[0168] Step 810: If the target recommended sentence cannot be determined from the recommended question and answer list according to the first sentence, then the step of obtaining the standard processing workflow and order information according to the first sentence is entered.

[0169] Among them, in order to improve the efficiency of human-computer question and answer, a recommended question and answer list can be set up, and the recommended question and answer list includes at least one recommended sentence and a corresponding recommended reply language, which can be determined in advance according to actual business needs, actual product needs or actual application scenarios. When the first sentence input by the user hits the recommended question and answer list, the hit recommended reply language can be directly determined as the first language corresponding to the first sentence.

[0170] Specifically, a recommended question and answer list is obtained, and a target recommended sentence matching the first sentence in the recommended question and answer list is determined based on the similarity between the first sentence and each recommended sentence in the recommended question and answer list. Further, a target recommended reply speech corresponding to the target recommended sentence is obtained, and the target recommended reply speech is determined as the first speech corresponding to the first sentence.

[0171] Among them, if the first sentence cannot hit the recommended question and answer list, the target recommended sentence cannot be determined from the recommended question and answer list according to the first sentence, and then the step of obtaining the standard processing workflow and order information according to the first sentence is entered.

[0172] It should be understood that, although the various steps in the above-mentioned flow chart are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps is not strictly limited in order, and these steps can be executed in other orders. Moreover, at least a part of the steps in the above-mentioned flow chart may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these sub-steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0173] In one embodiment, Figure 9 As shown, a human-machine question-answering device 900 is provided, comprising: a first receiving module 902, an acquisition module 904, a generation module 906, a second receiving module 908, a processing module 910, a detection module 912, a sending module 914 and a display module 916, wherein:

[0174] The first receiving module 902 is configured to receive a first sentence input by a user.

[0175] The acquisition module 904 is used to acquire the standard processing workflow and order information according to the first statement.

[0176] The generation module 906 is used to generate a first script of the first sentence according to the standard processing workflow and order information.

[0177] The second receiving module 908 is used to receive a second sentence input by the user for the first speech.

[0178] The processing module 910 is used to determine the current emotion corresponding to the user according to the second sentence.

[0179] The detection module 912 is used to obtain the current session identifier if the current emotion satisfies the target emotion.

[0180] A sending module 914, configured to send the current session identifier to the device where the human customer service is located, so that the device obtains the current session content according to the current session identifier, and generates a human reply script for the second statement according to the session content.

[0181] A display module 916, configured to display the human reply script.

[0182] In one embodiment, the obtaining module 904 extracts key information from the first statement to obtain statement keywords, obtains a standard processing workflow according to the statement keywords, obtains a user identifier corresponding to the user, and obtains an order identifier and an order status according to the user identifier.

[0183] In one embodiment, the generating module 906 sorts out the standard processing workflow to generate a process handling script, and fuses the order information and the process handling script to generate a first script for the first statement.

[0184] In one embodiment, the processing module 910 identifies the intention of the second statement to obtain a statement intention, identifies the tone of the second statement to obtain the user tone, and the user tone and the statement intention determine the current emotion corresponding to the user.

[0185] In one embodiment, the human-machine question-and-answer device 900 extracts keywords from the second statement to obtain second statement keywords, searches for a target workflow from the standard processing workflow according to the second statement keywords, generates a second script corresponding to the second statement according to the target workflow, and displays the second script.

[0186] In one embodiment, the sending module 914 sends the current session identifier to the device where the human customer service is located, so that the device enters the page where the current session is located according to the current session identifier, displays the current session content corresponding to the current session on the page, and receives the human reply script, where the human reply script is generated by the human customer service of the device according to the current session content for a reply editing operation.

[0187] In one embodiment, the human-machine question-and-answer device 900 obtains a recommended question-and-answer list, where the recommended question-and-answer list includes at least one recommended statement and a corresponding recommended reply script, determines a target recommended statement from the recommended question-and-answer list according to the first statement, obtains a target recommended reply script corresponding to the target recommended statement, determines the target recommended reply script as the first script corresponding to the first statement, and if the target recommended statement cannot be determined from the recommended question-and-answer list according to the first statement, the obtaining module 904 obtains a standard processing workflow and order information according to the first statement.

[0188] For the specific limitations of the human-machine Q&A device, reference can be made to the limitations of the human-machine Q&A method in the above text, which will not be elaborated here. Each module in the above human-machine Q&A device can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in the processor of the computer device in hardware form or be independent of it, or be stored in the memory of the computer device in software form, so as to facilitate the processor to call and execute the operations corresponding to the above respective modules.

[0189] In one embodiment, a computer device is provided. The computer device can be a server, and its internal structure diagram can be as Figure 10 shown. The computer device includes a processor, a memory, a network interface, and a database connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store standard processing workflows. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements a human-machine Q&A method.

[0190] In one embodiment, a computer device is provided. The computer device can be a terminal, and its internal structure diagram can be as Figure 11 shown. The computer device includes a processor, a memory, a network interface, a display screen, and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it implements a human-machine Q&A method. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen. The input device of the computer device can be a touch layer covering the display screen, or a button, a trackball, or a touchpad provided on the housing of the computer device, or an external keyboard, a touchpad, or a mouse, etc.

[0191] Those skilled in the art can understand that Figure 10 or Figure 11 the structures shown in are only block diagrams of some structures related to the solution of this application, and do not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0192] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the following steps are implemented: receiving a first statement input by a user, obtaining a standard processing workflow and order information according to the first statement, generating a first speech script of the first statement according to the standard processing workflow and the order information, receiving a second statement input by the user for the first speech script, determining the current emotion corresponding to the user according to the second statement, if the current emotion meets the target emotion, obtaining the current session identifier, sending the current session identifier to the device where the human customer service is located, obtaining the current session content according to the current session identifier through the device, generating an artificial reply speech script for the second statement according to the session content, and displaying the artificial reply speech script.

[0193] In one embodiment, when the processor executes the computer program, the following steps are further implemented: extracting key information from the first statement to obtain statement keywords, obtaining a standard processing workflow according to the statement keywords, obtaining a user identifier corresponding to the user, and obtaining an order identifier and an order status according to the user identifier.

[0194] In one embodiment, when the processor executes the computer program, the following steps are further implemented: sorting out the standard processing workflow to generate a process handling speech script, and fusing the order information and the process handling speech script to generate a first speech script of the first statement.

[0195] In one embodiment, when the processor executes the computer program, the following steps are further implemented: identifying the intention of the second statement to obtain the statement intention, identifying the tone of the second statement to obtain the user tone, and determining the current emotion corresponding to the user according to the user tone and the statement intention.

[0196] In one embodiment, when the processor executes the computer program, the following steps are further implemented: extracting keywords from the second statement to obtain second statement keywords, searching for a target workflow from the standard processing workflow according to the second statement keywords, generating a second speech script corresponding to the second statement according to the target workflow, and displaying the second speech script.

[0197] In one embodiment, when the processor executes the computer program, the following steps are further implemented: sending the current session identifier to the device where the human customer service is located, so that the device enters the page where the current session is located according to the current session identifier, displays the current session content corresponding to the current session on the page, receives the artificial reply speech script, and the artificial reply speech script is generated by the human customer service of the device according to the current session content for reply editing operations.

[0198] In one embodiment, when the processor executes the computer program, the following steps are further implemented: obtaining a recommended question-and-answer list, where the recommended question-and-answer list includes at least one recommended statement and the corresponding recommended reply words, determining a target recommended statement from the recommended question-and-answer list according to the first statement, obtaining the target recommended reply words corresponding to the target recommended statement, determining the target recommended reply words as the first words corresponding to the first statement, and if the target recommended statement cannot be determined from the recommended question-and-answer list according to the first statement, then entering the step of obtaining the standard processing workflow and order information according to the first statement.

[0199] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented: receiving a first statement input by a user, obtaining a standard processing workflow and order information according to the first statement, generating first words of the first statement according to the standard processing workflow and order information, receiving a second statement input by the user for the first words, determining the current emotion corresponding to the user according to the second statement, if the current emotion meets the target emotion, obtaining the current session identifier, sending the current session identifier to the device where the artificial customer service is located, obtaining the current session content according to the current session identifier by the device, generating an artificial reply words for the second statement according to the session content, and displaying the artificial reply words.

[0200] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented: extracting key information from the first statement to obtain statement keywords, obtaining a standard processing workflow according to the statement keywords, obtaining a user identifier corresponding to the user, and obtaining an order identifier and an order status according to the user identifier.

[0201] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented: sorting out the standard processing workflow to generate process processing words, and fusing the order information and the process processing words to generate the first words of the first statement.

[0202] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented: identifying the intention of the second statement to obtain the statement intention, identifying the tone of the second statement to obtain the user tone, and determining the current emotion corresponding to the user according to the user tone and the statement intention.

[0203] In one embodiment, when the computer program is executed by a processor, the following steps are further implemented: extracting keywords from the second statement to obtain second statement keywords, searching for a target workflow from the standard processing workflow according to the second statement keywords, generating second words corresponding to the second statement according to the target workflow, and displaying the second words.

[0204] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: sending the current session identifier to the device where the manual customer service is located, so that the device enters the page where the current session is located according to the current session identifier, displays the current session content corresponding to the current session on the page, and receives manual reply words. The manual reply words are generated by the manual customer service of the device performing a reply editing operation based on the current session content.

[0205] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: obtaining a recommended question and answer list, the recommended question and answer list includes at least one recommended sentence and a corresponding recommended reply speech, determining a target recommended sentence from the recommended question and answer list according to a first sentence, obtaining a target recommended reply speech corresponding to the target recommended sentence, determining the target recommended reply speech as the first speech corresponding to the first sentence, and if the target recommended sentence cannot be determined from the recommended question and answer list according to the first sentence, then entering the step of obtaining a standard processing workflow and order information according to the first sentence.

[0206] Those of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing related hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct RAMbus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0207] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0208] The above-described embodiments merely represent several implementation manners of the present application. The description thereof is relatively specific and detailed, but it should not be construed as a limitation to the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, several modifications and improvements can still be made, and these all fall within the protection scope of the present application. Therefore, the protection scope of the patent of the present application shall be subject to the appended claims.

Claims

1. A human-computer question-answering method, the method comprising: receiving a first sentence input by a user; Acquire standard processing workflow and order information according to the first statement; Generate a first script for the first sentence according to the standard processing workflow and the order information; receiving a second sentence input by the user in response to the first speech technique; Determining a current emotion corresponding to the user according to the second sentence; If the current emotion satisfies the target emotion, obtaining the current session identifier; Sending the current session identifier to a device where a human customer service representative is located, obtaining the current session content through the device according to the current session identifier, and generating a manual reply script for the second sentence according to the session content; Demonstrate the manual response script.

2. The method according to claim 1, characterized in that The acquiring of standard processing workflow and order information according to the first statement includes: Extract key information from the first sentence to obtain sentence keywords; Obtaining a standard processing workflow based on the sentence keywords; Obtaining a user ID corresponding to the user; The order ID and order status are obtained according to the user ID.

3. The method according to claim 1, characterized in that The first speech technique of generating the first sentence according to the standard processing workflow and the order information includes: Arrange the standard processing workflow and generate process processing scripts; The order information and the process processing script are integrated to generate the first script of the first sentence.

4. The method according to claim 1, characterized in that: The determining the current emotion corresponding to the user according to the second sentence includes: Performing intent recognition on the second sentence to obtain sentence intent; Performing tone recognition on the second sentence to obtain the user's tone; The current emotion corresponding to the user is determined according to the user tone and the sentence intention.

5. The method according to claim 1, characterized in that Before determining the current emotion corresponding to the user according to the second sentence, the method further includes: Perform keyword extraction on the second sentence to obtain second sentence keywords; searching for a target workflow from the standard processing workflow according to the second sentence keyword; Generate a second speech script corresponding to the second sentence according to the target workflow; Demonstrate the second speech technique.

6. The method according to claim 1, characterized in that The sending of the current session identifier to a device where a human customer service representative is located, obtaining the current session content through the device according to the current session identifier, and generating a human reply script for the second sentence according to the session content includes: Sending the current session identifier to a device where the manual customer service is located, so that the device enters a page where the current session is located according to the current session identifier, and displays the current session content corresponding to the current session on the page; A manual reply speech is received, where the manual reply speech is generated by the manual customer service of the device performing a reply editing operation based on the current conversation content and the reply editing operation.

7. The method according to claim 1, characterized in that After receiving the first sentence input by the user, the method further includes: Obtaining a recommended question and answer list, wherein the recommended question and answer list includes at least one recommended sentence and a corresponding recommended answering script; Determining a target recommended sentence from the recommended question and answer list according to the first sentence; Obtain the target recommendation reply words corresponding to the target recommendation sentence; Determining the target recommended reply speech as the first speech corresponding to the first sentence; If the target recommended sentence cannot be determined from the recommended question and answer list according to the first sentence, the step of obtaining the standard processing workflow and order information according to the first sentence is entered.

8. A human-machine question-answering device, characterized in that: The device comprises: A first receiving module, used for receiving a first sentence input by a user; An acquisition module, used for acquiring standard processing workflow and order information according to the first statement; A generating module, configured to generate a first script of the first sentence according to the standard processing workflow and the order information; A second receiving module, used for receiving a second sentence input by the user in response to the first speech technique; A processing module, configured to determine a current emotion corresponding to the user according to the second sentence; A detection module, configured to obtain a current session identifier if the current emotion satisfies a target emotion; A sending module, configured to send the current session identifier to a device where a human customer service representative is located, obtain the current session content through the device according to the current session identifier, and generate a manual reply script for the second sentence according to the session content; A display module is used to display the manual reply script.

9. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images