WO2020034928A1 - 客服会话切换方法、系统和存储介质 - Google Patents

客服会话切换方法、系统和存储介质 Download PDF

Info

Publication number
WO2020034928A1
WO2020034928A1 PCT/CN2019/100288 CN2019100288W WO2020034928A1 WO 2020034928 A1 WO2020034928 A1 WO 2020034928A1 CN 2019100288 W CN2019100288 W CN 2019100288W WO 2020034928 A1 WO2020034928 A1 WO 2020034928A1
Authority
WO
WIPO (PCT)
Prior art keywords
emergency
customer service
request information
model
state
Prior art date
Application number
PCT/CN2019/100288
Other languages
English (en)
French (fr)
Inventor
刘峰
刘云峰
吴悦
胡晓
汶林丁
Original Assignee
深圳追一科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳追一科技有限公司 filed Critical 深圳追一科技有限公司
Publication of WO2020034928A1 publication Critical patent/WO2020034928A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00Commerce
    • G06Q30/06Buying, selling or leasing transactions
    • G06Q30/0601Electronic shopping [e-shopping]
    • G06Q30/0613Third-party assisted
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F40/00Handling natural language data
    • G06F40/20Natural language analysis
    • G06F40/279Recognition of textual entities

Definitions

  • the present application relates to the technical field of customer service robots, and in particular, to a method, system, and non-volatile computer-readable storage medium for customer service session switching.
  • customer service robots After receiving customer questions, customer service robots generally use candidate models to sort candidate common questions (FAQ, Frequently Asked Questions) to determine the relevance of customer questions to each FAQ, and then based on the confidence The degree model judges the confidence level of the answer corresponding to the FAQ ranked first. If the confidence in the first place is high enough, the answer corresponding to the first-place FAQ is returned directly to the customer, otherwise, N N-likely FAQs are recommended for customers to choose on their own.
  • FQ Frequently Asked Questions
  • a method and system for switching customer service sessions and a non-volatile computer-readable storage medium are provided.
  • a customer service session switching method includes:
  • a customer service session switching system includes a processor and a memory connected to the processor; the memory is used to store a computer program, and the computer program is used to perform at least the following operations:
  • One or more non-volatile computer-readable storage media storing computer-readable instructions, which when executed by one or more processors, cause the one or more processors to perform the following operations: Receive request information from customers;
  • FIG. 1 is an application environment diagram of a method for switching customer service sessions according to some embodiments.
  • FIG. 2 is a schematic flowchart of a method for switching customer service sessions according to some embodiments.
  • FIG. 3 is a schematic structural diagram of an emergency warning model according to some embodiments.
  • FIG. 4 is a schematic flowchart of a method for switching customer service sessions according to some embodiments.
  • FIG. 5 is a schematic flowchart of training an emergency warning model according to some embodiments.
  • FIG. 6 is a schematic structural diagram of a customer service session switching device according to some embodiments.
  • FIG. 7 is a schematic structural diagram of an access module according to some embodiments.
  • FIG. 8 is a schematic structural diagram of a model training module according to some embodiments.
  • FIG. 9 is a schematic structural diagram of a customer service session switching system according to some embodiments.
  • the method for switching customer service sessions provided in this application can be applied to the application environment shown in FIG. 1.
  • the application environment includes a computer device 102.
  • the computer device 102 may receive the request information sent by the client; send the request information to the pre-trained emergency warning model, wherein the emergency warning model is trained by the emergency vocabulary and emergency labeling data; obtain the emergency request information output by the emergency warning model. Status; access to robotic customer service or human customer service based on the emergency status.
  • the computer device 102 is various personal computers, notebook computers, smart phones, tablet computers, portable wearable devices, and the like.
  • the computer device 102 may also be a server, and the server may be implemented by an independent server or a server cluster composed of multiple servers.
  • FIG. 2 is a schematic flowchart of a method for switching customer service sessions according to an embodiment of the present application.
  • a method for switching customer service sessions includes:
  • Operation 202 Receive request information sent by a client.
  • the request information is sent to a pre-trained emergency warning model, and the emergency warning model is obtained by training from an emergency vocabulary and emergency labeling data.
  • Operation 206 Obtain the emergency state of the request information output by the emergency warning model.
  • the robot customer service or the manual customer service is accessed according to the emergency status.
  • the emergency state of the information sent by the customer is obtained, and the robot customer service or manual customer service is accessed according to the emergency state. Based on this, the urgent expression of the customer can be effectively identified. According to the urgent expression of targeted service, the user experience will be better, and the unpredictable caused by the failure to identify the urgent expression of the customer can be effectively avoided. Bad consequences.
  • the request information sent by the customer may be, but is not limited to, text information, and the type of the request information may be adjusted according to the device where the customer inputs the request information.
  • the customer's input device is a phone
  • the customer's key behavior can be set to request information. Therefore, this solution can be applied not only to the scenario of mobile customer service conversation, but also to scenarios such as telephone customer service and video customer service.
  • the corresponding request information can be set to phone key behavior, key behavior of visual interface, voice information, or Gesture information can even be expression information.
  • the emergency early warning model may be obtained by training from an emergency vocabulary and emergency labeling data.
  • the emergency alert model includes a keyword list enhancement model 320, an emergency recognition model 340, and a classification layer 360.
  • the emergency alert model receives the above text information, and the keyword list enhancement model 320 and emergency recognition The model 340 will analyze the text information.
  • the result obtained by the keyword list enhancement model 320 is the keyword feature.
  • the result obtained by the emergency recognition model 340 is the emergency weight feature of each word in the text information. Both results will be input to the classification layer.
  • complete features are obtained by integration, and finally, the classification layer 360 divides the text information into non-urgent, mild emergency, or severe emergency according to the obtained complete features and outputs them.
  • FIG. 4 is a flowchart of training an emergency warning model according to an embodiment. As shown in Figure 4, the training process of the emergency warning model may include:
  • Operation 402 Obtain a manual customer service log.
  • Operation 404 Extract emergency vocabulary and emergency annotation data from the manual customer service log.
  • the emergency word list is used as training data to train a keyword list enhancement model.
  • Operation 408 Use the emergency labeling data as training data to train an emergency recognition model.
  • the emergency label data is used as training data to train a classification layer.
  • the keyword list enhanced model and the emergency recognition model are respectively connected with the classification layer to obtain an emergency early warning model.
  • the training process of the emergency warning model may be, but is not limited to, the training process described above.
  • the emergency state may include three levels, that is, three states of non-emergency, light emergency, and severe emergency.
  • three levels that is, three states of non-emergency, light emergency, and severe emergency.
  • the specific implementation is not limited to these three levels of this embodiment.
  • FIG. 5 is a flowchart of a method for switching customer service sessions according to an embodiment. As shown in FIG. 5, in one embodiment, a method for switching customer service sessions includes:
  • Operation 502 Receive request information sent by a client.
  • the request information is sent to a pre-trained emergency warning model; the emergency warning model is obtained by training from an emergency vocabulary and emergency labeling data.
  • Operation 506 Obtain the emergency state of the request information output by the emergency warning model.
  • the emergency state is a mild emergency state, determine whether the manual customer service is busy, and when the manual customer service is busy, access the robotic customer service; when the manual customer is not busy, access the manual customer service.
  • the request information sent by the client may include the scene information of the conversation, and the scene information may make the emergency warning model judge the emergency state more accurately.
  • FIG. 6 is a schematic structural diagram of a customer service session switching device provided in an embodiment.
  • the receiving module 610 is configured to receive request information sent by a client.
  • a sending module 620 is configured to send the request information to an emergency training model obtained in advance; the emergency warning model is obtained by training from an emergency vocabulary and emergency labeling data.
  • the first acquiring module 630 is configured to acquire an emergency state of the request information output by the emergency early warning model.
  • the access module 640 is configured to access the robotic customer service or the manual customer service according to the emergency status.
  • the emergency state may include a non-emergency state, a mild emergency state, and a severe emergency state.
  • the access module 640 may include a first access unit 641, a second access unit 642, and a third access unit 643.
  • the first access unit 641 is configured to access the robot customer service when the emergency state is a non-emergency state.
  • the second access unit 642 is configured to determine whether the artificial customer service is busy when the emergency state is a mild emergency; access the robot customer service when the artificial customer service is busy; and access the manual customer service when the artificial customer is not busy.
  • the third access unit 643 is configured to access the artificial customer service when the emergency state is a severe emergency state.
  • the provided customer service session switching device may further include a model training module 650.
  • the emergency warning model may include a keyword list enhancement model, an emergency recognition model, and a classification layer.
  • the model training module 650 may include:
  • the second obtaining unit 651 is configured to obtain a manual customer service log.
  • An extraction unit 652 is configured to extract emergency vocabulary and emergency annotation data from a manual customer service log.
  • the first training unit 653 is configured to use the emergency word list as training data to train a keyword list enhancement model.
  • the second training unit 654 is configured to use the emergency labeling data as training data to train an emergency recognition model.
  • the third training unit 655 is configured to use the emergency labeling data as training data to train a classification layer.
  • a connecting unit 656 is configured to connect the keyword list enhancement model and the emergency recognition model to the classification layer to obtain an emergency early warning model.
  • the request information sent by the customer may also include the scene information of the session, and the scene information can make the emergency warning model judge the emergency state more accurately.
  • Each module in the above-mentioned customer service session switching device may be implemented in whole or in part by software, hardware, and a combination thereof.
  • the above-mentioned modules may be embedded in the hardware form or independent of the processor in the computer device, or may be stored in the memory of the computer device in the form of software, so that the processor calls and performs the operations corresponding to the above modules.
  • a unit may be, but is not limited to being, a process running on a processor, a processor, an object, executable code, a thread of execution, a program, and / or a computer.
  • a unit may be, but is not limited to being, a process running on a processor, a processor, an object, executable code, a thread of execution, a program, and / or a computer.
  • the application running on the server and the server can be a unit.
  • One or more units can reside within a process and / or thread of execution and a unit can be localized on one computer and / or distributed between two or more computers.
  • FIG. 9 is a schematic structural diagram of a customer service session switching system according to an embodiment of the present application.
  • the provided customer service session switching system includes a processor 920 and a memory 940 connected to the processor 920.
  • the memory 940 is configured to store a computer program.
  • the processor 920 executes the following operations:
  • the emergency warning model is trained from the emergency vocabulary and emergency labeling data
  • the emergency state includes a non-emergency state, a mild emergency state, and a severe emergency state; when the processor 920 executes access to the robotic customer service or manual customer service according to the emergency state, it also performs:
  • the emergency is a mild emergency, determine whether the manual customer service is busy; when the manual customer service is busy, access the robotic customer service; when the manual customer is not busy, access the manual customer service;
  • the emergency warning model includes a keyword table enhancement model, an emergency recognition model, and a classification layer; when the computer program is executed by the processor 920, the processor 920 performs the following operations: obtaining a manual customer service log;
  • the keyword list augmentation model and the emergency recognition model are connected to the classification layer to obtain an emergency warning model.
  • the request information includes scene information.
  • the above-mentioned customer service session switching system may include a computer device.
  • the computer device may include a processor, a memory, a network interface, a display screen, and an input device connected through a system bus.
  • 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 running an operating system and computer programs in a non-volatile storage medium.
  • the network interface of the computer device is used to communicate with an external terminal through a network connection.
  • the computer program is executed by the processor to implement a method for switching customer service sessions.
  • the display screen of the computer device may be a liquid crystal display screen or an electronic ink display screen.
  • the input device of the computer device may be a touch layer covered on the display screen, or a button, a trackball, or a touchpad provided on the computer device casing. , Or an external keyboard, trackpad, or mouse.
  • the customer service session switching device provided in this application may be implemented in the form of a computer program, and the computer program may be run on the customer service session switching system shown in FIG. 9.
  • the memory of the customer service session switching system may store various program modules constituting the customer service session switching device.
  • the computer program constituted by each program module causes the processor to perform the operations in the customer service session switching method of the present embodiment described in this specification.
  • One or more non-volatile computer-readable storage media storing computer-readable instructions.
  • the one or more processors perform the following operations: Request information
  • the emergency warning model is trained from the emergency vocabulary and emergency labeling data
  • the emergency state includes a non-emergency state, a mild emergency state, and a severe emergency state; when the processor executes to access the robotic customer service or manual customer service according to the emergency state, it also performs the following operations:
  • the emergency is a mild emergency, determine whether the manual customer service is busy; when the manual customer service is busy, access the robotic customer service; when the manual customer is not busy, access the manual customer service;
  • the emergency warning model includes a keyword list enhancement model, an emergency recognition model, and a classification layer; when the computer program is executed by the processor, the processor further executes: obtaining a manual customer service log;
  • the keyword list augmentation model and the emergency recognition model are connected to the classification layer to obtain an emergency warning model.
  • the request information includes scene information.
  • Any process or method description in a flowchart or otherwise described herein can be understood as representing a module, fragment, or portion of code that includes one or more executable instructions for implementing the operation of a particular logical function or process
  • the scope of the preferred embodiments of this application includes additional implementations in which the functions may be performed out of the order shown or discussed, including performing the functions in a substantially simultaneous manner or in the reverse order according to the functions involved, which should It is understood by those skilled in the art to which the embodiments of the present application pertain.
  • each part of the application may be implemented by hardware, software, firmware, or a combination thereof.
  • multiple operations or methods may be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system.
  • a suitable instruction execution system For example, if implemented in hardware, as in another embodiment, it may be implemented using any one or a combination of the following techniques known in the art: Discrete logic circuits, application-specific integrated circuits with suitable combinational logic gate circuits, programmable gate arrays (PGA), field programmable gate arrays (FPGA), etc.
  • each functional unit in each embodiment of the present application may be integrated into one processing module, or each unit may exist separately physically, or two or more units may be integrated into one module.
  • the above integrated modules may be implemented in the form of hardware or software functional modules. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it may also be stored in a computer-readable storage medium.
  • the aforementioned storage medium may be a read-only memory, a magnetic disk, or an optical disk.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Physics & Mathematics (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • General Physics & Mathematics (AREA)
  • Marketing (AREA)
  • General Business, Economics & Management (AREA)
  • Strategic Management (AREA)
  • Economics (AREA)
  • Development Economics (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • General Engineering & Computer Science (AREA)
  • Information Transfer Between Computers (AREA)
  • Manipulator (AREA)

Abstract

一种客服会话切换方法,包括:接收客户发送的请求信息;将请求信息发送到预先训练得到的紧急预警模型;获取紧急预警模型输出的请求信息的紧急状态;根据紧急状态的等级接入机器人客服或人工客服。

Description

客服会话切换方法、系统和存储介质
本申请要求于2018年08月15日提交中国专利局、申请号为201810927997.3、发明名称为“客服会话切换的方法、装置及系统”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及客服机器人技术领域,具体涉及一种客服会话切换方法、系统和非易失性计算机可读存储介质。
背景技术
随着服务业的发展,客户越来越注重获取服务的便捷性,人们便研发出了客服机器人与客户进行沟通的方案,以此来加快对客户提出的疑问的响应速度。
目前,客服机器人一般都是在接收到客户的问题后,先使用排序模型对候选的常见问题(FAQ,Frequently Asked Questions)进行排序,判断客户的问题与每个FAQ的相关性,然后再根据置信度模型判断排在第一位的FAQ对应的答案的置信度。如果排在第一位的置信度足够高,则直接把排在第一位的FAQ对应的答案返回给客户,否则,就推荐N条可能性较高的FAQ供客户自主选择。
上述客服机器人与客户交互的方案虽然可以高效地解决客户的部分高频的问题,但是,当客户遇到了严重的突发情况,问题中伴随着一些紧急表达时,当前的交互方案并不能将这种紧急表达识别出来,也就不能根据紧急表达采用较为合适的方案来解决客户的问题,这样就会让客户觉得自己的问题 没有被妥善解决,有可能会损伤用户体验,甚至带来一些无法预估的恶劣后果。
发明内容
根据本申请的各种实施例,提供一种客服会话切换方法、系统和非易失性计算机可读存储介质。
一种客服会话切换方法,包括:
接收客户发送的请求信息;
将所述请求信息发送到预先训练得到的紧急预警模型,其中,所述紧急预警模型由紧急词表和紧急标注数据训练得到;
获取所述紧急预警模型输出的所述请求信息的紧急状态;及
根据所述紧急状态的等级接入机器人客服或人工客服。
一种客服会话切换系统,包括:处理器,以及与所述处理器相连接的存储器;所述存储器用于存储计算机程序,所述计算机程序至少用于执行如下操作:
接收客户发送的请求信息;
将所述请求信息发送到预先训练得到的紧急预警模型,其中,所述紧急预警模型由紧急词表和紧急标注数据训练得到;
获取所述紧急预警模型输出的所述请求信息的紧急状态;及
根据所述紧急状态的等级接入机器人客服或人工客服。
一个或多个存储有计算机可读指令的非易失性计算机可读存储介质,所述计算机可读指令被一个或多个处理器执行时,使得所述一个或多个处理器执行以下操作:接收客户发送的请求信息;
将所述请求信息发送到预先训练得到的紧急预警模型,其中,所述紧急预警模型由紧急词表和紧急标注数据训练得到;
获取所述紧急预警模型输出的所述请求信息的紧急状态;及
根据所述紧急状态的等级接入机器人客服或人工客服。
本申请的一个或多个实施例的细节在下面的附图和描述中提出。本申请的其它特征、目的和优点将从说明书、附图以及权利要求书变得明显。
附图说明
为了更清楚地说明本申请实施例或现有技术中的技术方案,下面将对实施例或现有技术描述中所需要使用的附图作简单地介绍,显而易见地,下面描述中的附图仅仅是本申请的一些实施例,对于本领域普通技术人员来讲,在不付出创造性劳动的前提下,还可以根据这些附图获得其他的附图。
为了更好地描述和说明本申请公开的实施例和/或示例,可以参考一副或者多副附图。用于描述附图的附加细节或示例不应当被认为是对所公开的发明、目前描述的实施例和/或示例以及目前理解的这些发明的最佳模式中的任何一者的范围的限制。
图1为根据一些实施例提供的客服会话切换方法的应用环境图。
图2为根据一些实施例提供的客服会话切换方法的流程示意图。
图3为根据一些实施例提供的紧急预警模型的结构示意图。
图4为根据一些实施例提供的客服会话切换方法的流程示意图。
图5为根据一些实施例提供的训练紧急预警模型的流程示意图。
图6为根据一些实施例提供的客服会话切换装置的结构示意图。
图7为根据一些实施例提供的接入模块的结构示意图。
图8为根据一些实施例提供的模型训练模块的结构示意图。
图9为根据一些实施例提供的客服会话切换系统的结构示意图。
具体实施方式
为使本申请的目的、技术方案和优点更加清楚,下面将对本申请的技术方案进行详细的描述。显然,所描述的实施例仅仅是本申请一部分实施例,而不是全部的实施例。基于本申请中的实施例,本领域普通技术人员在没有做出创造性劳动的前提下所得到的所有其它实施方式,都属于本申请所保护的范围。
本申请提供的客服会话切换方法,可以应用于如图1所示的应用环境中。该应用环境包括计算机设备102。计算机设备102可以接收客户发送的请求信息;将请求信息发送到预先训练得到的紧急预警模型,其中,紧急预警模型由紧急词表和紧急标注数据训练得到;获取紧急预警模型输出的请求信息的紧急状态;根据紧急状态的等级接入机器人客服或人工客服。其中,计算机设备102是各种个人计算机、笔记本电脑、智能手机、平板电脑和便携式可穿戴设备等。可选地,计算机设备102也可以是服务器,服务器可以用独立的服务器或者是多个服务器组成的服务器集群来实现。
图2是本申请实施例提供的一种客服会话切换方法的流程示意图。
如图2所示,在一个实施例中,提供的客服会话切换方法包括:
操作202,接收客户发送的请求信息。
操作204,将请求信息发送到预先训练得到的紧急预警模型,紧急预警模型由紧急词表和紧急标注数据训练得到。
操作206,获取紧急预警模型输出的请求信息的紧急状态。
操作208,根据紧急状态的等级接入机器人客服或人工客服。
由于将接收到的客户发送的信息发送到紧急预警模型,得到客户发送的信息的紧急状态,根据紧急状态接入机器人客服或者人工客服。基于此,便可以有效识别出客户的紧急的表达,根据紧急的表达有针对性的提供服务,用户的体验会更佳,也可以有效地避免因无法识别客户紧急的表达而产生的无法预估的恶劣后果。
其中,操作202中,客户发送的请求信息可以但不仅限于是文字信息,请求信息的类型可以根据客户输入请求信息的设备来调整。比如,当客户的输入设备是电话时,可以将客户的按键行为设置为请求信息。所以本方案不仅可以应用在移动终端的客服会话场景中,还可以应用在电话客服、视频客服等场景中,对应的请求信息就可以设置成电话按键行为、可视界面的按键行为、语音信息或者手势信息,甚至还可以是表情信息。
操作204中,紧急预警模型可以由紧急词表和紧急标注数据训练得到。
其中,紧急预警模型的结构如图3所示。紧急预警模型包括关键词表增强模型320、紧急识别模型340和分类层360,当客户发送的请求信息为文字信息时,紧急预警模型接收到上述文字信息后,关键词表增强模型320和紧急识别模型340都会对文字信息进行分析,关键词表增强模型320得到的结果是关键词特征,紧急识别模型340得到的结果是文字信息中每个词的紧急权重特征,两个结果都会输入到分类层360中进行整和得到完整特征,最后分类层360根据得到的完整特征将文字信息划分为不紧急、轻度紧急或者重度紧急等并输出。
图4为一个实施例中提供的训练紧急预警模型的流程图。如图4所示,紧急预警模型的训练过程可以包括:
操作402,获取人工客服日志。
操作404,从人工客服日志中提取紧急词表和紧急标注数据。
操作406,以紧急词表作为训练数据,训练关键词表增强模型。
操作408,以紧急标注数据作为训练数据,训练紧急识别模型。
操作410,以紧急标注数据作为训练数据,训练分类层。
操作412,将关键词表增强模型和紧急识别模型分别与分类层连接得到紧急预警模型。
在操作404中,从人工客服日志中提取紧急词表和紧急标注数据的方法 可以有多种,可以是人工提取的方法,也可以是使用预先训练的提取模型进行提取的方法。
需要说明的是,紧急预警模型的训练过程可以但不仅限于如上的训练过程。
在一个实施例中,紧急状态可以包括三个等级,即不紧急、轻度紧急和重度紧急三个状态,当然,具体实施时,不限于本实施例的这三个等级。
图5为一个实施例中客服会话切换方法的流程图。如图5所示,在一个实施例中,提供的客服会话切换方法包括:
操作502,接收客户发送的请求信息。
操作504,将请求信息发送到预先训练得到的紧急预警模型;紧急预警模型由紧急词表和紧急标注数据训练得到。
操作506,获取紧急预警模型输出的请求信息的紧急状态。
操作508,当紧急状态为不紧急状态时,接入机器人客服。
操作510,当紧急状态为轻度紧急状态时,判断人工客服是否繁忙,当人工客服繁忙时,接入机器人客服;当人工客户不繁忙时,接入人工客服。
操作512,当紧急状态为重度紧急状态时,接入人工客服。
需要说明的是,客户发送的请求信息中可以包括会话的场景信息,场景信息可以使紧急预警模型对紧急状态的判断更加准确。
另外,图6是一个实施例中提供的客服会话切换装置的结构示意图。
本实施例提供的一种客服会话切换装置可以包括:
接收模块610,用于接收客户发送的请求信息。
发送模块620,用于将请求信息发送到预先训练得到的紧急预警模型;紧急预警模型由紧急词表和紧急标注数据训练得到。
第一获取模块630,用于获取紧急预警模型输出的请求信息的紧急状态。
接入模块640,用于根据紧急状态的等级接入机器人客服或人工客服。
其中,紧急状态可以包括不紧急状态、轻度紧急状态和重度紧急状态。如图7所示,在一个实施例中,接入模块640可以包括第一接入单元641、第二接入单元642和第三接入单元643。
具体的,第一接入单元641,用于当紧急状态为不紧急状态时,接入机器人客服。
第二接入单元642,用于当紧急状态为轻度紧急状态,判断人工客服是否繁忙;当人工客服繁忙时,接入机器人客服;当人工客户不繁忙时,接入人工客服。
第三接入单元643,用于当紧急状态为重度紧急状态,接入人工客服。
在一个实施例中,提供的客服会话切换装置还可以包括模型训练模块650。需要说明的是,紧急预警模型可以包括关键词表增强模型、紧急识别模型和分类层。对应的,如图8所示,在一个实施例中,模型训练模块650可以包括:
第二获取单元651,用于获取人工客服日志。
提取单元652,用于从人工客服日志中提取紧急词表和紧急标注数据。
第一训练单元653,用于以紧急词表作为训练数据,训练关键词表增强模型。
第二训练单元654,用于以紧急标注数据作为训练数据,训练紧急识别模型。
第三训练单元655,用于以紧急标注数据作为训练数据,训练分类层。
连接单元656,用于将关键词表增强模型和紧急识别模型分别与分类层连接得到紧急预警模型。
另外,客户发送的请求信息还可以包括会话的场景信息,场景信息可以使紧急预警模型对紧急状态的判断更加准确。
上述客服会话切换装置中的各个模块可全部或部分通过软件、硬件及其 组合来实现。上述各模块可以硬件形式内嵌于或独立于计算机设备中的处理器中,也可以以软件形式存储于计算机设备中的存储器中,以便于处理器调用执行以上各个模块对应的操作。
如在本申请中所使用的,术语“单元”、“模块”和“系统”等旨在表示计算机相关的实体,它可以是硬件、硬件和软件的组合、软件、或者执行中的软件。例如,单元可以是但不限于是,在处理器上运行的进程、处理器、对象、可执行码、执行的线程、程序和/或计算机。作为说明,运行在服务器上的应用程序和服务器都可以是单元。一个或多个单元可以驻留在进程和/或执行的线程中,并且单元可以位于一个计算机内和/或分布在两个或更多的计算机之间。
图9是本申请的实施例提供的一种客服会话切换系统的结构示意图。如图9所示,在一个实施例中,提供的客服会话切换系统包括处理器920和与该处理器920相连的存储器940。
存储器940用于存储计算机程序,计算机程序被处理器920执行时,使得处理器920执行如下操作:
接收客户发送的请求信息;
将请求信息发送到预先训练得到的紧急预警模型;紧急预警模型由紧急词表和紧急标注数据训练得到;
获取紧急预警模型输出的请求信息的紧急状态;
根据紧急状态的等级接入机器人客服或人工客服。
在一个实施例中,紧急状态包括不紧急状态、轻度紧急状态和重度紧急状态;处理器920执行根据紧急状态接入机器人客服或人工客服时,还执行:
当紧急状态为不紧急状态时,接入机器人客服;
当紧急状态为轻度紧急状态时,判断人工客服是否繁忙;当人工客服繁忙时,接入机器人客服;当人工客户不繁忙时,接入人工客服;
当紧急状态为重度紧急状态,接入人工客服。
在一个实施例中,紧急预警模型包括关键词表增强模型、紧急识别模型和分类层;计算机程序被处理器920执行时,使得处理器920执行如下操作:获取人工客服日志;
从人工客服日志中提取紧急词表和紧急标注数据;
以紧急词表作为训练数据,训练关键词表增强模型;
以紧急标注数据作为训练数据,训练紧急识别模型;
以紧急标注数据作为训练数据,训练分类层;
将关键词表增强模型和紧急识别模型分别与分类层连接得到紧急预警模型。
在一个实施例中,请求信息包括场景信息。
在一个实施例中,上述客服会话切换系统可以包括计算机设备。计算机设备可以包括通过系统总线连接的处理器、存储器、网络接口、显示屏和输入装置。其中,该计算机设备的处理器用于提供计算和控制能力。该计算机设备的存储器包括非易失性存储介质、内存储器。该非易失性存储介质存储有操作系统和计算机程序。该内存储器为非易失性存储介质中的操作系统和计算机程序的运行提供环境。该计算机设备的网络接口用于与外部的终端通过网络连接通信。该计算机程序被处理器执行时以实现一种客服会话切换方法。该计算机设备的显示屏可以是液晶显示屏或者电子墨水显示屏,该计算机设备的输入装置可以是显示屏上覆盖的触摸层,也可以是计算机设备外壳上设置的按键、轨迹球或触控板,还可以是外接的键盘、触控板或鼠标等。
在一个实施例中,本申请提供的客服会话切换装置可以实现为一种计算机程序的形式,计算机程序可在如图9所示的客服会话切换系统上运行。客服会话切换系统的存储器中可存储组成该客服会话切换装置的各个程序模块。各个程序模块构成的计算机程序使得处理器执行本说明书中描述的本申 请各个实施例的客服会话切换方法中的操作。
一个或多个存储有计算机可读指令的非易失性计算机可读存储介质,计算机可读指令被一个或多个处理器执行时,使得一个或多个处理器执行以下操作:接收客户发送的请求信息;
将请求信息发送到预先训练得到的紧急预警模型;紧急预警模型由紧急词表和紧急标注数据训练得到;
获取紧急预警模型输出的请求信息的紧急状态;
根据紧急状态的等级接入机器人客服或人工客服。
可选的,紧急状态包括不紧急状态、轻度紧急状态和重度紧急状态;处理器执行根据紧急状态接入机器人客服或人工客服时,还执行如下操作:
当紧急状态为不紧急状态时,接入机器人客服;
当紧急状态为轻度紧急状态时,判断人工客服是否繁忙;当人工客服繁忙时,接入机器人客服;当人工客户不繁忙时,接入人工客服;
当紧急状态为重度紧急状态时,接入人工客服。
在一个实施例中,紧急预警模型包括关键词表增强模型、紧急识别模型和分类层;计算机程序被处理器执行时,处理器还执行:获取人工客服日志;
从人工客服日志中提取紧急词表和紧急标注数据;
以紧急词表作为训练数据,训练关键词表增强模型;
以紧急标注数据作为训练数据,训练紧急识别模型;
以紧急标注数据作为训练数据,训练分类层;
将关键词表增强模型和紧急识别模型分别与分类层连接得到紧急预警模型。
在一个实施例中,请求信息包括场景信息。
可以理解的是,上述各实施例中相同或相似部分可以相互参考,在一些实施例中未详细说明的内容可以参见其他实施例中相同或相似的内容。
需要说明的是,在本申请的描述中,术语“第一”、“第二”等仅用于描述目的,而不能理解为指示或暗示相对重要性。此外,在本申请的描述中,除非另有说明,“多个”的含义是指至少两个。
流程图中或在此以其他方式描述的任何过程或方法描述可以被理解为,表示包括一个或更多个用于实现特定逻辑功能或过程的操作的可执行指令的代码的模块、片段或部分,并且本申请的优选实施方式的范围包括另外的实现,其中可以不按所示出或讨论的顺序,包括根据所涉及的功能按基本同时的方式或按相反的顺序,来执行功能,这应被本申请的实施例所属技术领域的技术人员所理解。
应当理解,本申请的各部分可以用硬件、软件、固件或它们的组合来实现。在上述实施方式中,多个操作或方法可以用存储在存储器中且由合适的指令执行系统执行的软件或固件来实现。例如,如果用硬件来实现,和在另一实施方式中一样,可用本领域公知的下列技术中的任一项或他们的组合来实现:具有用于对数据信号实现逻辑功能的逻辑门电路的离散逻辑电路,具有合适的组合逻辑门电路的专用集成电路,可编程门阵列(PGA),现场可编程门阵列(FPGA)等。
本技术领域的普通技术人员可以理解实现上述实施例方法携带的全部或部分操作是可以通过程序来指令相关的硬件完成,的程序可以存储于一种计算机可读存储介质中,该程序在执行时,包括方法实施例的操作之一或其组合。
此外,在本申请各个实施例中的各功能单元可以集成在一个处理模块中,也可以是各个单元单独物理存在,也可以两个或两个以上单元集成在一个模块中。上述集成的模块既可以采用硬件的形式实现,也可以采用软件功能模块的形式实现。集成的模块如果以软件功能模块的形式实现并作为独立的产品销售或使用时,也可以存储在一个计算机可读取存储介质中。
上述提到的存储介质可以是只读存储器,磁盘或光盘等。
在本说明书的描述中,参考术语“一个实施例”、“一些实施例”、“示例”、“具体示例”、或“一些示例”等的描述意指结合该实施例或示例描述的具体特征、结构、材料或者特点包含于本申请的至少一个实施例或示例中。在本说明书中,对上述术语的示意性表述不一定指的是相同的实施例或示例。而且,描述的具体特征、结构、材料或者特点可以在任何的一个或多个实施例或示例中以合适的方式结合。尽管上面已经示出和描述了本申请的实施例,可以理解的是,上述实施例是示例性的,不能理解为对本申请的限制,本领域的普通技术人员在本申请的范围内可以对上述实施例进行变化、修改、替换和变型。

Claims (16)

  1. 一种客服会话切换方法,包括:
    接收客户发送的请求信息;
    将所述请求信息发送到预先训练得到的紧急预警模型,其中,所述紧急预警模型由紧急词表和紧急标注数据训练得到;
    获取所述紧急预警模型输出的所述请求信息的紧急状态;及
    根据所述紧急状态的等级接入机器人客服或人工客服。
  2. 根据权利要求1所述的方法,其特征在于,所述根据所述紧急状态接入机器人客服或人工客服,包括:
    当所述紧急状态为不紧急状态时,接入机器人客服;及
    当所述紧急状态为重度紧急状态时,接入人工客服。
  3. 根据权利要求2所述的方法,其特征在于,所述根据所述紧急状态接入机器人客服或人工客服,还包括:
    当所述紧急状态为轻度紧急状态,判断人工客服是否繁忙;
    当人工客服繁忙时,接入机器人客服;及
    当人工客户不繁忙时,接入人工客服。
  4. 根据权利要求1所述的方法,其特征在于,所述紧急预警模型包括关键词表增强模型、紧急识别模型和分类层;所述方法还包括:
    获取人工客服日志;
    从所述人工客服日志中提取所述紧急词表和所述紧急标注数据;
    以所述紧急词表作为训练数据,训练所述关键词表增强模型;
    以所述紧急标注数据作为训练数据,训练所述紧急识别模型;
    以所述紧急标注数据作为训练数据,训练所述分类层;及
    将所述关键词表增强模型和所述紧急识别模型分别与所述分类层连接得到所述紧急预警模型。
  5. 根据权利要求1所述的方法,其特征在于,当所述请求信息为文字信息时,所述获取所述紧急预警模型输出的所述请求信息的紧急状态,包括:
    通过所述紧急预警模型的关键词表增强模型对所述请求信息进行分析,得到关键词特征;
    通过所述紧急预警模型的紧急识别模型对所述请求信息进行分析,得到所述请求信息中每一个词对应的紧急权重特征;
    通过所述紧急预警模型的分类层将所述关键词特征及每一个词对应的紧急权重特征进行整合处理,得到完整特征;及
    通过所述分类层基于所述完整特征划分所述请求信息对应的紧急状态,并输出所述紧急状态。
  6. 根据权利要求1所述的方法,其特征在于,所述请求信息包括场景信息。
  7. 一种客服会话切换的系统,包括:处理器,以及与所述处理器相连接的存储器;所述存储器用于存储计算机程序,使得所述处理器执行如下操作:
    接收客户发送的请求信息;
    将所述请求信息发送到预先训练得到的紧急预警模型,其中,所述紧急预警模型由紧急词表和紧急标注数据训练得到;
    获取所述紧急预警模型输出的所述请求信息的紧急状态;及
    根据所述紧急状态的等级接入机器人客服或人工客服。
  8. 根据权利要求7所述的系统,其特征在于,所述处理器执行所述根据所述紧急状态接入机器人客服或人工客服时,还执行:
    当所述紧急状态为不紧急状态时,接入机器人客服;及
    当所述紧急状态为重度紧急状态时,接入人工客服。
  9. 根据权利要求8所述的系统,其特征在于,所述处理器执行所述根据所述紧急状态接入机器人客服或人工客服时,还执行:
    当所述紧急状态为轻度紧急状态,判断人工客服是否繁忙;
    当人工客服繁忙时,接入机器人客服;及
    当人工客户不繁忙时,接入人工客服。
  10. 根据权利要求1所述的系统,其特征在于,所述处理器还执行如下操作:
    获取人工客服日志;
    从所述人工客服日志中提取所述紧急词表和所述紧急标注数据;
    以所述紧急词表作为训练数据,训练关键词表增强模型;
    以所述紧急标注数据作为训练数据,训练紧急识别模型;
    以所述紧急标注数据作为训练数据,训练分类层;及
    将所述关键词表增强模型和所述紧急识别模型分别与所述分类层连接得到所述紧急预警模型。
  11. 根据权利要求1所述的系统,其特征在于,当所述请求信息为文字信息时,所述处理器执行所述获取所述紧急预警模型输出的所述请求信息的紧急状态时,还执行如下操作:
    通过所述紧急预警模型的关键词表增强模型对所述请求信息进行分析,得到关键词特征;
    通过所述紧急预警模型的紧急识别模型对所述请求信息进行分析,得到所述请求信息中每一个词对应的紧急权重特征;
    通过所述紧急预警模型的分类层将所述关键词特征及每一个词对应的紧急权重特征进行整合处理,得到完整特征;及
    通过所述分类层基于所述完整特征划分所述请求信息对应的紧急状态,并输出所述紧急状态。
  12. 一个或多个存储有计算机可读指令的非易失性计算机可读存储介质,所述计算机可读指令被一个或多个处理器执行时,使得所述一个或多个处理 器执行以下操作:
    接收客户发送的请求信息;
    将所述请求信息发送到预先训练得到的紧急预警模型,其中,所述紧急预警模型由紧急词表和紧急标注数据训练得到;
    获取所述紧急预警模型输出的所述请求信息的紧急状态;及
    根据所述紧急状态的等级接入机器人客服或人工客服。
  13. 根据权利要求12所述的非易失性计算机可读存储介质,其特征在于,所述处理器执行所述根据所述紧急状态接入机器人客服或人工客服时,还执行:
    当所述紧急状态为不紧急状态时,接入机器人客服;及
    当所述紧急状态为重度紧急状态时,接入人工客服。
  14. 根据权利要求13所述的非易失性计算机可读存储介质,其特征在于,所述处理器执行所述根据所述紧急状态接入机器人客服或人工客服时,还执行:
    当所述紧急状态为轻度紧急状态,判断人工客服是否繁忙;
    当人工客服繁忙时,接入机器人客服;及
    当人工客户不繁忙时,接入人工客服。
  15. 根据权利要求12所述的非易失性计算机可读存储介质,其特征在于,所述处理器还执行如下操作:
    获取人工客服日志;
    从所述人工客服日志中提取所述紧急词表和所述紧急标注数据;
    以所述紧急词表作为训练数据,训练关键词表增强模型;
    以所述紧急标注数据作为训练数据,训练紧急识别模型;
    以所述紧急标注数据作为训练数据,训练分类层;及
    将所述关键词表增强模型和所述紧急识别模型分别与所述分类层连接得 到所述紧急预警模型。
  16. 根据权利要求12所述的非易失性计算机可读存储介质,其特征在于,当所述请求信息为文字信息时,所述处理器执行所述获取所述紧急预警模型输出的所述请求信息的紧急状态时,还执行如下操作:
    通过所述紧急预警模型的关键词表增强模型对所述请求信息进行分析,得到关键词特征;
    通过所述紧急预警模型的紧急识别模型对所述请求信息进行分析,得到所述请求信息中每一个词对应的紧急权重特征;
    通过所述紧急预警模型的分类层将所述关键词特征及每一个词对应的紧急权重特征进行整合处理,得到完整特征;及
    通过所述分类层基于所述完整特征划分所述请求信息对应的紧急状态,并输出所述紧急状态。
PCT/CN2019/100288 2018-08-15 2019-08-13 客服会话切换方法、系统和存储介质 WO2020034928A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810927997.3 2018-08-15
CN201810927997.3A CN109087175A (zh) 2018-08-15 2018-08-15 客服会话切换的方法、装置及系统

Publications (1)

Publication Number Publication Date
WO2020034928A1 true WO2020034928A1 (zh) 2020-02-20

Family

ID=64793538

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2019/100288 WO2020034928A1 (zh) 2018-08-15 2019-08-13 客服会话切换方法、系统和存储介质

Country Status (2)

Country Link
CN (1) CN109087175A (zh)
WO (1) WO2020034928A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109087175A (zh) * 2018-08-15 2018-12-25 深圳追科技有限公司 客服会话切换的方法、装置及系统
CN111526253B (zh) * 2020-03-09 2022-02-08 深圳追一科技有限公司 呼叫控制方法、装置、计算机设备和存储介质
CN111565151A (zh) * 2020-04-27 2020-08-21 中国银行股份有限公司 一种客服线路路由方法和装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103309948A (zh) * 2013-05-20 2013-09-18 携程计算机技术(上海)有限公司 联络中心舆情监控分析和智能分配处理系统及方法
US20160269883A1 (en) * 2015-03-13 2016-09-15 Kapali Eswaran Automated Service Systems and Methods
CN107590159A (zh) * 2016-07-08 2018-01-16 阿里巴巴集团控股有限公司 机器人客服转人工客服的方法和装置
CN107979704A (zh) * 2017-12-01 2018-05-01 中国联合网络通信集团有限公司 排队方法、排队系统
CN109087175A (zh) * 2018-08-15 2018-12-25 深圳追科技有限公司 客服会话切换的方法、装置及系统

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105592237B (zh) * 2014-10-24 2019-02-05 中国移动通信集团公司 一种会话切换的方法、装置及智能客服机器人
CN106844750A (zh) * 2017-02-16 2017-06-13 深圳追科技有限公司 一种基于客服机器人中情感安抚的人机交互方法及系统

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103309948A (zh) * 2013-05-20 2013-09-18 携程计算机技术(上海)有限公司 联络中心舆情监控分析和智能分配处理系统及方法
US20160269883A1 (en) * 2015-03-13 2016-09-15 Kapali Eswaran Automated Service Systems and Methods
CN107590159A (zh) * 2016-07-08 2018-01-16 阿里巴巴集团控股有限公司 机器人客服转人工客服的方法和装置
CN107979704A (zh) * 2017-12-01 2018-05-01 中国联合网络通信集团有限公司 排队方法、排队系统
CN109087175A (zh) * 2018-08-15 2018-12-25 深圳追科技有限公司 客服会话切换的方法、装置及系统

Also Published As

Publication number Publication date
CN109087175A (zh) 2018-12-25

Similar Documents

Publication Publication Date Title
US10275448B2 (en) Automatic question generation and answering based on monitored messaging sessions
US11455981B2 (en) Method, apparatus, and system for conflict detection and resolution for competing intent classifiers in modular conversation system
US10997258B2 (en) Bot networks
US20230092702A1 (en) Database systems and methods of naming record groups
US11068518B2 (en) Reducing negative effects of service waiting time in humanmachine interaction to improve the user experience
WO2020034928A1 (zh) 客服会话切换方法、系统和存储介质
US10956480B2 (en) System and method for generating dialogue graphs
US20220207392A1 (en) Generating summary and next actions in real-time for multiple users from interaction records in natural language
WO2020048296A1 (zh) 机器学习方法、设备及存储介质
US10885119B2 (en) Method and system for processing multimodal user queries
WO2020038253A1 (zh) 关键词的提取方法、系统和存储介质
US20220237376A1 (en) Method, apparatus, electronic device and storage medium for text classification
CN114341864A (zh) 在人工智能对话系统中对主体值的自主学习
US11381683B2 (en) System, device, and method of performing data analytics for advising a sales representative during a voice call
US20210374339A1 (en) Generating sentiment analysis of content
EP3751493A1 (en) A system, device, and method of performing data analytics for advising a sales representative during a voice call
WO2020048295A1 (zh) 音频标签的设置方法、设备和存储介质
US20230013842A1 (en) Human assisted virtual agent support
US20230067274A1 (en) Incomplete problem description determination for virtual assistant user input handling
US20230162058A1 (en) Systems and methods for implementing playbooks
US11483262B2 (en) Contextually-aware personalized chatbot
CN110532565B (zh) 语句处理方法及装置、以及电子设备
US11449471B2 (en) Sharing a modified file
WO2020024699A1 (zh) 客服机器人的自检方法、电子设备和计算机可读存储介质
US20200151241A1 (en) Webpage component replication system and method

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 19849889

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 24.06.2021

122 Ep: pct application non-entry in european phase

Ref document number: 19849889

Country of ref document: EP

Kind code of ref document: A1