CN109150701B

CN109150701B - Customer service system service method

Info

Publication number: CN109150701B
Application number: CN201811075502.5A
Authority: CN
Inventors: 朱频频
Original assignee: Shanghai Xiaoi Robot Technology Co Ltd
Current assignee: Shanghai Xiaoi Robot Technology Co Ltd
Priority date: 2015-12-30
Filing date: 2015-12-30
Publication date: 2020-07-10
Anticipated expiration: 2035-12-30
Also published as: CN109150701A; CN109245995B; CN109245995A; CN105553833A; CN105553833B

Abstract

A customer service system service method, comprising: when the robot customer service receives a conversation message from a customer client, detecting whether the robot customer service is in a conversation suspension state; when the robot customer service detects that the robot customer service is in a session suspension state, executing the transmission operation of session messages between the customer client and the customer service client; when the robot customer service detects that the robot customer service is not in a conversation suspension state, the robot customer service autonomously processes the conversation message; and in the autonomous processing process, when a preset conversation suspension condition is met, executing a process of switching the robot customer service to a conversation suspension state. The scheme can realize the cooperative work of the manual customer service and the robot customer service, thereby enhancing the service effectiveness of the customer service system and improving the user experience of customers.

Description

Customer service system service method

The application is a divisional application with the application date of 2015, 12, and 30 and the application number of 201511025278.5, and the invention provides a customer service system, a service method thereof and robot customer service.

Technical Field

The present invention relates to the field of instant messaging technologies, and in particular, to a customer service system service method and a computer-readable storage medium.

Background

With the rapid development of the internet and the improvement of service consciousness of people, network customer service has been popularized in various industries and goes deep into various links of daily business service.

The existing customer service system generally comprises machine customer service and artificial customer service, wherein the machine customer service is generally an instant messaging tool based on a webpage, the artificial customer service is generally an instant messaging tool embedded in the webpage, and each artificial customer service client is manually maintained. In the customer service process, when a session message from a customer is received, the session message is served by a machine customer service. And when the customer thinks that the machine customer service cannot solve the problem provided by the machine customer service, the customer clicks the manual customer service manually again to consult the manual customer service.

The customer service system can quickly respond to the requirements of customers to a certain extent. However, in the actual application process, the machine customer service identifies the content of the conversation message input by the customer, searches the stored resource database for data matched with the content of the conversation message, and sends the matched data to the user client for response, and for the same question, repeated answer is usually performed mechanically, which is easy to irritate the customer and affect the customer experience. However, the manual customer service cannot be on-line all the time, and usually faces a plurality of customers at the same time, so that the customer service cannot respond in time. Therefore, the current customer service system is difficult to effectively meet the service requirements of customers, and the user experience of the customers is influenced.

Disclosure of Invention

The technical problem solved by the invention is how to enhance the service effectiveness of the customer service system and improve the user experience of customers.

In order to solve the above technical problem, an embodiment of the present invention provides a customer service system service method, including:

when the robot customer service receives a conversation message from a customer client, detecting whether the robot customer service is in a conversation suspension state;

when the robot customer service detects that the robot customer service is in a session suspension state, executing the transmission operation of session messages between the customer client and the customer service client;

when the robot customer service detects that the robot customer service is not in a conversation suspension state, the robot customer service autonomously processes the conversation message; in the autonomous processing process, when a preset conversation suspension condition is met, executing a process of switching the robot customer service to a conversation suspension state;

when the robot customer service and the session service of the customer client are in a session state and receives a session message sent to the customer client from the customer service client, the session state of the robot in the session service with the customer client is switched to a session suspension state, and the transfer operation of the session message between the customer service client and the customer client is executed.

Optionally, the autonomously processing, by the robot customer service, the session message when the robot customer service detects that the robot customer service is not in the session suspended state includes:

identifying content of the session message;

determining whether the conversation message can be autonomously processed or not according to the content of the conversation message;

and when determining that the conversation message can be autonomously processed by the client, extracting a response message matched with the content of the conversation message from the stored service resource and sending the response message to the client.

Optionally, in the autonomous processing, when a preset session suspension condition is met, executing a process of switching the robot service to a session suspension state includes:

when determining that the conversation message cannot be autonomously processed or identified, setting the state of the conversation message in the subsequent conversation service of the client side as a conversation suspension state, extracting a placating response message matched with the content of the conversation message, sending the placating response message to the client side, and sending the conversation message which cannot be autonomously processed or identified to a customer service client side.

Optionally, the method further comprises:

and the robot customer service sends the soothing response message corresponding to the session message which cannot be processed or identified autonomously and the session message which cannot be processed or identified autonomously to the customer service client.

Optionally, after the robot customer service detects that it is in the session suspended state, the method further includes:

and when the robot customer service does not receive the session message from the customer service client, the robot customer service identifies the content of the session message, extracts the soothing response message matched with the content of the session message and sends the soothing response message to the customer client.

Optionally, the method further comprises:

and when the session between the customer service client and the customer client is determined to be finished, restoring the state of the customer service client in the session service with the customer client to be a session state.

Optionally, the determining that the session between the customer service client and the customer client is over includes at least one of:

determining that no session message is transmitted between the customer service client and the customer client for a preset duration;

and detecting a session completion signal of the customer service client.

Optionally, the customer service client sends the session completion signal when detecting that the session window is closed or when detecting that the session message sent by the customer service client includes content indicating that the session is completed.

Optionally, the determining whether the session message can be autonomously processed according to the content of the session message includes:

determining whether the robot customer service can answer the content of the session message according to the language experience data;

when it is determined that the robot customer service cannot answer according to the language experience data, determining a scene category to which the content of the conversation message belongs;

and when the robot customer service is determined to be capable of autonomously answering according to the language experience data or the scene category to which the content of the session message belongs exists, determining that the robot customer service is capable of autonomously processing the session message.

To solve the above technical problem, the present invention further provides a computer-readable storage medium having a program stored thereon, the program implementing the steps of the customer service system service method described above when executed.

Compared with the prior art, the technical scheme of the embodiment of the invention has the following beneficial effects:

when the robot customer service receives the conversation message from the customer client, by detecting whether the robot customer service is in a conversation suspension state, and when determining that the client is in a session suspension state, executing the transmission operation of the session message between the client and the customer service client, that is, the customer is directly served by the manual customer service at the customer service client, and when determining that the customer is not in the session suspension state, the session message is handled autonomously by the robot customer service itself, and in the course of autonomous handling, when a preset conversation suspension condition is met, executing a process of switching the robot customer service to a conversation suspension state, namely, the robot customer service mainly serves the customer at the customer client independently, and when certain conditions are met, and the customer service is automatically switched to be manually served, so that the cooperative work of manual customer service and robot customer service is realized. On one hand, when the robot customer service is in a conversation suspension state, only the transmission of the conversation message is needed without any autonomous processing, so that the service resources of the robot customer service can be saved; on the other hand, the robot customer service can independently work only when the robot customer service is not in a session suspension state, so that the priority of the robot customer service is higher, the robot customer service can be prevented from seizing the session midway, and a customer can obtain consistent service experience, so that the service effectiveness of a customer service system can be improved, and the user experience of the customer can be improved.

Further, the robot customer service, when receiving the conversation message from the customer client and not being in the conversation suspension state, identifies the content of the conversation message and determines whether the robot customer service can autonomously process the conversation message. And when determining that the session message can be autonomously processed, extracting a response message matched with the content of the session message from the stored service resource and sending the response message to the client, so that the client can obtain a timely and effective response.

Further, when it is determined that the session message cannot be autonomously processed or identified, on one hand, the state of the client in the subsequent session service with the client is set to be a session suspension state, so that processing resources can be saved; on the other hand, by timely sending the soothing response message matched with the content of the session message, the problem that a customer cannot reasonably respond to the conversation message to generate a fidget mood can be avoided, the conversation message which cannot be autonomously processed or identified by the customer can be sent to the customer service client side without manual selection of the customer, and when the robot customer service cannot effectively process the conversation message, manual customer service timely intervention can be realized to effectively solve the problem of the customer. In a word, the scheme can timely and effectively solve the problem of the customer through the cooperative work of the robot customer service automation and the manual customer service, the customer cannot feel the difference between the robot customer service and the manual customer service, the emotion can be effectively placated in the process of waiting for response, and better interactive experience is obtained.

Further, the soothing response message corresponding to the session message which cannot be processed or identified autonomously and the session message which cannot be processed or identified autonomously are sent to the customer service client together, so that the artificial customer service can see the content of the response message replied by the robot customer service to the customer, the consistency and consistency of response are convenient to maintain, the consistent service experience can be optimized, and the user experience of the customer is improved.

Further, when the robot customer service is in a session suspension state, if the session message from the customer service client is not received, that is, the manual customer service is not yet intervened, the content of the session message is identified, and when the content of the session message can be identified, the corresponding soothing response message is extracted from the stored service resources, and in the process of interaction between the manual customer service and the customer, the robot customer service does not perform any identification, autonomous processing or response on the session message, but only transmits the session between the customer service client and the customer client, so that the degree of coordination between the robot customer service and the manual customer service can be further improved, and the anxiety of the customer in the waiting process can be further relieved.

Further, when the robot customer service is in a session state in the session service with the customer client and receives a session message from the customer service client, the session state of the robot customer service in the session service with the customer client is switched to a session suspension state, and the transmission operation of the session message between the customer service client and the customer client is executed.

Further, when the robot customer service determines that the session between the customer service client and the customer client is finished, the state of the robot customer service in the subsequent session service between the robot customer service client and the customer client is recovered to be a session state, namely, when the intervention of the manual customer service is finished, the state of the robot customer service is automatically switched back, the seamless connection between the manual customer service and the robot customer service is realized, and therefore the service experience of the customer can be further optimized.

Further, by determining that no session message is transmitted between the customer service client and the client for a preset time, or when a session window is detected to be closed or when a session message sent by the customer service client contains content for displaying session completion, a session completion signal automatically sent by the customer service client can automatically identify completion of manual customer service intervention, so that the manual customer service operation can be reduced, the workload of manual customer service is reduced, and convenience is provided for the manual customer service.

Furthermore, the robot customer service is subjected to big data training, so that the robot customer service can identify the scene type of the session message, and can rapidly determine whether the session message can be processed and carry out subsequent processing, thereby improving the response speed of the customer service and further enhancing the interactive experience of customers.

Further, when the robot customer service determines that the robot customer service can autonomously answer according to the language experience data, the response message matched with the content of the session message is extracted from the stored language experience data, and when the robot customer service determines that the robot customer service cannot autonomously answer according to the language experience data, the response message matched with the content of the session message in the scene is selected, so that the robot customer service can answer according to the language experience data and the customer service resource data, and the matched resource data corresponds to the scene of the customer service, so that the breadth of the response message which can be answered by the robot customer service can be expanded, and the accuracy of the response is improved.

Further, a soothing response message representing that the customer service is in a state of searching for answers to the problems is fed back to the customer and is executed simultaneously with the process that the customer service client searches for the contents of the session message and gives a solution, so that on one hand, the actual state of the customer service can be reflected without manual input of the customer service, the response operation time of the manual customer service is saved, and the mood of the customer can be placated; on the other hand, the manual customer service can put more time into solving the problems of the customers, so that the responses which are more satisfied to the customers can be more quickly given.

Drawings

FIG. 1 is a schematic diagram of a customer service system according to an embodiment of the present invention;

FIG. 2 is a flow chart of a customer service system service method according to an embodiment of the present invention;

FIG. 3 is a flow chart of another customer service system servicing method in an embodiment of the present invention;

FIG. 4 is a schematic diagram of a customer client interactive interface in accordance with an embodiment of the present invention;

FIG. 5 is a schematic diagram illustrating a session state switching process in a customer service process according to an embodiment of the present invention;

fig. 6 is a schematic structural diagram of a robot customer service in an embodiment of the present invention.

Detailed Description

As described above, in the customer service process of the current customer service system, the machine customer service and the manual customer service independently serve the customer, that is, the machine customer service first serves the customer. And when the customer thinks that the machine customer service cannot solve the problem provided by the machine customer service, the customer clicks the manual customer service manually again to consult the manual customer service. According to the scheme, the robot customer service system has the problems that mechanical repeated answering is easy to irritate customers, manual customer service cannot be online all the time, the customers are numerous in general, and the customers cannot reply in time, so that the service requirements of the customers cannot be effectively met by the existing customer service system.

In view of the above problem, an embodiment of the present invention provides a customer service system, which is shown in fig. 1, and includes: a customer client 11, a customer service client 12 adapted to have a session with said customer client 11, and a robot service 13, said customer service client and said customer client having a session through said robot service.

In the customer service system of the embodiment of the invention, the robot customer service and the manual customer service cooperate to receive session messages from the customer client side uniformly, the robot customer service executes the transmission operation of the session messages between the customer client side and the customer service client side by detecting whether the robot customer service is in a session suspension state or not and executing the transmission operation of the session messages between the customer client side and the customer service client side when the robot service is determined to be in the session suspension state, namely, the manual customer service at the customer service client side directly serves the customer, and when the robot service is determined not to be in the session suspension state, the robot customer service self processes the session messages autonomously, and in the autonomous processing process, when a preset session suspension condition is met, the process of switching the robot customer service to the session suspension state is executed, namely, the robot customer service at the customer client side is served by the robot service autonomously, and when a certain condition is met, the process of automatically switching to the manual customer service is executed, thereby realizing the cooperative work of the manual customer service and the robot customer service.

In order to make the aforementioned objects, features and advantages of the present invention comprehensible, embodiments accompanied with figures are described in detail below.

Referring to fig. 1 and fig. 2, a method for providing a service to a customer service system according to an embodiment of the present invention is described in detail below with specific steps.

S21, the robot service receives a session message from the customer client.

In particular implementations, the session message may take a variety of forms. For example, the language may be any one of a sentence composed of natural language words, a data link, voice, an image or a video, or a combination of multiple forms.

S22, the robot customer service detects whether the robot is in the conversation suspension state, and when the robot is detected to be in the conversation suspension state, the step S23 is executed; when detecting that itself is not in the session suspended state, step S24 is executed.

In a specific implementation, in order to save processing resources of the robot customer service and realize cooperative work of the robot customer service and the manual customer service, a state in which the robot customer service autonomously processes a session message from a client may be set to a "session state" or a "non-suspended state", and a state in which the robot customer service cannot autonomously process or cannot recognize the content of the session message may be set to a "session suspended state".

And S23, executing the transmission operation of the session message between the customer client and the customer service client.

S24, the robot customer service autonomously processes the conversation message.

In an embodiment of the present invention, a process of the robot customer service autonomously processing the session message is as follows: firstly, identifying the content of the session message, then determining whether the session message can be autonomously processed according to the content of the session message, and extracting a response message matched with the content of the session message from the stored service resources and sending the response message to the customer client when determining that the session message can be autonomously processed.

S25, detecting whether the preset conversation suspension condition is satisfied, if yes, executing the step S26; if not, step S24 is performed.

S26, executing the process of switching the robot service to the session suspended state.

In the autonomous processing process of the robot customer service, when a preset conversation suspension condition is met, the robot customer service can be switched to a conversation suspension state. For example, when it is detected that the session message cannot be autonomously processed or recognized, the following operations may be performed:

setting the state of the user in the subsequent session service of the client as a session suspension state, extracting a placation response message matched with the content of the session message and sending the placation response message to the client, and sending the session message which cannot be autonomously processed or identified to the customer service client.

The soothing response message is a message that can stabilize the uneasy and negative emotions of the customer, and in the specific implementation, the soothing response message can be presented in the forms of natural character language, pictures, voice, music and the like.

In specific implementation, the soothing response message corresponding to the session message which cannot be processed or identified autonomously and the session message which cannot be processed or identified autonomously can be sent to the customer service client together, so that the artificial customer service can see the content of the response message replied by the robot customer service to the customer, the consistency and consistency of response can be conveniently maintained, the consistent service experience can be optimized, and the user experience of the customer can be improved.

By adopting the embodiment, on one hand, when the robot customer service is in the conversation suspension state, only the transmission of the conversation message is needed without any autonomous processing, and the service resource of the robot customer service can be saved; on the other hand, the robot customer service can independently work only when the robot customer service is not in a session suspension state, so that the priority of the robot customer service is higher, the robot customer service can be prevented from seizing the session midway, and a customer can obtain consistent service experience, so that the service effectiveness of a customer service system can be improved, and the user experience of the customer can be improved.

The following explains the principle of cooperative work of robot customer service and manual customer service in the embodiment of the present invention by using another customer service method of a customer service system in the embodiment of the present invention, and with reference to fig. 3, the specific steps are as follows:

s31, receiving the session message from the customer client.

In particular implementations, the session message may take a variety of forms, as previously described. For example, the language may be any one of a sentence composed of natural language words, a data link, voice, an image or a video, or a combination of multiple forms.

In an embodiment of the present invention, the form of the conversation message includes a sentence composed of natural language words. The customer can describe the description object through the sentence composed by the natural language. For example, if the description object is a product, the session message may be a consultation message of product-related information such as size, shape, color, texture, production date, and shelf life of the product, or a consultation message of after-market service of the product. If the description object is a service, the session message may be service content, service duration, notice, and consulting information of service-related content such as service staff providing the service.

In a specific implementation, the conversation message can be presented in a form of voice, image or video besides being presented in a text form. For example, the customer may have a conversation with the customer service in the form of an audio or instant audio conversation or a video conversation.

In a specific implementation, the customer can also have a conversation with the customer service in the form of pictures through the customer client. The picture may be a static picture, a dynamic picture, or a combination thereof with semantic information, or a photograph or screen shot of the product or service object for which the customer session is directed.

In particular implementations, the customer may also send a link containing a product or service object in the form of a session message.

It is understood that, in a specific implementation, the conversation message may also be implemented by combining two or more forms of sentences composed of natural language words, data links, voice, images or videos. For example, a customer communicates with a customer service via voice or video, while sending the corresponding data link to the customer service via the customer client.

S32, determining the state of the user in the session service, and if the state is the session suspension state, executing the step S37; if the non-session suspended state exists, step S33 is executed.

If the robot service determines that the state of the robot service in the session service with the customer client is a non-session suspended state, step S33 is executed. If the robot service determines that the state of the robot service in the session service of the client is the session suspended state, the subsequent client questions from the client under the current session are not automatically replied, but are sent to the service client for manual service processing, i.e., step S37 is executed.

The following steps S33 to S36 are the robot customer service autonomous processing procedure, and step S37 is a procedure of triggering manual intervention.

S33, identifying the content of the conversation message, if it can be identified, executing step S34; otherwise, step S36 is executed.

In a specific implementation, the content of the conversation message can be identified by the robot customer service, and different identification methods can be adopted for identification processing according to different conversation message forms. For example, a sentence composed of natural language words may be recognized by word segmentation and semantic recognition. For another example, received voice can be converted into text, and then word segmentation and semantic recognition are further performed. If the conversation message contains an image, the text or pattern lines in the image can be extracted and further identified. If it can be identified, step S34 is performed.

It should be noted that some conversation messages may not be recognized, such as video or images with large information amount or complicated message form, which may exceed the recognition capability of the robot customer service, and then step S36 may be executed.

S34, determining whether autonomous processing is possible, if so, executing step S35; if not, step S36 is performed.

The robot service may determine whether autonomous processing is possible according to the contents of the recognized conversation message.

In particular implementations, the robot may confirm whether autonomous processing is possible based on stored service resources. The stored service resources may include stored language experience data and customer service resource data.

Specifically, it may be determined first whether the content of the session message can be answered autonomously from the language-experience data, and when it is determined that the content of the session message can be answered autonomously from the language-experience data, it is determined whether autonomous processing is possible; when it is determined that the robot customer service cannot autonomously answer the conversation message according to the language experience data, it may be further determined that a scene category to which the content of the conversation message belongs exists.

In order to improve the autonomous processing capability of the robot, big data machine training can be performed on the robot customer service to identify the scene type of the session message. In specific implementation, the session messages with the client side can be classified and split according to scenes in advance to obtain the customer service resource data of different scene categories.

In an embodiment of the invention, in the scene classification and machine training process, the session message not only comprises the session message between the robot customer service and the customer client, but also comprises the session message between the artificial customer service and the customer client, all the session messages are classified and split according to scenes, and the artificial customer service is not regarded as an independent individual but is used as an element of a specific scene in the whole service link, so that the whole customer service working mode is reconstructed, the working efficiency of the artificial customer service can be effectively improved under the condition of a large number of customers, and the customer service experience is optimized. Wherein elements of a scenario may include customer and customer service and session message content for both different service scenarios.

And S35, extracting the matched response message from the stored service resource and sending the response message to the customer client.

As previously described, the service resources may include stored language experience data and customer service resource data. Accordingly, in a specific implementation, when it is determined that a response can be autonomously made based on the language-experience data, extracting a response message matching the content of the conversation message from the stored language-experience data; and when determining that the answer cannot be made according to the language experience data autonomously, extracting an answer message matched with the content of the conversation message in the scene from the stored customer service resource data.

For example, when the content of the conversation message is identified as "hello", the response message that results in a match by extracting in the stored language experience data is: "good, on woollen cloth. What can you be served? ". As another example, when the content of the identified conversation message is: "did i ask me to purchase a blue dress for shipment? And if so, extracting order information corresponding to the customer from the language experience data and the customer service resource data to obtain a response message: "you are just! Your purchased dress has been shipped ". The response message may also be attached with logistics information of the order.

And S36, setting the self in a suspended state in the subsequent session service of the customer client, replying a soothing response message corresponding to the session message to the customer client, and sending the session message and the soothing response message to the customer service client.

The robot service can complete the process of switching to the session suspended state through the step S36, thereby realizing the cooperative work of the robot service and the manual service. The following is illustrated by a specific application scenario:

and when the robot cannot process the conversation message autonomously or cannot recognize the conversation message, the machine customer service the extracted soothing response message matched with the content of the conversation message. In particular implementations, the soothing response message is adapted to characterize the customer service as being in a state of seeking answers to questions in order to avoid a petitious mood resulting from the customer's conversation messages failing to respond for a long period of time. For example, when the content of the conversation message is identified as: when the robot customer service determines that the robot customer service cannot autonomously process the leather shoes, on one hand, the robot customer service sets the robot customer into a session suspension state in subsequent session service with the customer client; on the other hand, the matching result is that: kaki, good. Wait a little to see your shoes. Sending the placating response message to the customer client so that the customer can reply in time and have patience to wait, and sending a session message of 'I want to buy the double man leather shoes' to the customer service end so that manual service is handled. Some soothing response messages corresponding to the situation that the robot customer service cannot recognize the content of the specific conversation message can also be set in advance, for example, for the received pictures, when the contents cannot be recognized, the corresponding soothing response messages can be set to be 'please look for a little, and i see first'.

And S37, transmitting the session message to the corresponding customer service client.

When the robot customer service is in a conversation suspension state, the conversation messages from the customer client and the customer service client are respectively forwarded to the other side without any autonomous processing, and the robot customer service only forwards the conversation messages without any identification, processing or response through the intervention of the manual customer service, so that the service resources of the robot customer service can be saved.

It can be known from the above embodiments that, when receiving a session message from a client, a robot service first detects whether the robot service is in a session suspended state, and when determining that the robot service is not in the session suspended state, the robot service performs an operation of identifying the content of the session message, and when determining that the robot is in the session suspended state, the robot service performs session interaction with the client through the service client, thereby implementing better cooperative work between the robot service and the robot service, on one hand, when the robot service is in the session suspended state, only the transmission of the session message is required without performing any autonomous processing, which can save the service resources of the robot service, on the other hand, the robot service can work only when the robot service is not in the session suspended state, so that the priority of the robot service is higher, which can avoid the robot service seizing the midway session, therefore, the customer can obtain consistent service experience, and the user experience of the customer is improved.

In specific implementation, in order to further relieve the anxiety of customers in the waiting process, when the robot customer service is in a conversation suspension state, if the conversation message from the customer service client is not received, namely the manual customer service has not been subjected to intervention processing, the content of the conversation message can be identified, and when the content of the conversation message can be identified, a corresponding soothing response message is extracted from the stored service resources; after receiving the session message from the customer service client, namely the manual intervention processing, in the process of interaction between the manual customer service and the customer, the robot customer service does not perform any identification, autonomous processing or response on the session message, but only transmits the session between the customer service client and the customer client.

In a specific implementation, when it is determined that the session between the customer service client and the customer client is ended, the robot service may autonomously restore its own state in the session service with the customer client to a session state, i.e., a non-session suspended state.

When the intervention of the manual customer service is finished, the seamless connection between the manual customer service and the robot customer service is realized by automatically switching the state of the robot customer service back, so that the service experience of customers can be further optimized. In particular implementations, there are a number of ways to determine that a session between the customer service client and the customer client is over. For example, it is determined whether there is no session message transmission between the customer service client and the customer client for a preset duration, and when the preset duration is reached, it is determined that the session between the customer service client and the customer client is ended. For another example, whether a session completion signal from the customer service client is detected, when the session completion signal from the customer service client is detected, it is determined that the session between the customer service client and the customer client is ended. In an embodiment of the present invention, the customer service client may send the session completion signal when detecting that the customer service client closes the session window. In another embodiment of the present invention, the customer client may also send the session completion signal when detecting that the session message sent by the customer service client includes content indicating that the session is completed.

It will be appreciated that in particular embodiments, the methods described in the different embodiments above may be used in combination, as desired.

The method comprises the steps that whether conversation information is transmitted between the customer service client and the customer client for a preset time length or not is determined, or when a conversation window is detected to be closed or when a conversation message sent by the customer service client contains content for displaying conversation completion, a conversation completion signal sent by the customer service client automatically identifies completion of manual customer service intervention, so that manual customer service operation can be reduced, workload of manual customer service is reduced, and convenience is brought to manual customer service.

In specific implementation, when the client is in a session state in a session service with the client and receives a session message from the customer service client, the session state of the client in the session service with the client is switched to a session suspension state, and a transmission operation of the session message between the customer service client and the customer service client is executed, that is, in a process of cooperative work of manual customer service and robot customer service, the priority of the manual customer service is higher, and the manual customer service can actively intervene at any time as required, so that better service can be provided for the customer.

It is to be understood that, in the above embodiments, the processing flow when a session message is received from a client is described, and each received session message may be processed according to the above flow.

In order to enable those skilled in the art to better understand and implement the present invention, the following detailed description is provided by way of a specific session scenario.

Referring to fig. 4, a schematic diagram of a customer client interaction interface is shown. The interactive interface is a session interactive interface of the customer A and the customer service B, and is displayed through a customer client of the customer A. In the customer service system, no matter robot customer service or manual customer service, the customer A is displayed as a uniform image customer service B, and an actual service person may be the robot customer service or some manual customer service behind a customer service client. The boxes a and B respectively represent identification information of the customer a and the customer service B, and in a specific implementation, the boxes a and B can be displayed by using information such as customized head images and nicknames of the customer a and the customer service B. The time information "5:44 PM" and "5:50 PM" in the dashed

boxes

48, 49 indicate the time at which the session displayed on the customer client interface of customer A occurred. The right-side session message is the session message content sent by the customer A, and the left-side session message is the response message content of the customer service B. The following describes in detail the processing procedure executed by the customer service system during this session interaction:

the method comprises the steps that a customer A sends a conversation message 'hello' through a customer client, when the robot customer service confirms that the robot customer service does not stay in a conversation suspension state, the conversation message is identified, after identification is obtained, whether a response can be made according to language experience data is confirmed, and when the response can be made according to the language data, a matched response message 'hello', where the response can be made, is extracted. What can you be served? "and sent to customer client of customer a. And then the customer A sends a conversation message of ' i want to buy double-leather shoes ' through the customer client, and after the robot customer service confirms that the robot is not in a conversation suspension state and identifies the state, and confirms that the robot cannot process the information autonomously, the comfort response message ' kahou matched with the conversation message of ' i want to buy double-leather shoes ' is extracted. Wait a little to see your shoes. Meanwhile, the robot customer service may set a session state of the robot in a subsequent session service with the customer client of the customer a to a session suspended state, and simultaneously send the session message "i want to buy double leather shoes" to the customer service client to be processed by the manual customer service. Customer a replies to the session message "good, thank you".

It should be noted that, the manual customer service may not be in place, or because too many customers need to be serviced by the manual customer service cannot intervene in the process in time, after a while, customer a sends a session message "ask for a question" through the customer client? "is the robot service received the session message" asked? And after the message is detected to be in the session suspension state, identifying the content of the message, extracting and replying a corresponding soothing response message, namely 'slightly waiting, i still look like' and sending the message to the customer client of the customer A.

Then, a manual customer service intervenes, a conversation message "hello" is sent through a customer service client, and I feel that the following shoe is very suitable for you. http// www.tmall.com/", the robot customer service transmits the message sent by the artificial customer service to the customer client of the customer A without any processing, the customer A replies the session message of kay, good, i consider seeing after seeing the session message, and the artificial customer service replies the reply session message of kay, good after the customer service client receives the session message. You can just read you to contact me at any time ", and" thank you for you being on, 88 ".

According to the above processing flow, the

session messages

41 and 42 are messages which are processed and responded by the robot, the session messages 43 and 44 are placating response messages generated by the robot customer service, and the

session messages

45, 46 and 47 are session messages responded by the robot customer service.

It will be appreciated that in particular implementations, the

session messages

46 and 47 may also be automatically generated response messages for the robot service. For example, if the robot service does not receive the response message from the service client within a preset time after receiving the session message "kayeto, good, i consider to see" and can confirm that the manual intervention is finished, the session message of the service response may be identified. After identifying the content of the session message, in one embodiment of the invention, the "kay, ok" may be answered directly from the stored linguistic empirical data. You can just like to contact me at any time "and" thank you for you being on, 88 ". In another embodiment of the invention, after the content of the session message is identified, if the self-answering cannot be carried out according to the language experience data, the scene type corresponding to the session message is matched, and the matched answer message is extracted from the stored service resources. You can just like to contact me at any time "and" thank you for your light, 88 "and send them to the customer client of customer a in turn.

In order to make those skilled in the art better understand and implement the present invention, a state switching process of the robot service in the session service in the service process is described in detail below with reference to a schematic diagram of a session state switching process in the service process in the embodiment of the present invention shown in fig. 5. For example, the initial state of the robot service is the session state 51 (non-session suspended state), and when the robot service does not recognize the received session message, or the robot service cannot autonomously process the session message, or the human service intervenes in the session, the robot service switches to the session suspended state 52. And when the robot customer service is in the suspended state of conversation, if the signal of the intervention completion of the manual customer service is detected, the robot customer service can carry out conversation recovery 53, and through the conversation recovery 53, the robot customer service can continue the autonomous service, namely, the robot customer service is switched to the conversation state 51, and can autonomously serve the customer. After the robot service is in the session suspended state or the manual intervention is completed, if a new session message from the customer client is not received, the whole session process is ended, i.e. in the session end state 54.

In order to enable those skilled in the art to better understand and implement the present invention, an embodiment of the present invention further provides a robot customer service capable of implementing the customer service method described above, and referring to fig. 6, the robot customer service 60 may include: a first message interaction unit 61, a status detection unit 62, an autonomous processing unit 63, a status control unit 64, and a second message interaction unit 65, wherein:

the first message interaction unit 61 is adapted to receive a session message from the customer client and return a reply message matching the content of the session message.

A state detection unit 62 adapted to detect a state of the robot customer service in a conversation service with the customer client when the first message interaction unit 61 receives a conversation message from the customer client.

An autonomous processing unit 63 adapted to autonomously process the session message and return the obtained reply message to the customer client through the first message interaction unit when the state detection unit 62 detects that the robot customer service is not in a session suspended state.

A state control unit 64 adapted to control a state of the robot customer service in a subsequent session service with the customer client to be a session suspended state when a preset suspension condition is satisfied during the autonomous processing by the autonomous processing unit 63.

A second message interaction unit 65, adapted to transmit the session message received by the first message interaction unit 61 to the customer service client when the state detection unit 62 detects that the robot customer service is in the session suspension state; and transmits the session message from the customer service client to the corresponding customer client through the first message interaction unit 61.

In a specific implementation, the form of the session message may include at least one of: sentences composed of natural language words, data links, voice, images and video.

Wherein the image comprises at least one of: a still or moving picture with semantic information, or a photograph or screen shot of the product or service object targeted for the customer session.

By adopting the robot customer service, on one hand, when the robot customer service is in a conversation suspension state, only the transmission of conversation messages is needed without any autonomous processing, so that the service resources of the robot customer service can be saved; on the other hand, the robot customer service can independently work only when the robot customer service is not in a session suspension state, so that the priority of the artificial customer service is higher, the robot customer service can be prevented from seizing the session midway, the customer can obtain consistent service experience, the cooperative work of the robot customer service and the artificial customer service is realized, the service effectiveness of a customer service system can be improved, and the user experience of the customer is improved.

In a specific implementation, as shown in fig. 6, the autonomous processing unit 63 may include: a recognition unit 631, a capability judgment unit 632, a storage unit 633, and a first reply message generation unit 632, wherein:

an identifying unit 631 adapted to identify the content of the session message received by the first message interacting unit 61 when the state detecting unit 62 detects that the robot service is not in the session suspended state.

A capability judging unit 632, adapted to determine whether the robot customer service can autonomously process the conversation message according to the content of the conversation message identified by the identifying unit 631.

A storage unit 633 adapted to store service resources. The service resources may include: language experience data, customer service resource data, and the like.

A first response message generating unit 634, adapted to extract a response message matched with the content of the conversation message from the service resource stored in the storage subunit 633 when the ability judging unit 632 determines that the robot customer service can autonomously process the conversation message.

By adopting the robot customer service, when the conversation message from the customer client is received and the robot customer service is not in the conversation suspension state, the content of the conversation message is identified and whether the robot customer service can autonomously process the conversation message or not is determined. And when determining that the session message can be autonomously processed, extracting a response message matched with the content of the session message from the stored service resource and sending the response message to the client, so that the client can obtain a timely and effective response.

In an embodiment of the present invention, the robot customer service further includes a machine training unit 66, adapted to perform big data machine training on the robot to identify a scene category to which the session message belongs, where the scene category is obtained by classifying and splitting the session message with the client according to the scene.

By carrying out big data training on the robot customer service, the robot customer service can identify the scene type of the session message, so that whether the session message can be processed or not can be quickly determined and the subsequent processing can be carried out, the response speed of the customer service can be improved, and the interactive experience of customers can be further enhanced.

In a specific implementation, the capability determining unit 632 may include:

a first judging subunit (not shown in the figure), adapted to determine whether the robot customer service can autonomously answer the content of the session message according to the language experience data;

a second judging subunit (not shown in the figure), adapted to determine a scene category to which the content of the conversation message belongs when the first judging subunit determines that the robot customer service cannot autonomously answer according to the language experience data.

In a specific implementation, the first reply message generating unit 634 may include:

a first generating subunit (not shown) adapted to, when it is determined that the robot customer service can autonomously answer based on the language-experience data, extract a response message matching the content of the conversation message from the language-experience data stored in the storage unit 633;

a second generating subunit (not shown) adapted to, when it is determined that the robot service cannot answer according to the language-experience data, extract a response message matching the content of the conversation message in the scene from the service resource data stored in the storage unit 633.

By adopting the robot customer service in the embodiment, when the robot customer service determines that the robot customer service can answer according to the language experience data autonomously, the response message matched with the content of the session message is extracted from the stored language experience data, and when the robot customer service determines that the robot customer service cannot answer according to the language experience data autonomously, the response message matched with the content of the session message in the scene is selected, so that the robot customer service can answer according to the language experience data and the customer service resource data, and the matched resource data corresponds to the scene of the customer service, so that the breadth of the response message which can be answered by the robot customer service can be expanded, and the accuracy of the response can be improved.

In a specific implementation, as shown in fig. 6, the state control unit 64 is adapted to control the state of the robot service in the subsequent session service with the customer client to be a session suspended state when the capability determination unit determines that the robot service cannot autonomously process or recognize the session message.

Accordingly, the autonomous processing unit 63 may further include: a second response message generating unit 635 adapted to extract a soothing response message matching the content of the session message from the service resource stored in the storage unit 633 when the state control unit 64 controls the state of the robot service in the subsequent session service with the customer client to be a session suspended state. The first message interaction unit 634 is further adapted to return the placingability response message matching with the content of the conversation message to the customer client.

The soothing response message is a message that stabilizes the uneasy and negative emotions of the customer, and may be presented in a natural language, a picture, voice, music, or the like in specific implementation. In an embodiment of the invention, the placticity response message is adapted to indicate that the customer service is in a state of looking for answers to questions.

The method has the advantages that the placating response message representing the state that the customer service is searching for the answer to the question is fed back to the customer, and the process of searching for the conversation message and giving the solution by the customer service client is executed simultaneously, so that on one hand, the actual state of the customer service can be reflected without manual input of the customer service, the response operation time of the manual customer service is saved, and the mood of the customer can be placated; on the other hand, the manual customer service can put more time into solving the problems of the customers, so that the responses which are more satisfied to the customers can be more quickly given.

In the embodiment, when it is determined that the robot customer service cannot autonomously process or recognize the session message, on one hand, the state of the robot customer service in the subsequent session service with the customer client is set to be a session suspension state, so that processing resources can be saved; on the other hand, by timely sending the soothing response message matched with the content of the session message, the problem that a customer cannot reasonably respond to the conversation message to generate a fidget mood can be avoided, the conversation message which cannot be autonomously processed or identified by the customer can be sent to the customer service client side without manual selection of the customer, and when the robot customer service cannot effectively process the conversation message, manual customer service timely intervention can be realized to effectively solve the problem of the customer. In a word, the scheme can timely and effectively solve the problem of the customer through the cooperative work of the robot customer service automation and the manual customer service, the customer cannot feel the difference between the robot customer service and the manual customer service, the emotion can be effectively placated in the process of waiting for response, and better interactive experience is obtained.

In a specific implementation, the second message interacting unit 65 is further adapted to send a soothing response message corresponding to the session message that cannot be processed or identified by itself and the session message that cannot be processed or identified by itself to the customer service client, so that the artificial customer service can see the content of the response message replied by the robot customer service to the customer, which is convenient for maintaining consistency and consistency of the response, thereby optimizing consistent service experience and improving user experience of the customer.

The autonomous processing unit 63 is further adapted to, when the state detecting unit 62 detects that the state of the robot service in the session service with the customer client is a session suspension state, and the second message interacting unit 65 does not receive the session message from the customer service client, identify the content of the session message, extract a soothing response message matching the content of the session message, and send the soothing response message to the customer client.

By adopting the robot customer service, when the robot customer service is in a conversation suspension state, if the conversation message from the customer service client is not received, namely the artificial customer service is not intervened, the content of the conversation message is identified, when the content of the conversation message can be identified, the corresponding soothing response message is extracted from the stored service resource, and in the interaction process of the artificial customer service and the customer, the robot customer service does not perform any identification, autonomous processing or response on the conversation message, but only transmits the conversation between the customer service client and the customer client, so that the degree of coordination between the robot customer service and the artificial customer service can be further improved, and the anxiety of the customer in the waiting process can be further relieved.

In a specific implementation, the state control unit 64 is further adapted to switch the state of the robot service in a subsequent session service with the customer client to a session suspended state and control the second message interaction unit 65 to perform a message interaction operation when the robot service is in a session state in the session service with the customer client and the second message interaction unit 65 receives a session message from the customer client. In the cooperative working process of the manual work and the robot customer service, the priority of the manual customer service is higher through the cooperation of the state control unit and the second message interaction unit in the robot customer service, and the manual customer service can actively intervene at any time according to the needs, so that better service can be provided for customers.

In a specific implementation, the robot customer service may further include: a session end detection unit 67 adapted to determine whether a session between the customer service client and the customer client is ended; the state control unit 64 is further adapted to restore the state of the robot service in a subsequent session service with the customer client to a session state when it is determined that the session between the customer service client and the customer client is ended.

When the robot customer service determines that the session between the customer service client and the customer client is finished, the state of the robot customer service in the subsequent session service between the robot customer service client and the customer client is recovered to be a session state, namely, when the intervention of the manual customer service is finished, the state of the robot customer service is automatically switched back, the seamless connection between the manual customer service and the robot customer service is realized, and therefore the service experience of the customer can be further optimized.

In an embodiment of the present invention, the session end detection unit 67 includes at least one of the following:

a first detection subunit (not shown in the figures) adapted to detect whether the second message interaction unit 65 is idle for a preset duration, and determine that the session between the customer service client and the customer client is ended when the preset duration is reached;

a second detecting subunit (not shown in the figure), adapted to detect a session completion signal from the customer service client, and determine that the session between the customer service client and the customer client is ended when the session completion signal from the customer service client is detected.

For example, the first detecting subunit determines that there is no session message transmitted between the customer service client and the client for a preset duration, or the second detecting subunit, when detecting that the session window is closed or detecting that the session message sent by the customer service client includes content indicating that the session is completed, automatically sends a session completion signal to the customer service client, so that completion of manual customer service intervention can be automatically identified, thereby reducing operations of manual customer service, reducing workload of manual customer service, and providing convenience for manual customer service.

The embodiment of the invention also provides a customer service system capable of realizing the customer service method, which refers to the customer service system shown in FIG. 1.

In implementations, the customer client and the customer service client may each be presented in a variety of forms. For example, the customer client and the customer service client may be built into a program in a browser and run automatically upon detecting that a browser page is opened. For another example, the client and the service client are independent application clients, are built in a desktop computer, a tablet computer, a mobile phone, and a vehicle-mounted terminal having a general computer function, and can be started and operated under the operation of the client. For another example, the customer client and the customer service client may also be built in a dedicated device customized for the customer by a product provider or a service provider, and the dedicated device may be used for the customer to consult and feedback with the product provider or the service provider during the sales exhibition or the service experience, so as to better meet the customer needs.

It is understood that the customer client and the customer service client performing the call may be implemented in the same or different manners, as long as the session can be implemented by using the same communication protocol.

Specific implementation manners such as specific structures and working principles of the customer client 11, the customer service client 12, and the robot customer service 13 can be referred to the descriptions in the above embodiments, and are not described herein again.

In a specific implementation, one robot service may simultaneously talk to a plurality of customer clients, and the robot service may also forward the received session message to different customer service clients. When a plurality of client sides and a plurality of service client sides are in a session state, when the robot service determines that the session message cannot be processed independently or identified, the session message of the client can be replied as soon as possible, and the session message from the client side and the corresponding appeasing response message can be sent to the service client side with the least task amount according to the task amount of the service client side in the session state. For example, the background has 10 manual customer services, which correspond to 10 different customer service clients, respectively, and the customer service robot may uniformly distribute the received session message to each customer service client according to the number of the customer clients corresponding to the session message processed by each current customer service client.

In specific implementation, in order to enable customers to obtain better consistent service and improve the service efficiency of each manual customer service, the robot customer service is suitable for sending the session message from the customer client to the customer service client with the least task amount according to the task amount of the customer service client in the session state when a plurality of customer clients in the session state and the plurality of customer service clients are determined to be incapable of processing the session message autonomously or recognizing the session message. In an embodiment of the present invention, the robot customer service may further send the soothing response message corresponding to the session message that cannot be processed or identified to the customer service client that has processed the session message of the customer client before, so as to avoid that the experience of the customer is affected by the fact that the manual customer service does not know the communication condition with the customer.

It can be understood that, according to needs, in the process of forwarding the message to the customer service client by the robot customer service, multiple allocation policies may be considered at the same time, each allocation policy may be configured with different priorities, and in a specific forwarding process, matching may be performed according to a priority order, which is not described in detail herein.

Those skilled in the art will appreciate that all or part of the steps in the methods of the above embodiments may be implemented by associated hardware instructed by a program, which may be stored in a computer-readable storage medium, and the storage medium may include: ROM, RAM, magnetic or optical disks, and the like.

Although the present invention is disclosed above, the present invention is not limited thereto. Various changes and modifications may be effected therein by one skilled in the art without departing from the spirit and scope of the invention as defined in the appended claims.

Claims

1. A customer service system service method, comprising:

when the robot customer service detects that the robot customer service is not in a conversation suspension state, the robot customer service autonomously processes the conversation message; in the autonomous processing process, when a preset conversation suspension condition is met, executing a process of switching the robot customer service to a conversation suspension state; the preset session suspension condition comprises the following steps: the session message cannot be processed autonomously or identified;

when the robot customer service and the session service of the customer client are in a session state and receives a session message sent to the customer client from the customer service client, the session state of the robot in the session service of the customer client is switched to a session suspension state, the transfer operation of the session message between the customer service client and the customer client is executed, and the received session message is directly transmitted to the customer client after the session message of the customer service client is received.

2. A customer service system service method according to claim 1, wherein said autonomously processing by said robot customer service said session message when said robot customer service detects itself not in a session suspended state comprises:

identifying content of the session message;

3. The customer service system service method according to claim 2, wherein the performing of the process of switching the robot customer service to the session suspended state when a preset session suspended condition is satisfied in the autonomous process includes:

4. The customer service system service method of claim 3, further comprising:

5. The customer service system service method according to claim 1, wherein when the robot customer service detects that it is in a session suspended state, further comprising:

6. The customer service system service method of claim 1, further comprising:

7. The customer service system service method of claim 6, wherein said determining that the session between the customer service client and the customer client is over comprises at least one of:

and detecting a session completion signal of the customer service client.

8. The customer service system service method according to claim 7, wherein the customer service client sends the session completion signal when detecting that the session window is closed or when detecting that the session message sent by the customer service client contains a content indicating that the session is completed.

9. The customer service system service method according to claim 3, wherein said determining whether the conversation message can be handled autonomously according to the content of the conversation message comprises:

determining whether the robot customer service can autonomously answer the contents of the session message according to language experience data;

10. A computer-readable storage medium, on which a program is stored, characterized in that the program, when executed, carries out the steps of the customer service system service method according to any one of claims 1 to 9.