CN105553833A

CN105553833A - Customer service system and service method and robot customer service thereof

Info

Publication number: CN105553833A
Application number: CN201511025278.5A
Authority: CN
Inventors: 李思远; 朱频频
Original assignee: Shanghai Zhizhen Intelligent Network Technology Co Ltd
Current assignee: Guizhou little love robot technology Co., Ltd.; Shanghai Zhizhen Intelligent Network Technology Co Ltd
Priority date: 2015-12-30
Filing date: 2015-12-30
Publication date: 2016-05-04
Anticipated expiration: 2035-12-30
Also published as: CN105553833B; CN109150701B; CN109245995A; CN109150701A; CN109245995B

Abstract

The invention discloses a customer service system and a service method and a robot customer service thereof; the service method of the customer service system comprises the following steps: when the robot customer service receives a session message from a customer client-side, detecting whether the robot customer service is in a session answering state; when the robot customer service detects that the robot customer service is in the session answering state, performing transmission operation of the session message between the customer client-side and a customer service client-side; when the robot customer service detects that the robot customer service is not in the session answering state, handling the session message by the robot customer service; and in independent handling process, when the preset session answering conditions are met, performing the process of switching the robot customer service to the session answering state. According to above scheme, the cooperative work between a manual customer service and the robot customer service can be realized, the effectiveness of the service of the customer service system can be strengthened, and the user experience of a customer is improved.

Description

Customer service system and method for servicing thereof, robot customer service

Technical field

The present invention relates to instant messaging field, particularly relate to a kind of customer service system and method for servicing thereof, robot customer service.

Background technology

Along with the develop rapidly of the Internet, the raising of people's service awareness, network customer service spreads to all trades and professions, is deep into the links of everyday commerce service.

Current customer service system, usually be made up of machine customer service and artificial customer service, wherein machine customer service is generally the immediate communication tool based on webpage, and artificial customer service is the immediate communication tool be normally embedded in webpage, and each artificial customer service client all has manual maintenance.In customer service processes, when receiving the conversation message from client, first served by machine customer service.When client thinks that machine customer service cannot solve the problem of its proposition, more manually click artificial customer service, seek advice to artificial customer service.

Above-mentioned customer service system, can the demand of quick customer in response to a certain extent.But; in actual application; the content of the conversation message of machine customer service identification client input; the data with the content matching of described conversation message are searched from stored resource database; and send to subscription client to reply by mating the data obtained, for same problem, usually mechanically can carry out repeatability and answer; easily enrage client, affect customer experience.Artificial customer service then cannot Always Online", and usually wants simultaneously in the face of multiple client, thus there is situation about can not reply in time.Thus current customer service system is difficult to the demand for services effectively meeting client, affects the Consumer's Experience of client.

Summary of the invention

The technical problem that the present invention solves is the validity how strengthening customer service system service, improves the Consumer's Experience of client.

For solving the problems of the technologies described above, the embodiment of the present invention provides a kind of customer service system method of servicing, and described method comprises:

When robot customer service receives the conversation message from consumer user end, detect self and whether be in session suspended state;

When the customer service of described robot detect self be in session suspended state time, perform conversation message between described consumer user end and customer service client transmission operation;

When the customer service of described robot detect self be not in session suspended state time, independently process described conversation message by the customer service of described robot; In autonomous processing procedure, when meeting the session suspension condition preset, perform the process customer service of described robot being switched to session suspended state.

Optionally, described when the customer service of described robot detect self be not in session suspended state time, independently process described conversation message by the customer service of described robot, comprising:

Identify the content of described conversation message;

According to the content of described conversation message, determine whether self can independently process described conversation message;

When determining that self independently can process described conversation message, from stored Service Source, extract the response message of mating with the content of described conversation message, and send to described consumer user end.

Optionally, described in autonomous processing procedure, when meeting the session suspension condition preset, performing the process customer service of described robot being switched to session suspended state, comprising:

When determining self cannot independently to process or described in None-identified during conversation message, certainly session suspended state will be set in the state in the subsequent session business of described consumer user end, extract with the pacifying property response message of the content matching of described conversation message and be sent to described consumer user end, and self cannot independently to process described or the conversation message of None-identified is sent to customer service client.

Optionally, described method also comprises: the customer service of described robot by described cannot independently process or None-identified conversation message corresponding to pacifying property response message cannot independently to process with described or None-identified conversation message together be sent to described customer service client.

Optionally, after the customer service of described robot detects and self is in session suspended state, also comprise: when the customer service of described robot does not receive the conversation message from described customer service client, the content of conversation message described in the customer service identification of described robot, extracts with the pacifying property response message of the content matching of described conversation message and is sent to described consumer user end.

Optionally, described method also comprises: in the session service of the customer service of described robot and described consumer user end, be in session status, and when receiving the conversation message sent to described consumer user end from described customer service client, certainly will switch to session suspended state in the session status in the session service of described consumer user end, and the transmission performing conversation message between described customer service client and described consumer user end operates.

Optionally, described method also comprises: when determining the conversation end between described customer service client and described consumer user end, and the state recovered in the session service between described consumer user end is session status.

Optionally, the described conversation end determined between described customer service client and described consumer user end, comprises following at least one:

Determine that between described customer service client and described consumer user end, sessionless transmission of messages reaches preset duration;

The session settling signal of described customer service client detected.

Optionally, described customer service client sends described session settling signal when detecting that session window is closed or when detecting in the conversation message that described customer service client sends and comprise content that display session completes.

Optionally, pacifying property response message described in is suitable for characterizing customer service and is in the state finding problem answers.

Optionally, described Service Source comprises: language empirical data and customer service resource data.

Optionally, to be undertaken from the conversation message between consumer user end classifying the described customer service resource data splitting and obtain different scene type according to scene.

Optionally, described method also comprises: carry out the training of large data machine, to identify the scene type belonging to conversation message.

Optionally, the described content according to described conversation message, determines whether can independently process described conversation message, comprising:

Determine whether the customer service of described robot can independently answer according to described language empirical data the content of described conversation message;

When determining that the customer service of described robot cannot be answered according to described language empirical data, determine the scene type belonging to content of described conversation message;

When determining that the customer service of described robot can independently be answered according to described language empirical data, or when there is the scene type belonging to content of described conversation message, determine that the customer service of described robot can independently process described conversation message.

Optionally, described when determining independently to process described conversation message, from stored Service Source, extract the response message of mating with the content of described conversation message, comprising:

When determining that the customer service of described robot can independently be answered according to described language empirical data, from stored language empirical data, extract the response message of mating with the content of described conversation message;

When determining independently to answer according to described language empirical data, the response message of mating with the content of described conversation message under scene belonging to extracting from stored customer service resource data.

Optionally, the form of described conversation message comprises following at least one: the statement that natural language word forms, data link, voice, image and video.

Optionally, described image comprise following one of them plant: static state or the dynamic picture with semantic information, or for client's session for product or the photo of service object or screenshotss information.

For solving the problem, the embodiment of the present invention additionally provides the customer service of a kind of robot, and the customer service of described robot comprises:

First information interaction unit, is suitable for receiving the conversation message from consumer user end, and returns the response message of mating with the content of described conversation message;

State detection unit, is suitable for when described first information interaction unit receives the conversation message from consumer user end, detect the customer service of described robot with the state in the session service of described consumer user end;

From Main Processor Unit, be suitable for when described state detection unit detects that the customer service of described robot is not in session suspended state, the described conversation message of autonomous process, and obtained response message is back to described consumer user end by described first information interaction unit;

Status unit, is suitable for described in the autonomous processing procedure of Main Processor Unit, and when meeting the suspension condition preset, controlling the customer service of described robot is being session suspended state with the state in the subsequent session business of described consumer user end;

Second information interaction unit, is suitable for, when described state detection unit detects that the customer service of described robot is in session suspended state, the conversation message that described first information interaction unit receives being transferred to customer service client; And the conversation message from described customer service client is sent to corresponding consumer user end by described first information interaction unit.

Optionally, describedly to comprise from Main Processor Unit:

Recognition unit, is suitable for, when described state detection unit detects that the customer service of described robot is not in session suspended state, identifying the content of the conversation message received by described first information interaction unit;

Ability judging unit, is suitable for the content of the described conversation message obtained according to described recognition unit identification, determines whether the customer service of described robot can independently process described conversation message;

Memory cell, is suitable for stores service resource;

First response message generation unit, be suitable for when described ability judging unit determines that the customer service of described robot independently can process described conversation message, from the Service Source that described storing sub-units stores, extract the response message of mating with the content of described conversation message.

Optionally, described status unit, be suitable for when described ability judging unit determines that the customer service of described robot cannot independently process or described in None-identified during conversation message, controlling the customer service of described robot is being session suspended state with the state in the subsequent session business of described consumer user end;

Describedly also to comprise from Main Processor Unit: the second response message generation unit, being suitable for controlling the customer service of described robot when being session suspended state with the state in the subsequent session business of described consumer user end at described status unit, from the Service Source of described cell stores, extracting the pacifying property response message with the content matching of described conversation message;

Described first information interaction unit, is also suitable for the pacifying property response message of the content matching of described and described conversation message to be back to described consumer user end.

Optionally, described second information interaction unit be also suitable for by described cannot independently process or None-identified conversation message corresponding to pacifying property response message cannot independently to process with described or None-identified conversation message together be sent to described customer service client.

Optionally, described from Main Processor Unit, also be suitable for detecting that the customer service of described robot is being session suspended state with the state in the session service of described consumer user end when described state detection unit, and described second information interaction unit is not when receiving the conversation message from described customer service client, identify the content of described conversation message, extract with the pacifying property response message of the content matching of described conversation message and be sent to described consumer user end.

Optionally, described status unit, also be suitable for being in session status in the session service of the customer service of described robot and described consumer user end, and described second information interaction unit is when receiving the conversation message from described customer service client, the customer service of described robot is switched to session suspended state with the state in the subsequent session business of described consumer user end, and controls described second information interaction unit execution message exchange operations.

Optionally, the customer service of described robot also comprises: conversation end detecting unit, is suitable for judging whether the session between described customer service client and described consumer user end terminates;

Described status unit, is also suitable for when determining the conversation end between described customer service client and described consumer user end, and recovering the state of described robot customer service between described consumer user end in subsequent session business is session status.

Optionally, described conversation end detecting unit comprises following at least one:

First detection sub-unit, whether reach preset duration, and when reaching preset duration, determine the conversation end between described customer service client and described consumer user end if being suitable for detecting the described second information interaction unit free time;

Second detection sub-unit, is suitable for detecting the session settling signal from described customer service client, and when the session settling signal from described customer service client being detected, determines the conversation end between described customer service client and described consumer user end.

Optionally, the Service Source that described memory cell stores comprises: language empirical data and customer service resource data.

Optionally, the customer service of described robot also comprises: machine training unit, be suitable for carrying out the training of large data machine to described robot, to identify the scene type belonging to conversation message, described scene type is according to scene, the conversation message between consumer user end is carried out classification fractionation to obtain.

Optionally, described ability judging unit, comprising:

First judgment sub-unit, is suitable for determining whether the customer service of described robot can independently answer according to described language empirical data the content of described conversation message;

Second judgment sub-unit, is suitable for, when described first judgment sub-unit determines that the customer service of described robot cannot independently be answered according to described language empirical data, determining the scene type belonging to content of described conversation message.

Optionally, described first response message generation unit, comprising:

First generates subelement, is suitable for when determining that the customer service of described robot can independently be answered according to described language empirical data, extracts the response message of mating with the content of described conversation message from the language empirical data that described memory cell stores;

Second generates subelement, is suitable for when determining that the customer service of described robot cannot be answered according to described language empirical data, the response message of mating with the content of described conversation message under scene belonging to extracting from the customer service resource data that described memory cell stores.

For solving the problem, the embodiment of the present invention additionally provides a kind of customer service system, described customer service system comprises: consumer user end, be suitable for the customer service client that conversates with described consumer user end and any one robot customer service above-mentioned, and described customer service client and described consumer user end are by the customer service session of described robot.

Optionally, consumer user end and customer service client that the customer service of described robot is suitable for when being in session status are multiple, and determine self cannot independently to process or described in None-identified during conversation message, according to the task amount of customer service client being in session status, the pacifying property response message of the described conversation message from consumer user end and correspondence is sent to the minimum customer service client of task amount.

Optionally, consumer user end and customer service client that the customer service of described robot is suitable for when being in session status are multiple, and determine self cannot independently to process or described in None-identified during conversation message, according to conversation message record, self cannot independently to process described or the customer service client of described consumer user end conversation message is crossed in pre-treatment that the conversation message of None-identified is sent to.

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

When robot customer service receives the conversation message from consumer user end, session suspended state whether is in by detecting self, and when determining self to be in session suspended state, perform the transmission operation of conversation message between described consumer user end and customer service client, namely be directly customer service by the artificial customer service being positioned at customer service client, and when determining self not to be in session suspended state, by robot customer service self the independently described conversation message of process, and in autonomous processing procedure, when meeting the session suspension condition preset, perform the process customer service of described robot being switched to session suspended state, namely primarily of robot customer service independently for being positioned at the customer service of described consumer user end, and when meeting certain condition, automatically switch to by manually for customer service, thus realize artificial customer service and robot customer service collaborative work.On the one hand, when robot customer service is in session suspended state, the transmission of the message that only needs to conversate and need not carry out any autonomous process, can save the Service Source of robot customer service; On the other hand, robot customer service only could utonomous working when self not being in session suspended state, the priority of old friend's work customer service is higher, robot customer service midway can be avoided to seize session, make client can obtain consistent formula service experience, thus the validity of customer service system service can be improved, improve the Consumer's Experience of client.

Further, robot customer service is not at the conversation message received from consumer user end and when self being in session suspended state, by identifying the content of described conversation message and determining whether self can independently process described conversation message.When determining that self independently can process described conversation message, from stored Service Source, extracting the response message with the content matching of described conversation message, and send to described consumer user end, client can be replied timely and effectively.

Further, when determining self cannot independently to process or described in None-identified during conversation message, on the one hand, by from being set to session suspended state in the state in the subsequent session business of described consumer user end, process resource can being saved; On the other hand, by sending the pacifying property response message with the content matching of described conversation message in time, client can be avoided to can not get reasonably response and to produce impatience, and self cannot independently to process described or the conversation message of None-identified is sent to customer service client, need not client manually select, when the customer service of described robot cannot effectively process, artificial customer service can be realized and get involved in time and effectively solve the problem of described client.In a word, such scheme by robot customer service automatically with the collaborative work of artificial customer service, the problem of client can be solved timely and effectively, and client is the difference experienced less than robot customer service and artificial customer service, in the process of wait acknowledge, mood can effectively be pacified, and obtains better Interactive Experience.

Further, by by described cannot independently process or None-identified conversation message corresponding to pacifying property response message cannot independently to process with described or None-identified conversation message together be sent to described customer service client, make artificial customer service can see the content of robot customer service to the response message of customer replies, be convenient to continuity and the consistency of hold acknowledge, thus consistent formula service experience can be optimized, improve the Consumer's Experience of client.

Further, when robot customer service is in session suspended state, if do not receive the conversation message from customer service client, namely when artificial customer service does not also get involved process, the content of conversation message is identified, when the content of conversation message can be identified, pacifying property response message is accordingly extracted from stored Service Source, and in artificial customer service and client's reciprocal process, the customer service of described robot does not do any identification to conversation message, autonomous process or response, but only carry out the transmission of customer service client and consumer user end both sides session, the collaborative e-commerce of robot customer service and artificial customer service can be improved further, impatience in further alleviation client waiting process.

Further, when self being in session status with robot customer service described in the session service of described consumer user end, and when receiving the conversation message from described customer service client, robot customer service is being switched to session suspended state with the session status in the session service of described consumer user end, and the transmission performing conversation message between described customer service client and described consumer user end operates, in artificial and robot customer service cooperative working process, the priority of artificial customer service is higher, artificial customer service as required, can Active participation at any time, therefore better service can be provided for client.

Further, when the conversation end between described customer service client and described consumer user end is determined in robot customer service, the state recovered in the subsequent session business between described consumer user end is session status, namely when artificial customer service has got involved, the state of the robot customer service that automatically switches back, realize the seamless connection of artificial customer service and robot customer service, thus the service experience of client can be optimized further.

Further, by determining that between described customer service client and described consumer user end, sessionless transmission of messages reaches preset duration, or when detecting that session window is closed or detect in the conversation message that described customer service client sends when comprising content that display session completes, the session settling signal that described customer service client sends automatically, artificial customer service can be automatically identified got involved, thus the operation of artificial customer service can be reduced, reduce the workload of artificial customer service, for artificial customer service is provided convenience.

Further, by carrying out large data training to the customer service of described robot, make the customer service of described robot can identify scene type belonging to conversation message, thus can determine whether rapidly can process described conversation message and carry out subsequent treatment, customer service response speed can be improved, strengthen the Interactive Experience of client further.

Further, when the customer service of described robot is determined independently to answer according to language empirical data, the response message of mating with the content of described conversation message is extracted from stored language empirical data, and when determining independently to answer according to language empirical data, the response message of mating with the content of described conversation message under choosing affiliated scene, thus not only can answer according to language empirical data, and can answer according to customer service resource data, and due to coupling resource data corresponding with the scene of customer service, thus can expand the range of the response message that robot customer service can be answered and improve the accuracy of response.

Further, by characterizing to consumer feedback the pacifying property response message that customer service is in the state finding problem answers, to find for the content of conversation message with customer service client and the process providing solution performs simultaneously, the virtual condition of customer service can be reflected and need not customer service manually input on the one hand, both saved the response operation time of artificial customer service, and can play to client's mood the effect of pacifying again; Also make artificial customer service the more time can be put in the middle of the problem of solution client on the other hand, thus can more promptly provide the response more making Customer Satisfaction.

Accompanying drawing explanation

Fig. 1 is a kind of customer service system structural representation in the embodiment of the present invention;

Fig. 2 is a kind of customer service system method of servicing flow chart in the embodiment of the present invention;

Fig. 3 is another kind of customer service system method of servicing flow chart in the embodiment of the present invention;

Fig. 4 is a kind of consumer user end interactive interface schematic diagram in the embodiment of the present invention;

Fig. 5 is session status handoff procedure schematic diagram in a kind of customer service process in the embodiment of the present invention;

Fig. 6 is the structural representation of a kind of robot customer service in the embodiment of the present invention.

Embodiment

As previously mentioned, current customer service system, in customer service processes, machine customer service and artificial customer service are customer service independently of one another, are namely first served by machine customer service.When client thinks that machine customer service cannot solve the problem of its proposition, more manually click artificial customer service, seek advice to artificial customer service.Such scheme exists robot customer service mechanicalness to be repeated to answer and easily enrages client, and artificial customer service cannot Always Online" and client is numerous usually, and be difficult to timely reply, therefore current customer service system is difficult to the demand for services effectively meeting client.

For the problems referred to above, the embodiment of the present invention provides a kind of customer service system, with reference to the customer service system structural representation shown in Fig. 1, comprise: consumer user end 11, be suitable for the customer service client 12 that conversates with described consumer user end 11 and robot customer service 13, described customer service client and described consumer user end are conversated by the customer service of described robot.

In the customer service system of the embodiment of the present invention, by robot customer service and artificial customer service collaborative work, by the unified conversation message received from consumer user end of robot customer service, session suspended state whether is in by detecting self, and when determining self to be in session suspended state, perform the transmission operation of conversation message between described consumer user end and customer service client, namely be directly customer service by the artificial customer service being positioned at customer service client, and when determining self not to be in session suspended state, by robot customer service self the independently described conversation message of process, and in autonomous processing procedure, when meeting the session suspension condition preset, perform the process customer service of described robot being switched to session suspended state, namely primarily of robot customer service independently for being positioned at the customer service of described consumer user end, and when meeting certain condition, automatically switch to by manually for customer service, thus realize artificial customer service and robot customer service collaborative work.

For enabling above-mentioned purpose of the present invention, characteristic sum beneficial effect more becomes apparent, and is described in detail specific embodiments of the invention below in conjunction with accompanying drawing.

With reference to Fig. 1 and Fig. 2, in the embodiment of the present invention, a kind of customer service system method of servicing, is described in detail below by way of concrete steps.

S21, robot customer service receives the conversation message from consumer user end.

In concrete enforcement, the form of described conversation message can have multiple.Such as, can be wherein any one form such as the statement of natural language word composition, data link, voice, image or video, or the combination of various ways.

S22, whether what robot customer service detected self is in session suspended state, when detect self be in session suspended state time, perform step S23; When detect self be not in session suspended state time, perform step S24.

In concrete enforcement, in order to save the process resource of robot customer service, realize the collaborative work of robot customer service and artificial customer service, robot customer service can be set to " session status " state independently processed from the conversation message of consumer user end, or " non-suspended state ", and robot customer service cannot independently be processed or the content of None-identified conversation message time state be set to " session suspended state ".

S23, performs the transmission operation of conversation message between described consumer user end and customer service client.

S24, the customer service of described robot independently processes described conversation message.

In an embodiment of the present invention, the process that the customer service of described robot independently processes described conversation message is as follows: first, identify the content of described conversation message, afterwards, according to the content of described conversation message, determine whether self can independently process described conversation message, and when determining that self independently can process described conversation message, from stored Service Source, extract the response message of mating with the content of described conversation message, and send to described consumer user end.

S25, detects and whether meets the session suspension condition preset, and if so, then performs step S26; If not, then step S24 is performed.

S26, performs the process customer service of described robot being switched to session suspended state.

In the autonomous processing procedure of described robot customer service, when meeting the session suspension condition preset, the customer service of described robot can be switched to session suspended state.Such as, self independently cannot process when detecting or described in None-identified during conversation message, following operation can be performed:

Certainly session suspended state will be set in the state in the subsequent session business of described consumer user end, extract with the pacifying property response message of the content matching of described conversation message and be sent to described consumer user end, and self cannot independently to process described or the conversation message of None-identified is sent to customer service client.

Wherein, described in pacifying property response message, refer to the message that the uneasiness of client and negative feeling can be made to settle out, in concrete enforcement, can be presented by modes such as nature word language, picture, voice, music.

In concrete enforcement, can by described cannot independently process or None-identified conversation message corresponding to pacifying property response message cannot independently to process with described or None-identified conversation message together be sent to described customer service client, make artificial customer service can see the content of robot customer service to the response message of customer replies, be convenient to continuity and the consistency of hold acknowledge, thus consistent formula service experience can be optimized, improve the Consumer's Experience of client.

Adopt above-described embodiment, on the one hand, when robot customer service is in session suspended state, the transmission of the message that only needs to conversate and need not carry out any autonomous process, can save the Service Source of robot customer service; On the other hand, robot customer service only could utonomous working when self not being in session suspended state, the priority of old friend's work customer service is higher, robot customer service midway can be avoided to seize session, make client can obtain consistent formula service experience, thus the validity of customer service system service can be improved, improve the Consumer's Experience of client.

Client service method below by way of customer service system another kind of in the embodiment of the present invention illustrates the principle of robot customer service and artificial customer service collaborative work in the embodiment of the present invention, and with reference to Fig. 3, concrete steps are as follows:

S31, receives the conversation message from consumer user end.

In concrete enforcement, as previously mentioned, the form of described conversation message can have multiple.Such as, can be wherein any one form such as the statement of natural language word composition, data link, voice, image or video, or the combination of various ways.

In an embodiment of the present invention, the form of described conversation message comprises the statement of natural language word composition.The statement that client can consist of described natural language is described description object.Such as, if description object is product, described conversation message can be then the advisory message of the product related information such as size, shape, color, quality, date of manufacture, shelf-life to product, also can be the advisory message to after services of product.If description object is service, described conversation message can be then service content, service duration, points for attention, provides the attendant of service to wait the consultation information of service related content.

In concrete enforcement, above-mentioned conversation message, except presenting with written form, can also present with the form of voice, image or video.Such as, client is conversated by the form of audio frequency or instant audio session or video conversation and customer service.

In concrete enforcement, client can also be conversated with the form of picture and customer service by consumer user end.Described picture can be have the static images of semantic information, dynamic picture or its combine, or client's session for product or the photo of service object or screenshotss information.

In concrete enforcement, the link comprising product or service object can also be sent by the form of conversation message by client.

Be understandable that, in concrete enforcement, described conversation message can also be the statement of natural language word composition, two or more the form such as data link, voice, image or video combines and realize.Such as, client by voice or video and customer service session, while corresponding data link is sent to customer service by consumer user end.

S32, determines from the state in session service, if session suspended state, then performs step S37; As non-session suspended state, then perform step S33.

If robot customer service is determined from being non-session suspended state with the state in the session service of described consumer user end, then to perform step S33.If robot customer service is determined self to be that the state in the session service of described consumer user end is session suspended state, then will no longer independently reply from the follow-up Customer Problems of described consumer user end under current sessions, but be sent to customer service client and transfer to artificial customer service process, namely perform step S37.

Following steps S33 ~ S36 is the autonomous processing procedure of described robot customer service, and step S37 is the process triggering manpower intervention.

S33, identifies the content of described conversation message, if can identify, then performs step S34; Otherwise perform step S36.

In concrete enforcement, can, by the content of conversation message described in robot customer service identification, different recognition methodss can be adopted to carry out identifying processing according to different conversation message forms.Such as, can participle and semantics recognition be passed through, identify the statement of natural language word composition.And for example, by transferring the voice received to text, then participle and semantics recognition can be carried out further.If comprise image in described conversation message, then can to the word in image or pattern be capable extracts and do the operations such as further identification.If can identify, then perform step S34.

It should be noted that, some conversation messages may None-identified, such as video, or the form of message that the amount of information such as image is very large or complicated, may exceed the recognition capability of robot customer service, at this moment, can perform step S36.

S34, determines whether independently to process, and if so, then performs step S35; If not, then step S36 is performed.

Robot customer service can, according to the content of the conversation message identified, determine whether independently to process.

In concrete enforcement, robot can be confirmed whether according to stored Service Source independently to process.The Service Source stored can comprise stored language empirical data and customer service resource data.

Specifically, first can determine whether the content of described conversation message can independently answer according to described language empirical data, when determining independently to answer according to described language empirical data, determines whether independently to process; When determining independently to answer according to described language empirical data, then can determine the scene type belonging to content of described conversation message further, when determining the scene type existed belonging to the content of described conversation message, determine that the customer service of described robot can independently process described conversation message.

Wherein, for improving the ability of robot autonomous process, the training of large data machine can be carried out, to identify the scene type belonging to conversation message to robot customer service.In concrete enforcement, can be undertaken from the conversation message between consumer user end classifying the described customer service resource data splitting and obtain different scene type according to scene in advance.

In an embodiment of the present invention, in scene category classification and machine learning process, described conversation message not only comprises conversation message between robot customer service and consumer user end, also comprise the conversation message of artificial customer service and consumer user end, all conversation messages are carried out classification according to scene split, here artificial customer service is not taken as independently individual, but as the element of certain scene concrete in whole service link, thus by whole customer service mode of operation reconstruct, when a large amount of client, effectively can improve the operating efficiency of artificial customer service, optimize customer service to experience.Wherein, the element of scene can comprise client and customer service and the two the conversation message content for different service scenarios.

S35, extracts mated response message and is sent to described consumer user end from stored Service Source.

As previously mentioned, Service Source can comprise stored Service Source and can comprise stored language empirical data and customer service resource data.Correspondingly, in concrete enforcement, when determining independently to answer according to described language empirical data, from stored language empirical data, the response message of mating with the content of described conversation message is extracted; When determining independently to answer according to described language empirical data, the response message of mating with the content of described conversation message under scene belonging to extracting from stored customer service resource data.

Such as, when the content identifying described conversation message is " hello ", by extracting the response message obtaining mating in stored language empirical data be: " you are good,.Have what can for you service? "And for example, when the content of the described conversation message identified is: " one-piece dress that may I ask that part blueness that I buys is delivered? " time, then according to language empirical data and customer service resource data, can therefrom extract the sequence information that described client is corresponding, obtaining response message is: " you are good! The one-piece dress that you buy is shipped ".The logistics information of described order also can be enclosed in described response message.

S36, by from the subsequent session business of described consumer user end in be set to suspended state, and reply the pacifying property response message corresponding to described conversation message to described consumer user end, and by described conversation message and described in pacifying property response message be sent to customer service client.

Can complete by the customer service of step S36 robot the process switching to session suspended state, realize the collaborative work of robot customer service and artificial customer service.Be described below by way of an embody rule scene:

Robot cannot independently process or None-identified conversation message time, machine customer service extract with the pacifying property response message of described conversation message content matching.In concrete enforcement, the conversation message in order to avoid client can not get response for a long time and produces mood restless, described in pacifying property response message be suitable for characterizing customer service and be in the state finding problem answers.Such as, when the content identifying conversation message is: " I want buy two Men's Leather Shoes ", when robot customer service is determined independently to process, on the one hand by from the subsequent session business of described consumer user end in be set to session suspended state; On the other hand, coupling obtains pacifying property response message and is: ", good.Waiting, I removes to look at your shoes applicable.", described pacifying property response message is sent to consumer user end, makes client obtain replying in time and patient wait, and " I wants to buy two Men's Leather Shoes " this conversation message is sent to customer service customer side, process to make artificial customer service.The pacifying property response message that some are corresponding when robot customer service None-identified goes out the content of concrete conversation message also can be set in advance, such as the picture received, during None-identified, corresponding pacifying property response message can be set to " just a moment,please, and I first has a look ".

S37, transfers to corresponding customer service client by described conversation message.

When robot customer service is in session suspended state, only need to conversate message transmission and any autonomous process need not be carried out, conversation message from client's client and customer service client is transmitted to the other side respectively, process is got involved by artificial customer service, robot customer service only conversates the forwarding of message, and do not do any identification, process or response, thus the Service Source of robot customer service can be saved.

From above-described embodiment, when receiving the conversation message from consumer user end, first robot customer service self detects self and whether is in session suspended state, and when determining self not to be in session suspended state, the operation of the content identifying described conversation message is just performed by the customer service of described robot, and when determining self to be in session suspended state, then conversate alternately by artificial customer service by customer service client and client, thus realize artificial and robot customer service collaborative work better, on the one hand, when robot customer service is in session suspended state, only need to conversate message transmission and any autonomous process need not be carried out, the Service Source of robot customer service can be saved, on the other hand, robot customer service only could work when self not being in session suspended state, the priority of old friend's work customer service is higher, robot customer service midway can be avoided to seize session, make client can obtain consistent formula service experience, improve the Consumer's Experience of client.

In concrete enforcement, in order to alleviate the impatience in client's waiting process further, when robot customer service is in session suspended state, if do not receive the conversation message from customer service client, namely when artificial customer service does not also get involved process, can identify the content of conversation message, when the content of conversation message can be identified, from stored Service Source, extract pacifying property response message accordingly; And receiving the conversation message from customer service client, namely after manpower intervention process, in artificial customer service and client's reciprocal process, the customer service of described robot can not do any identification, autonomous process or response to conversation message, but only carries out the transmission of customer service client and consumer user end both sides session.

In concrete enforcement, when determining the conversation end between described customer service client and described consumer user end, the state that robot customer service can independently recover in the session service between described consumer user end is session status, i.e. non-session suspended state.

When artificial customer service has got involved, by the state of the robot customer service that automatically switches back, realize the seamless connection of artificial customer service and robot customer service, thus the service experience of client can have been optimized further.In concrete enforcement, the conversation end that various ways is determined between described customer service client and described consumer user end can be had.Such as, determining between described customer service client and described consumer user end, whether sessionless transmission of messages reaches preset duration, when reaching preset duration, then determining the conversation end between described customer service client and described consumer user end.And for example, the session settling signal from described customer service client whether being detected, when the session settling signal from described customer service client being detected, then determining the conversation end between described customer service client and described consumer user end.In an embodiment of the present invention, described customer service client can detect that described customer service client sends described session settling signal when session window is closed.In an alternative embodiment of the invention, when detect in the conversation message that described customer service client sends comprise content that display session completes time described consumer user end also can send described session settling signal.

Be understandable that, in concrete enforcement, as required, the method in above-mentioned different embodiment can be combined.

By determining that between described customer service client and described consumer user end, sessionless transmission of messages reaches preset duration, or when detecting that session window is closed or detect in the conversation message that described customer service client sends when comprising content that display session completes, the session settling signal that described customer service client sends automatically, artificial customer service can be automatically identified got involved, thus the operation of artificial customer service can be reduced, reduce the workload of artificial customer service, for artificial customer service is provided convenience.

In concrete enforcement, when with the session service of described consumer user end in self be in session status, and when receiving the conversation message from described customer service client, certainly session suspended state will be switched in the session status in the session service of described consumer user end, and the transmission performing conversation message between described customer service client and described consumer user end operates, namely in artificial and robot customer service cooperative working process, the priority of artificial customer service is higher, artificial customer service as required, can Active participation at any time, thus better service can be provided for client.

Be understandable that, introduce in above-described embodiment for receiving a handling process during conversation message from consumer user end, all can process according to above-mentioned flow process for each conversation message received.

For making those skilled in the art understand better and realize the present invention, be described in detail below by way of a concrete session context.

With reference to the consumer user end interactive interface schematic diagram shown in Fig. 4.Described interactive interface is the session interaction interface of client A and customer service B, is shown by the consumer user end of client A.Wherein, in customer service system, no matter robot customer service or artificial customer service, show client A as a unified vivid customer service B, active service person may be robot customer service, or the artificial customer service of customer service client below certain.Described square frame A and B represents the identification information of client A and customer service B respectively, in concrete enforcement, can be shown in described square frame A and B by described client A and the information such as the self-defining head portrait of customer service B and the pet name.Temporal information " 5:44 in afternoon " in dotted line frame 48,49, " 5:50 in afternoon " represent the moment that the session of the consumer user end interface display of client A occurs.The conversation message content that right side conversation message sends for client A, left side conversation message is the response message content of customer service B.Below in this session interaction process, the processing procedure performed by customer service system describes in detail:

Client A sends conversation message " hello " by consumer user end, when robot customer service confirms self not to be in session suspended state, described conversation message is identified, after identification draws, first be confirmed whether to answer according to language empirical data, when determining to answer according to language data, extract mated response message " you are good,.Have what can for you service? " and be sent to the consumer user end of client A.Client A then sends conversation message " I wants to buy two leather shoes " by consumer user end, robot customer service confirms self not to be in session suspended state and after identifying, confirmation cannot independently process, then extract mate with conversation message " I want buy pair leather shoes " pacifying property response message ", good.Waiting, I removes to look at your shoes applicable."; simultaneously; the customer service of described robot can will be session suspended state from the session state setting in the session service follow-up with the consumer user end of described client A, described conversation message " I wants to buy two leather shoes " is sent to customer service client simultaneously and is processed by artificial customer service.Client A replys conversation message " good, to thanks ".

It should be noted that, artificial customer service may not need the client of service too many in position or due to artificial customer service, process cannot be got involved in time for the moment, spend a period of time like this, client A again by consumer user end send conversation message " please say hello? " robot customer service receive described conversation message " please say hello? " after, detect and self be in session suspended state, then identify the content of described message, extract and reply pacifying property response message accordingly and " wait; I is also seeing ", and be sent to the consumer user end of described client A.

Afterwards, artificial customer service gets involved, and " you are good, and I thinks that these footwear are well suited for you below to send conversation message by customer service client.Http//www.tmall.com/ "; the message that described artificial customer service sends is not done any process by robot customer service; be directly passed to the consumer user end of described client A; after client A sees described conversation message, reply conversation message ", good, I considers to look at "; manually customer service is after customer service client receives this conversation message, successively reply response conversation message ", good.You are good can contact me at any time ", and " thanking to your presence, 88 ".

According to above-mentioned handling process, conversation message 41,42 is robot autonomous process and the message of response, the pacifying property response message that conversation message 43 and 44 generates for robot customer service, and conversation message 45,46 and 47 is the conversation message of artificial customer service response.

Be understandable that, in concrete enforcement, the response message that conversation message 46 and 47 also can generate for robot customer service automatically.Such as, if customer service response the customer service of conversation message robot receive conversation message ", good; I consider look at " after, in preset duration, do not receive the response message from customer service client, can confirm that manpower intervention terminates, then can identify the content of described conversation message.After the content identifying described conversation message, in an embodiment of the present invention, directly can to answer according to stored language empirical data ", good.You are good can contact me at any time " and " thanking to your presence, 88 ".In an alternative embodiment of the invention, after the content identifying described conversation message, determine independently to answer according to language empirical data, then mate the scene type that described conversation message is corresponding, and extract from stored Service Source coupling response message ", good.You are good can contact me at any time " and " thanking to your presence, 88 ", and send to the consumer user end of described client A successively.

Understand better for making those skilled in the art and realize the present invention, referring to session status handoff procedure schematic diagram in customer service process a kind of in the embodiment of the present invention shown in Fig. 5, the stateful switchover process of robot customer service in session service in service process is described in detail.Such as, robot customer service initial condition is session status 51 (non-session suspended state), the conversation message received when robot customer service is unidentified, or robot customer service cannot independently process described conversation message, or when artificial customer service gets involved session, then switch to session suspended state 52.And when robot customer service is in session suspended state, if when the signal that artificial customer service got involved being detected, then the customer service of described robot can conversate recovery 53,53 are recovered by session, robot customer service can continue autonomous service, namely switching to session status 51, can be independently customer service.After robot customer service is in session suspended state or manpower intervention completes, if do not receive the new conversation message from consumer user end, then whole conversation procedure terminates, and is namely in conversation end state 54.

Understand better for making those skilled in the art and realize the present invention, the embodiment of the present invention additionally provides a kind of robot customer service that can realize above-mentioned client service method, with reference to Fig. 6, robot customer service 60 can comprise: the first information interaction unit 61, state detection unit 62, from Main Processor Unit 63, status unit 64 and the second information interaction unit 65, wherein:

First information interaction unit 61, is suitable for receiving the conversation message from consumer user end, and returns the response message of mating with the content of described conversation message.

State detection unit 62, is suitable for when described first information interaction unit 61 receives the conversation message from consumer user end, detect the customer service of described robot with the state in the session service of described consumer user end.

From Main Processor Unit 63, be suitable for when described state detection unit 62 detects that the customer service of described robot is not in session suspended state, the described conversation message of autonomous process, and obtained response message is back to described consumer user end by described first information interaction unit.

Status unit 64, be suitable for described in the autonomous processing procedure of Main Processor Unit 63, when meeting the suspension condition preset, controlling the customer service of described robot is being session suspended state with the state in the subsequent session business of described consumer user end.

Second information interaction unit 65, is suitable for, when described state detection unit 62 detects that the customer service of described robot is in session suspended state, the conversation message that described first information interaction unit 61 receives being transferred to customer service client; And the conversation message from described customer service client is sent to corresponding consumer user end by described first information interaction unit 61.

In concrete enforcement, the form of described conversation message can comprise following at least one: the statement that natural language word forms, data link, voice, image and video.

Wherein, described image comprise following one of them plant: static state or the dynamic picture with semantic information, or for client's session for product or the photo of service object or screenshotss information.

Adopt the customer service of above-mentioned robot, on the one hand, when robot customer service is in session suspended state, the transmission of the message that only needs to conversate and need not carry out any autonomous process, can save the Service Source of robot customer service; On the other hand, robot customer service only could utonomous working when self not being in session suspended state, the priority of old friend's work customer service is higher, robot customer service midway can be avoided to seize session, make client can obtain consistent formula service experience, realize robot customer service and artificial customer service collaborative work, thus the validity of customer service system service can be improved, improve the Consumer's Experience of client.

In concrete enforcement, as shown in Figure 6, describedly can to comprise from Main Processor Unit 63: recognition unit 631, ability judging unit 632, memory cell 633 and the first response message generation unit 632, wherein:

Recognition unit 631, is suitable for, when described state detection unit 62 detects that the customer service of described robot is not in session suspended state, identifying the content of the conversation message received by described first information interaction unit 61.

Ability judging unit 632, is suitable for the content identifying the described conversation message obtained according to described recognition unit 631, determines whether the customer service of described robot can independently process described conversation message.

Memory cell 633, is suitable for stores service resource.Wherein, Service Source can comprise: language empirical data and customer service resource data etc.

First response message generation unit 634, be suitable for when described ability judging unit 632 determines that the customer service of described robot independently can process described conversation message, from the Service Source that described storing sub-units 633 stores, extract the response message of mating with the content of described conversation message.

Adopt the customer service of above-mentioned robot, at the conversation message received from consumer user end and when self not being in session suspended state, by identifying the content of described conversation message and determining whether self can independently process described conversation message.When determining that self independently can process described conversation message, from stored Service Source, extracting the response message with the content matching of described conversation message, and send to described consumer user end, client can be replied timely and effectively.

In an embodiment of the present invention, the customer service of described robot also comprises machine training unit 66, be suitable for carrying out the training of large data machine to described robot, to identify the scene type belonging to conversation message, described scene type is according to scene, the conversation message between consumer user end is carried out classification fractionation to obtain.

By carrying out large data training to the customer service of described robot, make the customer service of described robot can identify scene type belonging to conversation message, thus can determine whether rapidly can process described conversation message and carry out subsequent treatment, customer service response speed can be improved, strengthen the Interactive Experience of client further.

In concrete enforcement, described ability judging unit 632 can comprise:

First judgment sub-unit (not shown), is suitable for determining whether the customer service of described robot can independently answer according to described language empirical data the content of described conversation message;

Second judgment sub-unit (not shown), is suitable for, when described first judgment sub-unit determines that the customer service of described robot cannot independently be answered according to described language empirical data, determining the scene type belonging to content of described conversation message.

In concrete enforcement, described first response message generation unit 634 can comprise:

First generates subelement (not shown), being suitable for when determining that the customer service of described robot can independently be answered according to described language empirical data, from the language empirical data that described memory cell 633 stores, extracting the response message of mating with the content of described conversation message;

Second generates subelement (not shown), be suitable for when determining that the customer service of described robot cannot be answered according to described language empirical data, the response message of mating with the content of described conversation message under scene belonging to extracting from the customer service resource data that described memory cell 633 stores.

Adopt the robot customer service in above-described embodiment, when the customer service of described robot is determined independently to answer according to language empirical data, the response message of mating with the content of described conversation message is extracted from stored language empirical data, and when determining independently to answer according to language empirical data, the response message of mating with the content of described conversation message under choosing affiliated scene, not only can answer according to language empirical data, and can answer according to customer service resource data, and due to coupling resource data corresponding with the scene of customer service, thus can expand the range of the response message that robot customer service can be answered and improve the accuracy of response.

In concrete enforcement, as shown in Figure 6, described status unit 64, be suitable for when described ability judging unit determines that the customer service of described robot cannot independently process or described in None-identified during conversation message, controlling the customer service of described robot is being session suspended state with the state in the subsequent session business of described consumer user end.

Correspondingly, describedly also can to comprise from Main Processor Unit 63: the second response message generation unit 635, being suitable for controlling the customer service of described robot when being session suspended state with the state in the subsequent session business of described consumer user end at described status unit 64, from the Service Source that described memory cell 633 stores, extracting the pacifying property response message with the content matching of described conversation message.Described first information interaction unit 634, is also suitable for the pacifying property response message of the content matching of described and described conversation message to be back to described consumer user end.

Described pacifying property response message, refers to the message that the uneasiness of client and negative feeling can be made to settle out, and in concrete enforcement, can be presented by modes such as nature word language, picture, voice, music.In an embodiment of the present invention, pacifying property response message described in is suitable for characterizing customer service and is in the state finding problem answers.

By characterizing to consumer feedback the pacifying property response message that customer service is in the state finding problem answers, to find for the content of conversation message with customer service client and the process providing solution performs simultaneously, the virtual condition of customer service can be reflected and need not customer service manually input on the one hand, both saved the response operation time of artificial customer service, and can play to client's mood the effect of pacifying again; Also make artificial customer service the more time can be put in the middle of the problem of solution client on the other hand, thus can more promptly provide the response more making Customer Satisfaction.

Robot customer service in above-described embodiment, when determining self cannot independently to process or described in None-identified during conversation message, on the one hand, by from being set to session suspended state in the state in the subsequent session business of described consumer user end, process resource can be saved; On the other hand, by sending the pacifying property response message with the content matching of described conversation message in time, client can be avoided to can not get reasonably response and to produce impatience, and self cannot independently to process described or the conversation message of None-identified is sent to customer service client, need not client manually select, when the customer service of described robot cannot effectively process, artificial customer service can be realized and get involved in time and effectively solve the problem of described client.In a word, such scheme by robot customer service automatically with the collaborative work of artificial customer service, the problem of client can be solved timely and effectively, and client is the difference experienced less than robot customer service and artificial customer service, in the process of wait acknowledge, mood can effectively be pacified, and obtains better Interactive Experience.

In concrete enforcement, described second information interaction unit 65 be also suitable for by described cannot independently process or None-identified conversation message corresponding to pacifying property response message cannot independently to process with described or None-identified conversation message together be sent to described customer service client, make artificial customer service can see the content of robot customer service to the response message of customer replies, be convenient to continuity and the consistency of hold acknowledge, thus consistent formula service experience can be optimized, improve the Consumer's Experience of client.

Described from Main Processor Unit 63, also be suitable for detecting that the customer service of described robot is being session suspended state with the state in the session service of described consumer user end when described state detection unit 62, and described second information interaction unit 65 is not when receiving the conversation message from described customer service client, identify the content of described conversation message, extract with the pacifying property response message of the content matching of described conversation message and be sent to described consumer user end.

Adopt the customer service of above-mentioned robot, when robot customer service is in session suspended state, if do not receive the conversation message from customer service client, namely when artificial customer service does not also get involved process, the content of conversation message is identified, when the content of conversation message can be identified, pacifying property response message is accordingly extracted from stored Service Source, and in artificial customer service and client's reciprocal process, the customer service of described robot does not do any identification to conversation message, autonomous process or response, but only carry out the transmission of customer service client and consumer user end both sides session, the collaborative e-commerce of robot customer service and artificial customer service can be improved further, impatience in further alleviation client waiting process.

In concrete enforcement, described status unit 64, also be suitable for being in session status in the session service of the customer service of described robot and described consumer user end, and described second information interaction unit 65 is when receiving the conversation message from described customer service client, the customer service of described robot is being switched to session suspended state with the state in the subsequent session business of described consumer user end, and is controlling described second information interaction unit 65 and perform message exchange operations.In artificial and robot customer service cooperative working process, by the cooperation of status unit and the second information interaction unit in the customer service of described robot, make the priority of artificial customer service higher, artificial customer service as required, can Active participation at any time, therefore better service can be provided for client.

In concrete enforcement, the customer service of described robot also can comprise: conversation end detecting unit 67, is suitable for judging whether the session between described customer service client and described consumer user end terminates; Described status unit 64, is also suitable for when determining the conversation end between described customer service client and described consumer user end, and recovering the state of described robot customer service between described consumer user end in subsequent session business is session status.

When the conversation end between described customer service client and described consumer user end is determined in robot customer service, the state recovered in the subsequent session business between described consumer user end is session status, namely when artificial customer service has got involved, the state of the robot customer service that automatically switches back, realize the seamless connection of artificial customer service and robot customer service, thus the service experience of client can be optimized further.

In an embodiment of the present invention, described conversation end detecting unit 67 comprises following at least one:

First detection sub-unit (not shown), whether reach preset duration, and when reaching preset duration, determine the conversation end between described customer service client and described consumer user end if being suitable for detecting described second information interaction unit 65 free time;

Second detection sub-unit (not shown), be suitable for detecting the session settling signal from described customer service client, and when the session settling signal from described customer service client being detected, determine the conversation end between described customer service client and described consumer user end.

Such as, described first detection sub-unit determines that between described customer service client and described consumer user end, sessionless transmission of messages reaches preset duration, or the second detection sub-unit is when detecting that session window is closed or detect in the conversation message that described customer service client sends when comprising content that display session completes, the session settling signal that described customer service client sends automatically, artificial customer service can be automatically identified got involved, thus the operation of artificial customer service can be reduced, reduce the workload of artificial customer service, for artificial customer service is provided convenience.

The embodiment of the present invention additionally provides the customer service system that can realize above-mentioned client service method, with reference to the customer service system shown in Fig. 1.

In concrete enforcement, described consumer user end and customer service client all can present in a variety of forms.Such as, automatically run when detecting that browser page is opened during described consumer user end and customer service client can be built in browser program.And for example, described consumer user end and customer service client are independently applications client, be built in there is all-purpose computer function desktop computer, panel computer, mobile phone, can start under the operation of client and run in car-mounted terminal.For another example, described consumer user end and customer service client can also for being built in special equipment that product supplier or ISP customize for client, described special equipment can be used for client and product supplier or ISP in sales exhibition or service experience process to carry out seeking advice from and feeding back, to meet customer demand better.

Be understandable that, the way of realization of carrying out consumer user end and the customer service client conversed can be identical, also can be different, as long as compatible identical communication protocol can realize session.

Wherein the specific implementation such as concrete structure and operation principle of consumer user end 11, customer service client 12 and robot customer service 13 all see the introduction in the various embodiments described above, can repeat no more here.

In concrete enforcement, robot customer service can simultaneously with the session of multiple consumer user end, the conversation message received also can be forwarded to different customer service clients by robot customer service.When the consumer user end and customer service client that are in session status are multiple, when robot customer service is determining independently to process or described in None-identified during conversation message, replied as early as possible to make the conversation message of client, according to the task amount of customer service client being in session status, the pacifying property response message of the described conversation message from consumer user end and correspondence can be sent to the minimum customer service client of task amount.Such as, backstage has 10 artificial customer services, the customer service client that correspondence 10 is different respectively, can the quantity of consumer user end corresponding to the conversation message of each customer service client process current, described customer service robot can by the conversation message uniform distribution that receives to each customer service client.

In concrete enforcement, better consistent formula service is obtained in order to make client, improve the efficiency of service of each artificial customer service simultaneously, consumer user end and customer service client that the customer service of described robot is suitable for when being in session status are multiple, and determine self cannot independently to process or described in None-identified during conversation message, according to the task amount of customer service client being in session status, the described conversation message from consumer user end is sent to the minimum customer service client of task amount.In an embodiment of the present invention, the customer service of described robot can also just describedly cannot process or the pacifying property response message corresponding to conversation message of None-identified is sent to the customer service client that described consumer user end conversation message is crossed in described pre-treatment together, does not understand and described exchanging information of client and affect the experience of client to avoid artificial customer service.

Be understandable that, as required, in the customer service of described robot, message is forwarded in customer service client process, plurality of distribution strategy can be considered simultaneously, each allocation strategy can configure different priority, in concrete repeating process, can mate according to priority orders, be not described in detail here.

One of ordinary skill in the art will appreciate that all or part of step in the various methods of above-described embodiment is that the hardware that can carry out instruction relevant by program has come, this program can be stored in a computer-readable recording medium, and storage medium can comprise: ROM, RAM, disk or CD etc.

Although the present invention discloses as above, the present invention is not defined in this.Any those skilled in the art, without departing from the spirit and scope of the present invention, all can make various changes or modifications, and therefore protection scope of the present invention should be as the criterion with claim limited range.

Claims

1. a customer service system method of servicing, is characterized in that, comprising:

2. customer service system method of servicing according to claim 1, is characterized in that, described when the customer service of described robot detect self be not in session suspended state time, independently process described conversation message by the customer service of described robot, comprising:

Identify the content of described conversation message;

3. customer service system method of servicing according to claim 2, is characterized in that, described in autonomous processing procedure, when meeting the session suspension condition preset, performing the process customer service of described robot being switched to session suspended state, comprising:

4. customer service system method of servicing according to claim 3, is characterized in that, also comprise:

The customer service of described robot by described cannot independently process or None-identified conversation message corresponding to pacifying property response message cannot independently to process with described or None-identified conversation message together be sent to described customer service client.

5. customer service system method of servicing according to claim 1, is characterized in that, after the customer service of described robot detects self to be in session suspended state, also comprises:

When the customer service of described robot does not receive the conversation message from described customer service client, the content of conversation message described in the customer service identification of described robot, extracts with the pacifying property response message of the content matching of described conversation message and is sent to described consumer user end.

6. customer service system method of servicing according to claim 1, is characterized in that, also comprise:

Session status is in the session service of the customer service of described robot and described consumer user end, and when receiving the conversation message sent to described consumer user end from described customer service client, certainly will switch to session suspended state in the session status in the session service of described consumer user end, and the transmission performing conversation message between described customer service client and described consumer user end operates.

7. customer service system method of servicing according to claim 6, is characterized in that, also comprise:

When determining the conversation end between described customer service client and described consumer user end, the state recovered in the session service between described consumer user end is session status.

8. customer service system method of servicing according to claim 7, is characterized in that, the described conversation end determined between described customer service client and described consumer user end, comprises following at least one:

The session settling signal of described customer service client detected.

9. customer service system method of servicing according to claim 8, it is characterized in that, described customer service client sends described session settling signal when detecting that session window is closed or when detecting in the conversation message that described customer service client sends and comprise content that display session completes.

10. customer service system method of servicing according to claim 3, is characterized in that, described in pacifying property response message be suitable for characterizing customer service and be in and find the state of problem answers.

11. customer service system method of servicing according to claim 3, it is characterized in that, described Service Source comprises: language empirical data and customer service resource data.

12. customer service system method of servicing according to claim 11, is characterized in that, to be undertaken from the conversation message between consumer user end classifying the described customer service resource data splitting and obtain different scene type according to scene.

13. customer service system method of servicing according to claim 12, is characterized in that, also comprise: carry out the training of large data machine, to identify the scene type belonging to conversation message.

14. customer service system method of servicing according to claim 12 or 13, is characterized in that, the described content according to described conversation message, determine whether can independently process described conversation message, comprising:

15. customer service system method of servicing according to claim 14, is characterized in that, described when determining independently to process described conversation message, extract the response message of mating with the content of described conversation message, comprising from stored Service Source:

16. customer service system method of servicing according to claim 1, is characterized in that, the form of described conversation message comprises following at least one: the statement that natural language word forms, data link, voice, image and video.

17. customer service system method of servicing according to claim 16, it is characterized in that, described image comprise following one of them plant: static state or the dynamic picture with semantic information, or for client's session for product or the photo of service object or screenshotss information.

18. 1 kinds of robot customer services, is characterized in that, comprising:

The customer service of 19. robots according to claim 18, is characterized in that, describedly comprises from Main Processor Unit:

Memory cell, is suitable for stores service resource;

First response message generation unit, be suitable for when described ability judging unit determines that the customer service of described robot independently can process described conversation message, from the Service Source of described cell stores, extract the response message of mating with the content of described conversation message.

The customer service of 20. robots according to claim 19, it is characterized in that, described status unit, be suitable for when described ability judging unit determines that the customer service of described robot cannot independently process or described in None-identified during conversation message, controlling the customer service of described robot is being session suspended state with the state in the subsequent session business of described consumer user end;

The customer service of 21. robots according to claim 20, it is characterized in that, described second information interaction unit be also suitable for by described cannot independently process or None-identified conversation message corresponding to pacifying property response message cannot independently to process with described or None-identified conversation message together be sent to described customer service client.

The customer service of 22. robots according to claim 18, it is characterized in that, described from Main Processor Unit, also be suitable for detecting that the customer service of described robot is being session suspended state with the state in the session service of described consumer user end when described state detection unit, and described second information interaction unit is not when receiving the conversation message from described customer service client, identify the content of described conversation message, extract with the pacifying property response message of the content matching of described conversation message and be sent to described consumer user end.

The customer service of 23. robots according to claim 18, it is characterized in that, described status unit, also be suitable for being in session status in the session service of the customer service of described robot and described consumer user end, and described second information interaction unit is when receiving the conversation message from described customer service client, the customer service of described robot is switched to session suspended state with the state in the subsequent session business of described consumer user end, and controls described second information interaction unit execution message exchange operations.

The customer service of 24. robots according to claim 23, is characterized in that, also comprise: conversation end detecting unit, is suitable for judging whether the session between described customer service client and described consumer user end terminates; Described status unit, is also suitable for when determining the conversation end between described customer service client and described consumer user end, and recovering the state of described robot customer service between described consumer user end in subsequent session business is session status.

The customer service of 25. robots according to claim 24, is characterized in that, described conversation end detecting unit comprises following at least one:

First detection sub-unit, whether reach preset duration, and when reaching preset duration, determine the conversation end between described customer service client and described consumer user end if being suitable for detecting the described second information interaction unit free time; Second detection sub-unit, is suitable for detecting the session settling signal from described customer service client, and when the session settling signal from described customer service client being detected, determines the conversation end between described customer service client and described consumer user end.

The customer service of 26. robots according to claim 20, it is characterized in that, the Service Source that described memory cell stores comprises: language empirical data and customer service resource data.

The customer service of 27. robots according to claim 26, it is characterized in that, also comprise machine training unit, be suitable for carrying out the training of large data machine to described robot, to identify the scene type belonging to conversation message, described scene type is according to scene, the conversation message between consumer user end is carried out classification fractionation to obtain.

28. robot customer services according to claim 26 or 27, it is characterized in that, described ability judging unit, comprising:

The customer service of 29. robots according to claim 28, is characterized in that, described first response message generation unit, comprising:

The customer service of 30. robots according to claim 20, is characterized in that, described in pacifying property response message be suitable for characterizing customer service and be in and find the state of problem answers.

The customer service of 31. robots according to claim 18, is characterized in that, the form of described conversation message comprises following at least one: the statement that natural language word forms, data link, voice, image and video.

The customer service of 32. robots according to claim 31, is characterized in that, described image comprise following one of them plant: static state or the dynamic picture with semantic information, or for client's session for product or the photo of service object or screenshotss information.

33. 1 kinds of customer service systems, it is characterized in that, comprise: consumer user end, be suitable for the customer service client that conversates with described consumer user end and the robot customer service described in any one of claim 18-32, described customer service client and described consumer user end are by the customer service session of described robot.

34. customer service systems according to claim 33, it is characterized in that, consumer user end and customer service client that the customer service of described robot is suitable for when being in session status are multiple, and determine self cannot independently to process or described in None-identified during conversation message, according to the task amount of customer service client being in session status, the pacifying property response message of the described conversation message from consumer user end and correspondence is sent to the minimum customer service client of task amount.

35. customer service systems according to claim 33, it is characterized in that, consumer user end and customer service client that the customer service of described robot is suitable for when being in session status are multiple, and determine self cannot independently to process or described in None-identified during conversation message, according to conversation message record, self cannot independently to process described or the customer service client of described consumer user end conversation message is crossed in pre-treatment that the conversation message of None-identified is sent to.