Customer Service Information Processing Method, Device and Equipment

Abstract

The present application discloses a customer service information processing method, apparatus and device. Among them, the method includes: obtaining a customer service conversation message stream associated with an object to be queried; determining at least one customer service session and a plurality of conversation messages corresponding to the customer service session according to the conversation message stream; determining problem scenario information of the customer service session according to the plurality of conversation messages corresponding to the customer service session; and determining problem scenario distribution information of the object to be queried according to the problem scenario information of the customer service session. By adopting this processing method, the disordered customer service conversation messages can be converted into consultation scenario insights available for merchants to analyze, such as supporting analysis in multiple dimensions of time, merchant, consultation scenario and session details; therefore, the accuracy of the consumption problem consultation distribution information can be effectively improved, and the analysis and processing efficiency can be improved.

Description

Technical Field

[0001] This application relates to the technical field of data processing, and in particular to a customer service information processing method and apparatus, and an electronic device. Background Art

[0002] An e-commerce platform includes a transaction system, a logistics system, and a customer service dialogue system. Consumers can consult various consumption problems with the customer service staff of merchants through the customer service dialogue system, such as when this cosmetic will be shipped, whether the size of this piece of clothing is too large, too small, or of regular size, and so on. Merchants on the platform need to understand the distribution of consumption problem consultations to reasonably allocate customer service resources, etc. Currently, the traditional database query and retrieval method is mainly used to obtain the distribution information of consumption problem consultations.

[0003] However, in the process of implementing the present invention, the inventors found that the above solution has at least the following problems: the accuracy and processing efficiency of the distribution information of consumption problem consultations are both relatively low. Summary of the Invention

[0004] This application provides a customer service information processing method to solve the problem that the accuracy and processing efficiency of the distribution information of consumption problem consultations in the prior art are both relatively low. This application also provides a customer service information processing apparatus and an electronic device.

[0005] This application provides a method for processing commodity objects, including:

[0006] Obtaining a customer service dialogue message stream associated with the object to be queried;

[0007] Determining at least one customer service session and a plurality of dialogue messages corresponding to the customer service session according to the dialogue message stream;

[0008] Determining the problem scenario information of the customer service session according to the plurality of dialogue messages corresponding to the customer service session;

[0009] Determining the problem scenario distribution information of the object to be queried according to the problem scenario information of the customer service session.

[0010] Optionally, the determining at least one customer service session and a plurality of dialogue messages corresponding to the customer service session according to the dialogue message stream includes:

[0011] Obtaining first user information, second user information, the sending date information, and the sending time information corresponding to the dialogue message;

[0012] Determining the at least one customer service session according to the first user information, the second user information, and the sending date information;

[0013] Obtain multiple conversation messages corresponding to the customer service session according to the first user information, second user information, sending date information, and sending time information.

[0014] Optionally, determining at least one customer service session and multiple conversation messages corresponding to the customer service session according to the conversation message flow includes:

[0015] Obtain the first user information, second user information, sending date information, and sending time information corresponding to the conversation message;

[0016] Obtain a session greeting and a session closing statement according to the first user information, second user information, and sending date information;

[0017] Determine the at least one customer service session according to the session greeting and the session closing statement;

[0018] Obtain multiple conversation messages corresponding to the customer service session according to the first user information, second user information, sending date information, and sending time information.

[0019] Optionally, determining the problem scenario information of the customer service session according to multiple conversation messages corresponding to the customer service session includes:

[0020] Obtain the first user conversation message according to multiple conversation messages corresponding to the customer service session;

[0021] Determine the problem scenario information of the first user conversation message;

[0022] Determine the problem scenario information of the customer service session according to the problem scenario information of the first user conversation message.

[0023] Optionally, obtaining the first user conversation message according to multiple conversation messages corresponding to the customer service session includes:

[0024] Obtain the sender identity information of the conversation message;

[0025] Use the conversation message with the sender identity information being the first user as the first user conversation message.

[0026] Optionally, determining the problem scenario information of the first user conversation message includes:

[0027] Determine the problem scenario information of the first user conversation message through a problem category prediction model.

[0028] Optionally, determining the problem scenario distribution information of the object to be queried according to the problem scenario information of the customer service session includes:

[0029] Obtain the number of user inquiries and / or the number of customer service sessions of the object to be queried in at least one time period and in different problem scenarios according to the problem scenario information of the customer service session.

[0030] Optionally, it further includes at least one of the following steps:

[0031] Obtain the total number of user inquiries and / or the total number of customer service sessions of the object to be queried in at least one time period and in multiple problem scenarios according to the problem scenario information of the customer service session;

[0032] Obtain the accuracy of the problem scenario information of the customer service session, and obtain the ranking information of multiple customer service sessions in the same problem scenario in the same time period according to the accuracy.

[0033] Optionally, it further includes:

[0034] Determine the second user scheduling information according to the problem scenario distribution information of the object to be queried.

[0035] Optionally, execute the method at preset time intervals.

[0036] Optionally, it further includes at least one of the following steps:

[0037] Filter out the system messages in the conversation message stream;

[0038] Filter out the duplicate messages in the conversation message stream;

[0039] If it is determined that there are no unreceived conversation messages of the same customer service session before the currently received conversation message, determine the session end time information of the corresponding customer service session according to the sending time information of the currently received conversation message;

[0040] Filter out the invalid customer service sessions in the at least one customer service session.

[0041] This application also provides a customer service information processing method, including:

[0042] Send a request for obtaining problem scenario distribution information;

[0043] Receive the problem scenario distribution information of the object to be queried;

[0044] Display the problem scenario distribution information of the object to be queried;

[0045] Among them, the problem scenario distribution information of the object to be queried is determined in the following manner:

[0046] Obtain the customer service conversation message flow associated with the object to be queried; determine at least one customer service session and multiple conversation messages corresponding to the customer service session according to the conversation message flow; determine the problem scenario information of the customer service session according to the multiple conversation messages corresponding to the customer service session; in response to the request, obtain the problem scenario distribution information of the object to be queried according to the problem scenario information of the customer service session of the object to be queried.

[0047] Optionally, the problem scenario distribution information includes: the customer service consultation user volume and / or the customer service session volume of the object to be queried in at least one time period and in the problem scenario.

[0048] This application also provides an electronic device, including:

[0049] A processor; and

[0050] A memory for storing a program for implementing the above-mentioned customer service information processing method, and the terminal is powered on and runs the program of this method through the processor.

[0051] This application also provides a computer-readable storage medium, in which instructions are stored, and when it runs on a computer, it causes the computer to execute the above various methods.

[0052] This application also provides a computer program product including instructions, and when it runs on a computer, it causes the computer to execute the above various methods.

[0053] Compared with the prior art, this application has the following advantages:

[0054] The customer service information processing method provided by the embodiments of this application obtains the customer service conversation message flow associated with the object to be queried; determines at least one customer service session and multiple conversation messages corresponding to the customer service session according to the conversation message flow; determines the problem scenario information of the customer service session according to the multiple conversation messages corresponding to the customer service session; determines the problem scenario distribution information of the object to be queried (merchant) according to the problem scenario information of the customer service session. By adopting this processing method, the customer service session can be restored based on the disordered customer service conversation messages, and the problem scenario of the session can be determined, and then the problem scenario distribution can be analyzed based on the problem scenario information of the customer service session, thereby converting the disordered messages into consultation scenario insights that can be analyzed by merchants, such as supporting analysis in multiple dimensions of time, merchant, consultation scenario, and session details; therefore, it can effectively improve the accuracy of the consumption problem consultation distribution information and improve the analysis and processing efficiency. Description of the Drawings

[0055] Figure 1 The flowchart of the commodity object processing method provided by this application;

[0056] Figure 2Scenario schematic diagram of the commodity object processing method provided by this application;

[0057] Figure 3 Device interaction schematic diagram of the commodity object processing method provided by this application. Detailed implementation manners

[0058] Numerous specific details are set forth in the following description in order to provide a thorough understanding of this application. However, this application can be implemented in many other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the connotation of this application. Therefore, this application is not limited by the specific implementations disclosed below.

[0059] In this application, a customer service information processing method and apparatus, as well as an electronic device, are provided. Each of the various solutions will be described in detail in the following embodiments.

[0060] First Embodiment

[0061] Please refer to Figure 1 , which is a schematic flowchart of an embodiment of the customer service information processing method of this application. In this embodiment, the method may include the following steps:

[0062] Step S101: Obtain the customer service conversation message stream associated with the object to be queried.

[0063] There are a large number of real-time consultation messages from consumers and merchants in the e-commerce service operation scenario. The method provided by the embodiment of this application can utilize the real-time customer service conversation messages to help merchants more directly understand the distribution of consumer problem consultations in real time, display the current status of merchant service reception, assist merchants in reasonably scheduling customer service resources, and maximize the value of customer service resources.

[0064] Such as Figure 2As shown in the figure, a system includes a server, a first user terminal (consumer terminal), a second user terminal (customer service staff terminal), and a terminal of the object to be queried (merchant terminal). The server can be a server cluster. A customer service dialogue system and a customer service information processing system are deployed on the server. The customer service dialogue system and the customer service information processing system can be deployed on different servers. Consumers (the first users) can use their terminals (such as smartphones, tablets, personal computers, etc.) to consult various consumption problems with the customer service staff (the second users) of the merchant (the object to be queried) through the customer service dialogue system, such as "When will this cosmetic be shipped?", "Is the size of this piece of clothing larger, smaller, or regular size?", and so on. The customer service staff (the second users) can use the second user terminal (such as smartphones, tablets, personal computers, etc.) to answer the questions raised by consumers (the first users) through the customer service dialogue system, such as "Okay, let me check for you", and so on. Merchants (the objects to be queried) on the e-commerce platform can use their terminals (such as smartphones, tablets, personal computers, etc.) to understand the distribution of consumption problem consultations through the customer service information processing system to reasonably allocate customer service resources, etc.

[0065] The dialogue messages are the communication chats between consumers and merchant customer services during the daily shopping process, which can be text messages, voice messages, or card messages, picture messages, etc.

[0066] The dialogue message stream may include multiple dialogue messages in which messages of different customer service sessions associated with the object to be queried are interspersed. The dialogue message stream can be an ordered dialogue message stream, that is, the messages in the dialogue message stream are in the order of the actual sending time of the messages. The dialogue message stream can also be an unordered dialogue message stream, that is, the messages in the dialogue message stream may not be in the order of the actual message sending time. For example, the actual sending time of a later obtained dialogue message may be earlier than the actual sending time of an earlier obtained dialogue message. Table 1 below shows the dialogue message stream of merchant B with consumers A, C, D, and E on the morning of March 11, 2022 in this embodiment.

[0067]

[0068]

[0069] Table 1. Dialogue Message Stream

[0070] From the data in Table 1, it can be seen that the dialogue message stream includes multiple dialogue messages in which messages of different customer service sessions are interspersed. Among them, the message ID represents the message identifier; the message sending time is the actual sending time, not the actual receiving time; the sender is the initiator of the message, which can be the second user information (merchant customer service) or the first user information (consumer); the receiver is the recipient of the message. Among them, the actual customer service can be determined jointly by the merchant name and the customer service name.

[0071] In one example, the method may further include the following steps: filtering out system messages in the conversation message stream. The system messages include but are not limited to: system stress test messages, card messages, and evaluation information. For example, message 1005 in Table 1 is a card message "Are you satisfied with my service?" sent by the customer service robot 3, which can be filtered out. By adopting this processing method, the conversation content that is not related to the problem scenario and does not belong to the conversation between the consumer and the customer service is filtered out, which can effectively improve the analysis and processing efficiency and also improve the accuracy of the consumption problem consultation distribution information.

[0072] In one example, the method may further include the following steps: filtering out duplicate messages in the conversation message stream. For example, if message 1006 is sent twice, only one message needs to be retained to avoid the situation where the same consumer sends multiple messages to the same customer service at the same time. By adopting this processing method, the effectiveness of the customer service conversation messages can be effectively improved.

[0073] When specifically implemented, the filtering out of duplicate messages in the conversation message stream can be achieved in the following manner: constructing a message primary key according to the first user information, the second user information corresponding to the conversation message, and the first identification information (such as the message ID in Table 1) and the sending time information of the conversation message. The sending time can be in seconds or milliseconds, etc. By adopting this processing method, the message primary key is constructed based on the merchant, the consumer, the customer service, the message time, and the message ID, which can effectively avoid the situation of duplicate messages within the same second, and thus can effectively ensure the effectiveness of the customer service conversation messages. Table 2 shows the message primary keys corresponding to the messages in Table 1:

[0074] Message primary key Message ID 1001_A_B_1_2022-3-11 10:00:00 1001 1002_A_B_1_2022-3-11 10:00:05 1002 1003_A_B_2_2022-3-11 10:01:00 1003 1004_A_B_2_2022-3-11 10:01:05 1004 1005_A_B_3_2022-3-11 10:02:00 1005 1006_C_B_1_2022-3-11 10:30:00 1006 1007_C_B_2_2022-3-11 10:30:00 1007 1008_E_B_2_2022-3-11 10:45:00 1008 …

[0075] Table 2. Message primary key data

[0076] Step S103: Determine at least one customer service conversation and multiple conversation messages corresponding to the customer service conversation according to the conversation message stream.

[0077] The customer service conversation refers to a complete conversation process between a consumer and a merchant's customer service. One conversation may include multiple conversation messages, and one conversation corresponds to a merchant, a second user (customer service), and a first user (consumer). Table 3 shows the conversation restoration information in this embodiment:

[0078] Message primary key Message ID Session identifier 1001_A_B_1_2022-3-11 10:00:00 1001 1 1002_A_B_1_2022-3-11 10:00:05 1002 1 1003_A_B_2_2022-3-11 10:01:00 1003 2 1004_A_B_2_2022-3-11 10:01:05 1004 2 1006_C_B_1_2022-3-11 10:30:00 1006 3 1007_C_B_2_2022-3-11 10:30:00 1007 4 1008_E_B_2_2022-3-11 10:45:00 1008 5 …

[0079] Table 3. Conversation data

[0080] As can be seen from Table 3, messages 1001 and 1002 form Session 1, which is a session involving "Consumer A, Merchant B, and Customer Service 1" on March 11; messages 1003 and 1004 form Session 2, which is a session involving "Consumer A, Merchant B, and Customer Service 2" on March 11, and so on.

[0081] In one example, step S103 may include the following sub-steps:

[0082] Step S1031: Obtain the first user information, second user information, the sending date information, and the sending time information corresponding to the conversation message.

[0083] In specific implementation, the corresponding merchant information, second user information, first user information, sending date information, and sending time information can be obtained from the attribute data of the conversation message. For example, various information corresponding to the conversation message can be obtained through the data provided in Table 1.

[0084] In specific implementation, the first user information, second user information, merchant information, the sending date information, and the sending time information corresponding to the conversation message can also be obtained according to the above message primary key.

[0085] Step S1033: Determine the at least one customer service session according to the first user information, second user information, and sending date information.

[0086] In this embodiment, the at least one customer service session can be determined according to the first user information, second user information, and sending date information. By adopting this processing method, the whole-day communication between the service recipient and the service provider is regarded as a session service. The service provider is the provider of the service, such as the customer service of a certain merchant, and the service recipient is the demander of the service, such as the consumer buyer.

[0087] Step S1035: Obtain multiple conversation messages corresponding to the customer service session according to the first user information, second user information, sending date information, and sending time information.

[0088] Regarding the conversation messages of the whole-day communication between the service recipient and the service provider as the conversation messages included in a session service. For example, all the conversation messages of "Consumer A, Merchant B, and Customer Service 1" in one day form the content of one session.

[0089] In another example, step S103 may include the following sub-steps:

[0090] Step S1031': Obtain the first user information, second user information, the sending date information, and the sending time information corresponding to the conversation message.

[0091] Step S1033’: Obtain a session greeting and a session closing message according to the first user information, the second user information, and the sending date information.

[0092] Step S1035’: Determine the at least one customer service session according to the session greeting and the session closing message;

[0093] Step S1037’: Obtain a plurality of conversation messages corresponding to the customer service session according to the first user information, the second user information, the sending date information, and the sending time information.

[0094] By adopting this processing method, the greeting (such as "Hello") is used as the starting message of the session, the closing message (such as "Goodbye") is used as the ending message of the session, and the communication during this period is regarded as a session service. In this way, for the same group of merchants, customer service representatives, and consumers, multiple session services within a day can be restored. Therefore, the accuracy of session restoration can be effectively improved.

[0095] In specific implementation, the customer service session information restored based on the conversation message flow may further include the session start time and the session end time. In practical applications, the session start time and the session end time can be shown to the merchant to enable it to understand more detailed session information.

[0096] In the process of implementing the present invention, the inventor found that in practical applications, multiple messages of the same session may not be transmitted in sequence. For example, session 1 includes message 1001 and message 1002, the sending time of message 1001 is earlier than that of message 1002, but in the actual transmission process, message 1002 is received first, and then message 1001 is received. If the session end time is simply updated according to the sending time of the currently received message, the session end time will be the sending time of message 1001, which is not the true end time of this session. Therefore, the session end time cannot be accurately restored.

[0097] To accurately restore the session end time, in one example, the method may further include the following steps: If it is determined that there are no unreceived conversation messages of the same customer service session before the currently received conversation message, then determine the session end time information of the corresponding customer service session according to the sending time information of the currently received conversation message. By adopting this processing method, after a conversation message is obtained in real time, only when it is determined that there are no unobtained conversation messages of the same customer service session before this message, will the session end time information of the corresponding customer service session be updated according to the sending time information of this message; Therefore, the session end time can be accurately restored.

[0098] During specific implementation, a waiting duration threshold (such as 1 minute) can be set. If, after receiving a conversation message, no other conversation messages in the same conversation that occurred before this message are received after the waiting duration threshold is reached, then this conversation message is regarded as the latest message of the conversation, and the end time of the conversation is updated according to the sending time of the latest message.

[0099] During specific implementation, it is also possible to group by the primary key of the message and retrieve the latest field information of each message in real time to ensure the orderliness of real-time chat messages.

[0100] Step S105: Determine the problem scenario information of the customer service conversation based on multiple conversation messages corresponding to the customer service conversation.

[0101] The problem scenario is also known as the consultation scenario, consultation category, or problem category, which is to classify the problems of a conversation. In the e-commerce customer service scenario, the problem scenarios include but are not limited to: logistics delivery, refund and after-sales, and product consultation. For example, if the buyer's questions in a conversation are mainly about logistics delivery, then the problem scenario of this conversation is logistics delivery; if the buyer's questions are mainly about refund and after-sales, then the problem scenario is refund and after-sales; if the buyer's questions are mainly about asking for product details, then the problem scenario is product consultation.

[0102] In one example, step S105 may include the following sub-steps:

[0103] Step S1051: Retrieve the first user's conversation message based on multiple conversation messages corresponding to the customer service conversation.

[0104] During specific implementation, the sender identity of each conversation message can be determined to distinguish the identities of the service provider (merchant customer service, i.e., the second user) and the service recipient (consumer, i.e., the first user), and then the first user's conversation message is extracted from all the messages of a conversation, that is, the messages sent by the consumer are extracted.

[0105] In one example, step S1051 can be implemented in the following way: Obtain the sender identity information of the conversation message, and the sender identity can be the first user (consumer) or the second user (customer service); The conversation message with the sender identity information being the first user is used as the first user's conversation message. For example, the sender identity can be determined according to "whether the sender is a consumer" in Table 2. If it is "yes", the sender identity is the first user; if it is "no", the sender identity is the second user, and the second user can be a real person or a customer service robot.

[0106] Step S1053: Determine the problem scenario information of the first user's conversation message.

[0107] In specific implementation, step S1053 can be implemented in the following manner: Determine the problem scenario information of the first user conversation message through the problem category prediction model. By adopting this processing method, for the messages with the problem initiator being the consumer, the problem scenario of the consumer's consultation is predicted through the algorithm model, and the consultation scenario label is marked for each buyer message. For example, when the consumer asks: When will this cosmetic be shipped, the corresponding problem scenario is: Logistics delivery. Table 4 shows the session problem scenario information of this embodiment.

[0108]

[0109] Table 4. Problem scenario data

[0110] Step S1055: Determine the problem scenario information of the customer service session according to the problem scenario information of the first user conversation message.

[0111] After obtaining the scenario information corresponding to each problem of the first user in a session, the problem scenario information of the customer service session can be determined. For example, in Table 1, 1001 and 1002 are the messages included in a session. Since message 1001 belongs to the logistics delivery problem, it can be determined that the problem scenario of this session is logistics delivery.

[0112] By adopting the processing methods of the above steps S1051 to S1055, the problem scenario of a session is determined based on the consumer's conversation message. Therefore, the determination efficiency and accuracy of the session problem scenario can be effectively improved.

[0113] In an example, the method may further include the following steps: Filter out the invalid customer service sessions in the at least one customer service session. By adopting this processing method, for the real-time session information, the invalid scenarios without a problem scenario, with the problem scenario being "user chat" and "others" can be filtered. For example, since the problem scenario corresponding to session 4 in Table 3 is "others", session 4 is filtered out.

[0114] Step S107: Determine the problem scenario distribution information of the object to be queried according to the problem scenario information of the customer service session.

[0115] The problem scenario distribution information of the object to be queried refers to the distribution of consumers of a merchant consulting questions with the merchant. The problem scenario distribution information of the object to be queried includes, but is not limited to, at least one of the following: the number of customer service consulting users of the object to be queried in at least one time period and in different problem scenarios, and the number of customer service conversations of the object to be queried in at least one time period and in different problem scenarios. The number of customer service consulting users is also called the number of consulting people, that is, the number of buyers who have conversations with the customer service. The number of customer service conversations is also called the number of consulting conversations, that is, the number of times of customer service conversations. The time period can be a time period in days, such as the most recent day. The time period can also be a time period in hours, such as the most recent hour.

[0116] In one example, step S107 can be implemented in the following manner: Determine the number of customer service consulting users of the object to be queried in at least one time period and in different problem scenarios according to the problem scenario information of the customer service conversation, or obtain the number of customer service conversations, or obtain the number of customer service consulting users and the number of customer service conversations. Adopting this processing method enables analysis in multiple dimensions of time, merchant, consulting scenario, and conversation details; therefore, it can help merchants gain insights into the consulting scenario.

[0117] For example, according to the data in Tables 1 to 4 above, the number of consulting people regarding the "logistics delivery" consulting scenario of merchant B on March 11, 2011 is 2, and the number of consulting conversations is 2. Regarding the "refund and after-sales" consulting scenario, the number of consulting people is 2, and the number of consulting conversations is 2.

[0118] For another example, according to the data in Tables 1 to 4 above, the number of consulting people regarding the "logistics delivery" consulting scenario of merchant B in the most recent 1 hour (current 11:10:00 on March 11, 2022) on March 11, 2011 is 1, and the number of consulting conversations is 1. Regarding the "refund and after-sales" consulting scenario, the number of consulting people is 1, and the number of consulting conversations is 1.

[0119] In one example, the method may further include the following steps: Determine the total number of customer service consulting users and / or the total number of customer service conversations of the object to be queried in at least one time period and in multiple problem scenarios according to the problem scenario information of the customer service conversation. For example, according to the data in Tables 1 to 4 above, the number of consulting people of merchant B on March 11, 2011 is 4, and the number of consulting conversations is 5. For example, according to the data in Tables 1 to 4 above, the number of consulting people of merchant B in the most recent 1 hour (current 11:10:00 on March 11, 2022) on March 11, 2011 is 3, and the number of consulting conversations is 3.

[0120] In one example, the method may further include the following steps: obtaining the accuracy of the problem scenario information of the customer service session; and obtaining the ranking information of multiple customer service sessions in the same problem scenario within the same time period according to the accuracy. By adopting this processing method, the credibility of session restoration can be provided to the merchant, and the merchant can make decisions on processing methods such as customer service scheduling based on this.

[0121] Specifically in implementation, the above problem category prediction model can give the accuracy (such as probability) of the predicted problem scenario information, and this accuracy can be used as the accuracy of the problem scenario information of the customer service session, and this accuracy can be used to rank multiple customer service sessions of a merchant.

[0122] In one example, the method may further include the following steps: determining the second user scheduling information according to the problem scenario distribution information of the object to be queried. For example, if the merchant finds that there are more sessions in a certain problem scenario according to the problem scenario distribution information, the number of customer services for this problem scenario can be increased. By adopting this processing method, the efficiency and accuracy of customer service scheduling can be effectively improved.

[0123] In one example, the method may be executed according to a preset time interval to realize real-time statistics of data such as the customer service consultation volume of the merchant at a fixed sliding frequency. The time interval can be a time interval at the minute level, such as the time interval is 1 minute. By adopting this processing method, the customer service session is restored based on the disordered real-time customer service dialogue messages, and the problem scenario of the session is determined. Then, a multi-dimensional analysis and statistics are constructed based on the problem scenario information of the customer service session, thereby converting the disordered real-time messages into real-time multi-dimensional consultation scenario insights available for the merchant to analyze, which can support analysis in multiple dimensions of time, merchant, consultation scenario, and details at the same time; therefore, the timeliness of the consumer problem consultation distribution information can be effectively improved.

[0124] For example, statistics are made according to the dimensions of day and merchant: from 0:00 on the current day to the current time, the number of consulting users and consulting session volumes of the top 10 problem scenarios are statistically counted according to the merchant granularity, and the top 100 real-time sessions with the accuracy of each problem statistic (that is, the accuracy of predicting the problem scenario to which the session belongs) are supported for the merchant to view the detailed content of the session.

[0125] For another example, statistics are made according to the dimensions of the last 1 hour, merchant, and consultation scenario: the consumer consultation volume and consultation session volume are statistically counted according to the consultation scenario of the merchant in the last 1 hour, and the window is updated by sliding every 1 minute; the top 20 real-time sessions with the accuracy of each problem scenario statistic are provided, and the merchant is supported to view the detailed content of the session. That is, the ranking of consumer consultation problems in each dimension is shown to the merchant, and the high-frequency problem cases and all messages of a single session can be viewed according to different granularities of the session, helping the merchant better locate the consumer problems.

[0126] In this embodiment, through multi-dimensional statistics and data output multi-dimensional analysis of real-time message preprocessing (such as filtering system messages, duplicate messages, ensuring the normal order of messages, etc.), tagging of consultation questions (determining the problem scenarios of consumers' questions), real-time scenario session restoration (such as the communication of the same group of users in one day is regarded as one session), and sliding window (such as executing the method provided in this embodiment of the application once every minute), it helps merchants directly understand the distribution of consumer problem consultations more real-timely, display the current status of merchant service acceptance, assist merchants in reasonably scheduling customer service resources, and maximize the value of customer service resources.

[0127] In one example, the method may further include the following steps: output the statistically real-time data to a database (such as Lindorm) for storage according to the dimensions of time, session, and session content, as shown in Table 5 below:

[0128]

[0129]

[0130] Table 5. Multi-dimensional real-time window statistical analysis information

[0131] In one example, step S107 can be implemented in the following manner: in response to a problem scenario distribution information acquisition request, according to the problem scenario information of the customer service session, obtain the customer service consultation user volume and / or customer service session volume of the object to be queried in at least one time period and in different problem scenarios.

[0132] Specifically, it can be that the object to be queried sends a problem scenario distribution information acquisition request to the server through the terminal device it uses. The object to be queried can input the query date and time dimension type (such as day or the last 1 hour). The server executes the method to obtain the top 10 problem distribution information of the object to be queried in real time on the same day, or obtain the top 10 problem distribution information of the object to be queried in real time in the last 1 hour, or simultaneously obtain the top 10 problem distribution information of the object to be queried in real time on the same day and in the last 1 hour.

[0133] In one example, the method may further include the following steps: in response to a customer service dialogue acquisition request, obtain the customer service session information of the object to be queried in the target time period and in the target problem scenario; in addition, it may further include the following steps: in response to a dialogue content acquisition request, obtain the content information of the target customer service session of the object to be queried. For example, when the merchant clicks on a specific consultation scenario, the server uses range query to obtain and display the top 100 session cases stored in the database and display the session details.

[0134] As can be seen from the above embodiments, the customer service information processing method provided by the embodiments of the present application obtains the customer service conversation message stream associated with the object to be queried; determines at least one customer service session and multiple conversation messages corresponding to the customer service session according to the conversation message stream; determines the problem scenario information of the customer service session according to the multiple conversation messages corresponding to the customer service session; and determines the problem scenario distribution information of the object to be queried (merchant) according to the problem scenario information of the customer service session. By adopting this processing method, the customer service session is restored based on the unordered customer service conversation messages, the problem scenario of the session is determined, and then the problem scenario distribution is analyzed based on the problem scenario information of the customer service session, thereby converting the unordered messages into consultation scenario insights that can be analyzed by merchants, such as supporting analysis in multiple dimensions of time, merchant, consultation scenario, and session details; therefore, the accuracy of the consultation distribution information of consumption problems can be effectively improved, and the analysis and processing efficiency can be improved.

[0135] Second Embodiment

[0136] In the above embodiment, a customer service information processing method is provided. Correspondingly, the present application also provides a customer service information processing device. This device corresponds to the embodiment of the above method. The parts of this embodiment that are the same as those in the first embodiment will not be described again. Please refer to the corresponding parts in the first embodiment.

[0137] A customer service information processing device provided by the present application includes: a message stream acquisition unit, a customer service session determination unit, a problem scenario determination unit, and a problem scenario distribution determination unit.

[0138] The message stream acquisition unit is used to acquire the customer service conversation message stream associated with the object to be queried; the customer service session determination unit is used to determine at least one customer service session and multiple conversation messages corresponding to the customer service session according to the conversation message stream; the problem scenario determination unit is used to determine the problem scenario information of the customer service session according to the multiple conversation messages corresponding to the customer service session; the problem scenario distribution determination unit is used to determine the problem scenario distribution information of the object to be queried according to the problem scenario information of the customer service session.

[0139] In an example, the customer service session determination unit is specifically configured to acquire the first user information, the second user information, the sending date information, and the sending time information corresponding to the conversation message; determine the at least one customer service session according to the first user information, the second user information, and the sending date information; and acquire multiple conversation messages corresponding to the customer service session according to the first user information, the second user information, the sending date information, and the sending time information.

[0140] In one example, the customer service session determination unit is specifically configured to obtain first user information, second user information, the sending date information, and the sending time information corresponding to the conversation message; obtain a session greeting and a session closing message according to the first user information, the second user information, and the sending date information; determine the at least one customer service session according to the session greeting and the session closing message; and obtain a plurality of conversation messages corresponding to the customer service session according to the first user information, the second user information, the sending date information, and the sending time information.

[0141] In one example, the problem scenario determination unit is specifically configured to obtain a first user conversation message according to a plurality of conversation messages corresponding to the customer service session; determine the problem scenario information of the first user conversation message; and determine the problem scenario information of the customer service session according to the problem scenario information of the first user conversation message.

[0142] In one example, the problem scenario determination unit is specifically configured to obtain the sender identity information of the conversation message; and use the conversation message with the sender identity information being the first user as the first user conversation message.

[0143] In one example, the problem scenario determination unit is specifically configured to determine the problem scenario information of the first user conversation message through a problem category prediction model.

[0144] In one example, the problem scenario distribution determination unit is specifically configured to obtain the customer service consultation user volume and / or the customer service session volume of the object to be queried in at least one time period and in different problem scenarios according to the problem scenario information of the customer service session.

[0145] In one example, the device may further include at least one of the following units:

[0146] The data summary unit is configured to obtain the total customer service consultation user volume and / or the total customer service session volume of the object to be queried in at least one time period and in multiple problem scenarios according to the problem scenario information of the customer service session.

[0147] The session ranking unit is configured to obtain the accuracy of the problem scenario information of the customer service session, and obtain the ranking information of a plurality of customer service sessions in the same problem scenario in the same time period according to the accuracy.

[0148] In one example, the device may further include: a customer service scheduling unit, configured to determine second user scheduling information according to the problem scenario distribution information of the object to be queried.

[0149] In one example, the device may further include at least one of the following units:

[0150] A system message filtering unit, configured to filter system messages in the conversation message stream;

[0151] A duplicate message filtering unit, configured to filter duplicate messages in the conversation message stream;

[0152] A session time determination unit, configured to, if it is determined that there are no un - received conversation messages of the same customer service session before the currently received conversation message, determine the session end time information of the corresponding customer service session according to the sending time information of the currently received conversation message;

[0153] An invalid session filtering unit, configured to filter invalid customer service sessions in the at least one customer service session.

[0154] The third embodiment

[0155] This application also provides an electronic device. Since the device embodiments are basically similar to the method embodiments, the description is relatively simple. For related parts, please refer to the partial description of the method embodiments. The device embodiments described below are only illustrative.

[0156] An electronic device according to this embodiment, the electronic device includes: a processor and a memory; the memory is used to store a program for implementing the customer service information processing method, and the device is powered on and runs the program of this method through the processor.

[0157] The fourth embodiment

[0158] In the above - mentioned embodiment, a customer service information processing method is provided. Correspondingly, this application also provides a customer service information processing method, which can be used for the client of the object to be queried. This method corresponds to the system embodiment above. For parts that are the same as those in the first embodiment, they will not be repeated here. Please refer to the corresponding parts in the first embodiment.

[0159] The customer service information processing method of this application may include the following steps:

[0160] Step 1: Send a request for obtaining problem scenario distribution information.

[0161] Step 2: Receive the problem scenario distribution information of the object to be queried.

[0162] Step 3: Display the problem scenario distribution information of the object to be queried.

[0163] Such as Figure 3As shown, it is to determine the interaction process between the client and the server of the object to be queried. Among them, the problem scenario distribution information of the object to be queried is determined in the following manner: Obtain the customer service conversation message stream associated with the object to be queried; According to the conversation message stream, determine at least one customer service session and multiple conversation messages corresponding to the customer service session; According to the multiple conversation messages corresponding to the customer service session, determine the problem scenario information of the customer service session; In response to the request, obtain the problem scenario distribution information of the object to be queried according to the problem scenario information of the customer service session of the object to be queried. The problem scenario distribution information includes, but is not limited to: the customer service consultation user volume and / or the customer service session volume of the object to be queried in at least one time period and in the problem scenario.

[0164] Although this application is disclosed above in preferred embodiments, it is not used to limit this application. Any person skilled in the art can make possible changes and modifications without departing from the spirit and scope of this application. Therefore, the protection scope of this application should be subject to the scope defined by the claims of this application.

[0165] In a typical configuration, a computing device includes one or more processors (CPUs), an input / output interface, a network interface, and memory.

[0166] The memory may include non-permanent memory in the computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of, for example, read-only memory (ROM) or flash memory (flash RAM). The memory is an example of a computer-readable medium.

[0167] 1. Computer-readable media include permanent and non-permanent, removable and non-removable media that can store information by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette tapes, magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transitory media such as modulated data signals and carrier waves.

[0168] 2. Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code.

Claims

1. A customer service information processing method, characterized in that, Including: Obtain the customer service conversation message flow associated with the object to be queried. The conversation messages correspond to a first user and a second user, and the conversation message flow includes a plurality of disordered conversation messages in which the messages of different customer service sessions associated with the object to be queried are interspersed; The object to be queried is a merchant on an e-commerce platform, and the second user is the customer service staff of the merchant; According to the conversation message flow, determine at least one customer service session and a plurality of conversation messages corresponding to the customer service session; According to the plurality of conversation messages corresponding to the customer service session, determine the problem scenario information of the customer service session; According to the problem scenario information of the customer service session, determine the problem scenario distribution information of the object to be queried; According to the problem scenario distribution information of the object to be queried, determine the second user scheduling information.

2. The method according to claim 1, wherein The step of determining at least one customer service session and a plurality of conversation messages corresponding to the customer service session according to the conversation message flow includes: Obtain the first user information, second user information, the sending date information and sending time information of the conversation message corresponding to the conversation message; According to the first user information, second user information and sending date information, determine the at least one customer service session; According to the first user information, second user information, sending date information and sending time information, obtain a plurality of conversation messages corresponding to the customer service session.

3. The method according to claim 1, wherein The step of determining at least one customer service session and a plurality of conversation messages corresponding to the customer service session according to the conversation message flow includes: Obtain the first user information, second user information, the sending date information and sending time information of the conversation message corresponding to the conversation message; According to the first user information, second user information and sending date information, obtain the session greeting and session closing; According to the session greeting and session closing, determine the at least one customer service session; According to the first user information, second user information, sending date information and sending time information, obtain a plurality of conversation messages corresponding to the customer service session.

4. The method according to claim 1, wherein The step of determining the problem scenario information of the customer service session according to the plurality of conversation messages corresponding to the customer service session includes: According to the plurality of conversation messages corresponding to the customer service session, obtain the first user conversation message; Determine the problem scenario information of the first user conversation message; According to the problem scenario information of the first user conversation message, determine the problem scenario information of the customer service session.

5. The method according to claim 4, characterized in that, The step of obtaining the first user conversation message according to the plurality of conversation messages corresponding to the customer service session includes: Obtain the sender identity information of the conversation message; Use the conversation message with the sender identity information being the first user as the first user conversation message.

6. The method according to claim 4, wherein The step of determining the problem scenario information of the first user conversation message includes: Determine the problem scenario information of the first user conversation message through a problem category prediction model.

7. The method according to claim 1, wherein The step of determining the problem scenario distribution information of the object to be queried according to the problem scenario information of the customer service session includes: According to the problem scenario information of the customer service session, obtain the customer service consultation user volume and / or customer service session volume of the object to be queried in at least one time period and in different problem scenarios.

8. The method according to claim 1, wherein It further includes at least one of the following steps: Obtain the total number of customer service consulting users and / or the total number of customer service sessions of the object to be queried in at least one time period and in multiple problem scenarios according to the problem scenario information of the customer service session; Obtain the accuracy of the problem scenario information of the customer service session, and obtain the ranking information of multiple customer service sessions in the same problem scenario within the same time period according to the accuracy.

9. The method according to claim 1, characterized in that, Execute the method at a preset time interval.

10. The method according to claim 1, characterized in that, It further includes at least one of the following steps: Filter out system messages in the conversation message stream; Filter out duplicate messages in the conversation message stream; If it is determined that there are no unreceived conversation messages of the same customer service session before the currently received conversation message, determine the session end time information of the corresponding customer service session according to the sending time information of the currently received conversation message; Filter out invalid customer service sessions in the at least one customer service session.

11. A customer service information processing method, characterized in that It includes: Send a request to obtain problem scenario distribution information; Receive the problem scenario distribution information of the object to be queried; Display the problem scenario distribution information of the object to be queried; Among them, the problem scenario distribution information of the object to be queried is determined in the following manner: Obtain the customer service conversation message stream associated with the object to be queried; according to the conversation message stream, determine at least one customer service session and multiple conversation messages corresponding to the customer service session, the conversation messages corresponding to the first user information and the second user information, and the conversation message stream includes multiple disordered conversation messages in which the messages of different customer service sessions associated with the object to be queried are interspersed; The object to be queried is a merchant on an e-commerce platform, and the second user is the customer service staff of the merchant; according to multiple conversation messages corresponding to the customer service session, determine the problem scenario information of the customer service session; in response to the request, obtain the problem scenario distribution information of the object to be queried according to the problem scenario information of the customer service session of the object to be queried; determine the second user scheduling information according to the problem scenario distribution information of the object to be queried.

12. The method according to claim 11, wherein The problem scenario distribution information includes: the number of customer service consulting users and / or the number of customer service sessions of the object to be queried in at least one time period and in the problem scenario.

13. An electronic device, characterized in that, It includes: A processor; And A memory for storing a program for implementing the customer service information processing method according to any one of claims 1-12, and the terminal is powered on and runs the program of this method through the processor.

Images