A bank customer service interaction information control method, device, system and storage medium

Abstract

The application provides a bank customer service interaction information control method, device and system and a storage medium, and relates to the technical field of financial science and technology services. The bank customer service interaction information control method comprises the following steps: receiving IM information sent by a seat end; identifying the IM information through a message forwarding center, and judging whether the IM information contains a blacklist field; if yes, calling back the IM information to the seat end, and sending a withdrawal prompt. The application realizes real-time monitoring and withdrawal of non-compliant messages by means of a customer service system, all data generation and storage are completed through a system background service, and the application does not rely on manual identification and withdrawal of messages, thereby improving monitoring and statistical efficiency, avoiding the defects of needing to invest a lot of manpower for manual monitoring and withdrawal and occupying a large number of server resources for storing historical chat records.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of financial technology services, and more particularly to a bank customer service interaction information control method, device, system and storage medium. BACKGROUND

[0002] Bank customer service refers to the work of a staff member who provides bank counter service in a financial institution, patiently answers customers' questions, accepts customers' inquiries, enhances customers' service experience, and recommends and introduces existing financial products of the bank to customers. IM (Instant Messaging) is a channel for outputting customer service capabilities, and the language of the staff member needs to comply with the bank's specifications.

[0003] During the actual business processing of a financial institution, the bank customer service staff member on the staff member side may output non-compliant words, including inducement language, sensitive and non-compliant words, impolite words, and internal confidential information. Due to subjective or objective reasons, non-compliant messages often cannot be monitored and withdrawn in real time. For example, the staff member on the staff member side sends messages containing derogatory or insulting content to customers due to emotional fluctuations. These behaviors may leave a negative impression on customers, and may even lead to the loss of bank customers.

[0004] Currently, the conventional IM customer service system in the industry focuses more on the efficiency of instant communication, but lacks evaluation and review of the service attitude and professional quality of the staff member. Under the existing IM customer service system, the method for the customer service supervisor to check the performance of the staff member is usually to manually review the IM chat history of the staff member and each customer, which is complicated and not conducive to statistics. Not only a lot of manpower needs to be invested, but also a large amount of server resources needs to be occupied to store historical chat records. SUMMARY

[0005] Therefore, in view of the above technical problems, the present application provides a bank customer service interaction information control method which can be used in the field of financial technology or other related fields, comprising:

[0006] receiving IM information sent by the staff member side;

[0007] identifying the IM information through a message forwarding center to determine whether the IM information contains a black list field;

[0008] if so, the IM information is called back to the staff member side, and a withdrawal prompt corresponding to the IM information is sent.

[0009] Preferably, the identification of the IM information through the message forwarding center to determine whether the IM information contains a black list field comprises:

[0010] establish a message sending long connection corresponding to the agent end, and listen to IM information accessed by the message sending long connection through the message forwarding center, and confirm whether the blacklist field is contained in the IM information through the listening.

[0011] Preferably, the establishment of the message sending long connection corresponding to the agent end, and the listening to the IM information accessed by the message sending long connection through the message forwarding center, and the confirmation of whether the blacklist field is contained in the IM information through the listening, comprises:

[0012] determining whether the IM information includes an image file;

[0013] if the IM information does not include an image file, clustering the IM information based on the message sending long connection and using a trained field recognition model;

[0014] obtaining a characteristic field set according to the clustering result, and matching a meaning label according to the characteristic field set;

[0015] extracting a target label in the meaning label, and comparing a field in a characteristic field set corresponding to the target label with a blacklist field in a database, so as to determine whether the characteristic field set contains a same field as the blacklist field.

[0016] Preferably, after the determination of whether the IM information includes an image file, the method further comprises:

[0017] if the IM information includes an image file, determining whether the IM information is a dynamic image;

[0018] if yes, extracting a unit frame of the dynamic image;

[0019] detecting each unit frame using a direction detection model to determine a text direction of the unit frame;

[0020] correcting the unit frame according to the text direction to obtain an aligned frame;

[0021] extracting a text region in each aligned frame, and composing a text image set frame in a time stamp order of the unit frame;

[0022] recognizing the text image set frame using a trained text recognition model to obtain a characteristic field;

[0023] comparing the characteristic field with a blacklist field in a database, so as to determine whether the characteristic field set contains a same field as the blacklist field.

[0024] Preferably, the sending of the withdrawal prompt corresponding to the IM information comprises:

[0025] Obtaining a blacklist category label of the blacklist field corresponding to the IM information;

[0026] Generating the withdrawal prompt corresponding to the IM information according to the category label.

[0027] Preferably, after the generating of the withdrawal prompt corresponding to the IM information according to the category label, the method further comprises:

[0028] Recording a callback times of the callback of the IM information to the agent end;

[0029] Judging whether the callback times reaches a preset threshold value;

[0030] If yes, obtaining an agent ID of the agent end, and generating a warning information corresponding to the agent ID, so as to send the warning information to a server end.

[0031] Preferably, after the generating of the withdrawal prompt corresponding to the IM information according to the category label, the method further comprises:

[0032] Constructing an information statistical table corresponding to the agent end; the information statistical table comprises the callback times and the blacklist category label of the agent end;

[0033] Constructing performance portrait information corresponding to the agent end according to the information statistical table.

[0034] In addition, to solve the above problems, the application further provides a bank customer service interactive information control device, comprising:

[0035] A receiving module, configured to receive an IM information sent by an agent end;

[0036] A judging module, configured to identify the IM information through a message forwarding center, and judge whether the IM information contains a blacklist field;

[0037] A callback module, configured to, when the IM information contains the blacklist field, callback the IM information to the agent end, and send a withdrawal prompt corresponding to the IM information.

[0038] In addition, to solve the above problems, the application further provides a bank customer service interactive information control system, comprising a memory and a processor; the memory stores a bank customer service interactive information control program; and the processor runs the bank customer service interactive information control program to enable the bank customer service interactive information control system to execute the bank customer service interactive information control method as described above.

[0039] In addition, to solve the above problems, the application further provides a computer readable storage medium, wherein the computer readable storage medium stores a bank customer service interactive information control program, and the bank customer service interactive information control program is executed by a processor to realize the bank customer service interactive information control method.

[0040] The application provides a bank customer service interactive information control method, device, system and storage medium, wherein the method comprises the following steps: receiving IM information sent by a client end; identifying the IM information through a message forwarding center, and judging whether the IM information contains a blacklist field; if yes, calling back the IM information to the client end, and sending a withdrawal prompt corresponding to the IM information. The application acquires IM information of the client end through a service end platform of a customer service system in real time, identifies the IM information to judge whether the IM information contains a blacklist field, calls back the information and sends a withdrawal prompt if the blacklist field exists, thereby realizing real-time monitoring and withdrawal of non-compliant messages by means of the customer service system, improving monitoring and statistical efficiency, avoiding the defects of occupying a large number of server resources to store historical chat records, and avoiding the need to invest a lot of manpower to manually monitor and withdraw messages. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1 A structural schematic diagram of a hardware running environment related to the bank customer service interactive information control method embodiment of the application;

[0042] Figure 2 A flowchart of the first embodiment of the bank customer service interactive information control method of the application;

[0043] Figure 3 A flowchart of the step S200 in the second embodiment of the bank customer service interactive information control method of the application;

[0044] Figure 4 A flowchart of the step S210 in the second embodiment of the bank customer service interactive information control method of the application (the judgment result is no);

[0045] Figure 5 A flowchart of the step S210 in the second embodiment of the bank customer service interactive information control method of the application (the judgment result is yes);

[0046] Figure 6 A flowchart of the step S300 in the third embodiment of the bank customer service interactive information control method of the application;

[0047] Figure 7 A flowchart of the step of recording the callback times after the step 320 in the third embodiment of the bank customer service interactive information control method of the application;

[0048] Figure 8 This is a schematic diagram of the process for constructing performance profile information after step 320 in the third embodiment of the bank customer service interaction information control method of the present invention.

[0049] Figure 9 This is a schematic diagram of the module connections for the bank customer service interactive information control device of the present invention.

[0050] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0051] The embodiments of the present invention are described in detail below, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout.

[0052] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0053] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0054] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0055] like Figure 1 The diagram shown is a structural schematic of the hardware operating environment of the terminal involved in an embodiment of the present invention.

[0056] The bank customer service interaction information control system of the embodiment of the present application can be a PC, a smart phone, a tablet computer, a portable computer or a mobile terminal device. The bank customer service interaction information control system can include a processor 1001, for example, a CPU, a network interface 1004, a user interface 1003, a memory 1005 and a communication bus 1002. The communication bus 1002 is used to realize the connection and communication between the components. The user interface 1003 can include a display screen, an input unit such as a keyboard, a remote controller, and the optional user interface 1003 can further include a standard wired interface, a wireless interface. The network interface 1004 can optionally include a standard wired interface, a wireless interface (such as a WI-FI interface). The memory 1005 can be a high-speed RAM memory or a stable memory such as a disk memory. The memory 1005 can optionally be a storage device independent of the aforementioned processor 1001. Optionally, the bank customer service interaction information control system can further include an RF (Radio Frequency, radio frequency) circuit, an audio circuit, a WiFi module and the like. In addition, the bank customer service interaction information control system can also be configured with a gyroscope, a barometer, a hygrometer, a thermometer, an infrared sensor and other sensors, which will not be described here.

[0057] Those skilled in the art can understand that Figure 1 The bank customer service interaction information control system shown in the above embodiment does not constitute a limitation thereon, and can include more or fewer components than shown, or combine certain components, or different component arrangements. For example, Figure 1 As shown, the memory 1005 as a computer readable storage medium can include an operating system, a data interface control program, a network connection program and a bank customer service interaction information control program.

[0058] In summary, the present application can realize real-time monitoring and withdrawal of non-compliant messages by the customer service system by acquiring IM information of the agent end in real time through the service end platform of the customer service system, identifying the IM information to determine whether it contains a blacklist field, and sending a withdrawal prompt if the information is called back. All data generation and storage are completed through the system background service, and do not rely on manual identification and withdrawal of messages, thereby improving the monitoring and statistical efficiency, avoiding the need to invest a lot of manpower for manual monitoring and withdrawal, and avoiding the defect of occupying a large amount of server resources to store historical chat records.

[0059] Embodiment 1

[0060] With reference to Figure 2 The first embodiment of the present application provides a bank customer service interaction information control method, comprising:

[0061] Step S100, receiving the IM information sent by the agent end;

[0062] The bank customer service interaction information control method provided in the embodiment can be applied between the agent end, the service end and the client end.

[0063] The agent end is an agent, i.e., a customer service personnel, who performs manual feedback on the information of the client end in real time or at a fixed time and sends IM information (Instant Messaging, instant messaging information). The service end receives the IM information of the agent end and processes the IM information.

[0064] In step S200, the message forwarding center identifies the IM information to determine whether the IM information contains a blacklist field.

[0065] The service end is provided with a message forwarding center for processing and transmission of information. After the service end obtains the IM information, the message forwarding center identifies the obtained IM information to monitor specific fields in the IM information.

[0066] The IM information is identified to determine the field information in the IM information and determine whether the field information contains a specific blacklist field.

[0067] The blacklist field can be preset information stored in a database.

[0068] Before the step S100, the method can further include:

[0069] The service end constructs a field database, and the field database includes a blacklist field.

[0070] The field database is a pre-set database that includes a blacklist field.

[0071] On the platform of the service end, a platform administrator adds and modifies blacklist words through a blacklist word editing interface to form a list, and each blacklist field is provided with a corresponding label.

[0072] During the identification, a neural network processing model can be used for text recognition to determine whether the sent IM information contains a specific blacklist field.

[0073] In step S300, if yes, the IM information is called back to the agent end, and a withdrawal prompt corresponding to the IM information is sent.

[0074] If no, the IM information is sent to the client end.

[0075] The above, in the judgment, if there is a blacklist field, the IM information is intercepted by the message forwarding center of the server, and is called back to the agent end, in addition, the corresponding withdrawal prompt is sent, for example: "message hits ABC blacklist, has been withdrawn", so as to remind the agent end that the information is illegal.

[0076] The embodiment acquires the IM information of the agent end through the service end platform of the customer service system in real time, and identifies the IM information to judge whether the blacklist field is contained therein. If the blacklist field exists, the information is called back and the withdrawal prompt is sent, so that the non-compliant message can be monitored and withdrawn in real time by relying on the customer service system. All data generation and storage are completed through the system background service, and do not depend on manual identification and withdrawal of messages, so as to improve the monitoring and statistical efficiency, avoid the defect that a lot of manpower is needed for manual monitoring and withdrawal, and a large amount of server resources are occupied to store historical chat records.

[0077] Embodiment 2:

[0078] Referring to Figure 3 The second embodiment of the present application provides a bank customer service interaction information control method based on the above embodiment 1. In step S200, the message forwarding center identifies the IM information to judge whether the blacklist field is contained in the IM information, including:

[0079] In step S210, a message sending long connection corresponding to the agent end is established, and the message forwarding center listens to the IM information accessed by the message sending long connection to confirm whether the blacklist field is contained in the IM information.

[0080] The above, when the agent end interacts with the client through the message forwarding center, that is, the call is over, the conversation is in progress, first, the message sending long connection is established, so as to build the relationship between the agent end and the message forwarding center, and the message forwarding center and the client.

[0081] The message forwarding center listens to the sent IM message, so as to monitor and filter the information in real time.

[0082] The above, the blacklist field can be a field pre-stored in the database, including for example, inducement speech, non-standard vocabulary, uncivilized words and the like. Through the listening process, the content of the sent IM information is confirmed, so as to judge whether there is a field matched with the blacklist field from the content, and the real-time monitoring of the agent end is realized.

[0083] Further, referring to Figure 4The step S210 establishes a message sending long connection corresponding to the agent end, and listens to IM information accessed by the message sending long connection through the message forwarding center, and confirms whether the blacklist field is contained in the IM information through the listening, and the step S210 comprises the following steps.

[0084] The step S211 judges whether the IM information comprises an image file.

[0085] When the agent end communicates with the client end, real-time information interaction is needed, which can be realized by including text information, static image containing text and dynamic image containing text. First, it is needed to judge whether the IM information comprises an image file, so as to respectively extract and identify the information, and further improve the identification efficiency after classification.

[0086] The step S212, if the IM information does not comprise an image file, clusters the IM information based on the message sending long connection and by using a trained field identification model.

[0087] If the image file is not contained, it is judged that the current interaction mode is pure text information, and the field identification model is used for clustering in the message sending long connection.

[0088] The step S213 obtains a characteristic field set according to the clustering result, and matches a meaning label according to the characteristic field set.

[0089] The above clustering is a technology for finding the internal structure of data. Clustering organizes all data instances into similar groups, and these similar groups are called clusters. Data instances in the same cluster are the same as each other, and instances in different clusters are different from each other.

[0090] In the existing method, all IM information is matched with the preset field in the database, so as to determine whether the information completely matches the preset field. In the matching process, there are a large number of irrelevant and meaningless contents in actual application, such as mood words, exclamation words, greetings, bank business routine words, routine opening words and routine closing words, etc. These information leads to complete screening and identification in the identification process, thereby increasing a large amount of system resources.

[0091] In this embodiment, in order to further improve the efficiency of character recognition, the parts of the fields which are irrelevant to the actual application, have no actual meaning, have no recognition meaning, etc. are distinguished from the fields which need to be recognized, and are classified by using a clustering algorithm, so as to determine different characteristic field sets, and each field set is assigned a corresponding meaning label. For example, characteristic field set A, field label: conventional opening speech; characteristic field set B, bank business conventional set speech; characteristic field set C, target recognition language.

[0092] By clustering all the languages which need to be recognized, the characteristic field set with the meaning label is obtained, so as to filter out the characteristic field set which needs to be recognized for further matching, greatly improving the recognition efficiency, reducing the waste of system resources by the corpus fields which have no actual meaning and no recognition meaning, and improving the recognition efficiency.

[0093] Step S214, extracting the target label in the meaning label, and comparing the fields in the characteristic field set corresponding to the target label with the black list fields in the database, so as to determine whether the characteristic field set contains the same fields as the black list fields.

[0094] The above-mentioned meaning label includes the target label and other labels except the target label. For example, conventional opening speech, bank business conventional set speech, greeting, exclamation, conventional ending speech, etc. are other labels except the target label. The target recognition is used for the target label corresponding to the field content which needs to be further matched and filtered.

[0095] After extracting the target label, the fields in the characteristic field set corresponding to each target label are obtained, and are matched with the black list fields in the database respectively. By matching the fields, it is determined whether there are the same fields as the black list fields in the database. If there are the same fields, it is determined that message recall is needed.

[0096] Further, referring to Figure 5 , the step S211 of judging whether the IM information includes an image file further includes:

[0097] Step S215, if the IM information includes an image file, judging whether the IM information is a dynamic image;

[0098] In the real-time information communication between the agent end and the client end, in addition to the text communication, the communication through the image file is also possible. In order to further monitor the IM information sent by the agent end and avoid the sending of irregular information due to the special fields in the image file, in this embodiment, the field recognition is further performed after the classification of the image file.

[0099] The image file includes a dynamic image and a static image. The dynamic image can be, for example, a GIF image or a short video image. The static image can be in JPG, PNG, BMP format, and the like.

[0100] In step S216, if yes, a unit frame of the dynamic image is extracted.

[0101] If the image file is a dynamic image, the file of the dynamic image is split into a plurality of unit frames based on a preset time interval, and can be time-stamped and sorted to form a sequence of unit frames.

[0102] In step S217, a direction detection model is used to detect each unit frame to determine the text direction of the unit frame, and the unit frame is corrected according to the text direction to obtain an aligned frame.

[0103] As described above, in the image file, the dynamic image can appear in the form of a video or a chat expression, and the text direction therein is not necessarily completely in an identifiable vertical direction, for example, it can be 30° right-tilted or 60° left-tilted, or even in a reversed direction. The text with an inclined angle in the dynamic image can interfere with further text recognition.

[0104] In this embodiment, for each unit frame of the dynamic image, the direction is adjusted, the text content in each unit frame is detected by a direction detection model to determine the direction of the text content, and the inclined correction is performed based on the detected text direction, so that the text content with an inclined angle in the dynamic image is adjusted to a further recognizable vertical direction by inclined correction as an aligned frame.

[0105] In step S218, a text region in each aligned frame is extracted, and a text image collection frame is formed in the time stamp order of the unit frame. A trained text recognition model is used to recognize the text image collection frame to obtain a feature field.

[0106] In step S219, the feature field is compared with a blacklist field in a database to determine whether the feature field set contains the same field as the blacklist field.

[0107] The alignment frame is the unit frame adjusted by the tilt correction. The alignment frame formed by the sequence of time stamps is the text image collection frame, which has a time sequence. For example, in a dynamic GIF image as an expression, two fields appearing in a flash are "first information" and "second information", and the interval time between the two information is 0.2 seconds. If the preset time for intercepting the unit frame is 0.2 seconds, two unit frames are obtained, which are the unit frame containing "first information" (time stamp 0 seconds) and the unit frame containing "second information" (0.2 seconds). After the tilt correction, the two unit frames become two alignment frames. The above two alignment frames form a complete text image collection frame according to the time stamp sequence, and each alignment frame is included in the frame.

[0108] The trained text recognition model is used to recognize the content in the text image collection frame, and the feature field in the text image collection frame is obtained. The feature field realizes the recognition process of dynamic image-unit frame-alignment frame-text image collection frame-feature field. Further, the matching between the feature field and the blacklist field in the database is determined to determine whether the matching field in the blacklist field is included.

[0109] In addition, if it is determined in step that the IM information is a dynamic image, the result is:

[0110] If not, the image corresponding to the IM information is recognized by using the trained text recognition model to recognize the text image collection frame, and the feature field is obtained. The feature field is compared with the blacklist field in the database to determine whether the feature field set contains the same field as the blacklist field.

[0111] In the embodiment, in order to avoid the defect that the agent end and the client end communicate through the image file containing text information, avoid the monitoring of the text recognition model, and cannot achieve the monitoring effect, the image file is distinguished into dynamic image and static image, the recognition process of dynamic image-unit frame-alignment frame-text image collection frame-feature field is realized, and the finally obtained field is matched to determine the blacklist field therein, so that the monitoring text range is greatly improved, and the dynamic image and the expression package file also realize the real-time monitoring of whether there is illegal content therein.

[0112] Embodiment 3

[0113] Referring to Figure 6 The third embodiment of the present application provides a bank customer service interaction information control method, based on the above embodiment 1, the step S300 of sending a withdrawal prompt corresponding to the IM information, comprising:

[0114] Step S310, obtaining the blacklist category label of the blacklist field corresponding to the IM information;

[0115] As described above, in the database, each blacklist field corresponds to a blacklist category label.

[0116] In the database, the blacklist category label is used to distinguish the categories of different fields, for example, the following correspondence can be used:

[0117] Table 1, correspondence table of blacklist field and blacklist category label

[0118]

[0119] Step S320, generating a withdrawal prompt corresponding to the IM information according to the category label.

[0120] As described above, the message forwarding center calls the withdrawal interface to generate a withdrawal prompt, wherein the withdrawal prompt includes the triggered category label of the IM information sent by the agent end and the corresponding withdrawal prompt, for example: "Field J" sent by you is a violation of sales inducement language, hits the blacklist, and has been withdrawn. Thus, the customer service personnel of the agent end is reminded according to the withdrawal prompt.

[0121] Further, referring to Figure 7 , the step S320 of generating a withdrawal prompt corresponding to the IM information according to the category label further includes:

[0122] Step S330, recording the callback times of the IM information callback to the agent end;

[0123] As described above, if the number of times of the appearance of the blacklist field of the agent end exceeds a certain range, the management personnel corresponding to the agent end needs to be notified to remind the management personnel of the current situation of the agent end.

[0124] As described above, the callback times are the total number of times of the message callback processing and the sending of the withdrawal prompt by the message forwarding center after the appearance of the blacklist field of the agent end.

[0125] Step S340, judging whether the callback times reach a preset threshold;

[0126] Step S350, if yes, obtaining the agent ID of the agent end and generating a warning information corresponding to the agent ID, so as to send the warning information to the server.

[0127] If the callback times reach the preset threshold, a warning needs to be made through the server. For example, the preset threshold is 5; when the callback times of the agent end = 5, the agent ID of the agent end is obtained, which is 123456789, and a warning information is generated: "current agent ID: 123456789, reach callback times, please know". And the above warning information is sent to the server, that is, the warning information can be sent to the supervisor through the mail, and the supervisor can intervene in time, for example, connect the IM call, or temporarily withdraw the agent permission and the like.

[0128] If the callback times do not reach the preset threshold, the step S330 is returned, and the counting is continued.

[0129] Further, with reference to Figure 8 , the step S320, after generating the withdrawal prompt corresponding to the IM information according to the category label, further comprises:

[0130] Step S360, constructing an information statistical table corresponding to the agent end; the information statistical table includes the callback times and the blacklist category label of the agent end;

[0131] Step S370, constructing the performance portrait information corresponding to the agent end according to the information statistical table.

[0132] The information statistical table can include the related information of the agent end corresponding to the agent ID, including but not limited to the agent ID, daily callback times, triggered IM information, blacklist category label and the like, and the performance portrait information can be obtained through the above information, so that the work situation of the customer service personnel can be mastered.

[0133] For example, the system generates a statistical report regularly, tracks the frequency and type of non-compliant messages triggered by each agent. From the compliance professional accomplishment, service attitude, reply timeliness, reception person times and other dimensions, the performance portrait is generated.

[0134] In addition, with reference to Figure 9 , the embodiment also provides a bank customer service interaction information control device, comprising:

[0135] The receiving module 10 is used for receiving the IM information sent by the agent end;

[0136] The judging module 20 is used for identifying the IM information through the message forwarding center, and judging whether the blacklist field is contained in the IM information;

[0137] The callback module 30 is used for, when the blacklist field is contained in the IM information, calling back the IM information to the agent end, and sending the withdrawal prompt corresponding to the IM information.

[0138] Further, the embodiment also provides a bank customer service interaction information control system, comprising a memory and a processor, the memory stores a bank customer service interaction information control program, and the processor runs the bank customer service interaction information control program to enable the bank customer service interaction information control system to execute the bank customer service interaction information control method as described above.

[0139] Further, the embodiment also provides a computer readable storage medium, which stores a bank customer service interaction information control program, and the bank customer service interaction information control program is executed by a processor to implement the bank customer service interaction information control method as described above.

[0140] In summary, the application acquires IM information of the agent end through the service end platform of the customer service system in real time, and identifies the IM information to determine whether the IM information contains a blacklist field, and if so, the information is called back and a withdrawal prompt is sent, so that the non-compliant message can be monitored and withdrawn in real time by relying on the customer service system, all data generation and storage are completed through the system background service, and the monitoring and statistical efficiency is improved, and the defect of occupying a large number of server resources to store historical chat records is avoided.

[0141] The above embodiment numbers of the application are only for description, and do not represent the advantages and disadvantages of the embodiments.

[0142] Through the above description of the embodiments, those skilled in the art can clearly understand that the above-mentioned embodiment method can be realized by means of software and the necessary general hardware platform, of course, it can also be realized by hardware, but in many cases the former is a better embodiment. Based on such understanding, the technical solutions of the application can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a plurality of instructions for making a terminal device (which can be a mobile phone, computer, server, or network device, etc.) execute the method described in each embodiment of the application. The above is only the preferred embodiment of the application, and does not limit the patent scope of the application, and any equivalent structure or equivalent flow transformation made by using the content of the specification and drawings, or directly or indirectly applied to other related technical fields, are also included in the patent protection scope of the application.

Claims

1. A method for controlling interactive information in bank customer service, characterized in that, include: Receive IM messages sent by the agent; Determine whether the IM information includes image files; If the IM information does not include image files, the trained field recognition model is used to cluster the IM information; a set of feature fields is obtained based on the clustering results, and meaning tags are matched based on the set of feature fields; target tags are extracted from the meaning tags, and the fields in the set of feature fields corresponding to the target tags are compared with blacklist fields in the database to determine whether the set of feature fields contains the same fields as the blacklist fields. If so, the IM information is sent back to the agent, and the blacklist category tag of the blacklist field corresponding to the IM information is obtained; a recall prompt corresponding to the IM information is generated based on the category tag.

2. The bank customer service interaction information control method as described in claim 1, characterized in that, After determining whether the IM information includes an image file, the process further includes: If the IM information includes an image file, then determine whether the IM information is a dynamic image; If so, then extract the unit frame of the dynamic image; Each unit frame is detected using an orientation detection model to determine the text orientation of the unit frame; the unit frame is then tilted and corrected based on the text orientation to obtain an aligned frame. Extract the text region from each aligned frame and assemble it into a text image set frame according to the timestamp order of the unit frames; The trained text recognition model is used to recognize the text image set frames to obtain feature fields; the feature fields are compared with blacklist fields in the database to determine whether the feature field set contains fields that are the same as the blacklist fields.

3. The bank customer service interaction information control method as described in claim 1, characterized in that, After generating the recall notification corresponding to the IM information based on the category label, the method further includes: Record the number of times the IM information is returned to the agent's terminal; Determine whether the number of callbacks has reached a preset threshold; If so, the agent ID of the agent terminal is obtained, and a warning message corresponding to the agent ID is generated so that the warning message can be sent to the server.

4. The bank customer service interaction information control method as described in claim 3, characterized in that, After generating the recall notification corresponding to the IM information based on the category label, the method further includes: Construct an information statistics table corresponding to the agent terminal; the information statistics table includes the number of callbacks and the blacklist category label of the agent terminal; Based on the information statistics table, construct the performance profile information corresponding to the agent terminal.

5. A bank customer service interactive information control device, characterized in that, include: The receiving module is used to receive IM information sent by the agent. The determination module is used to determine whether the IM information includes an image file; If the IM information does not include image files, the trained field recognition model is used to cluster the IM information; a set of feature fields is obtained based on the clustering results, and meaning tags are matched based on the set of feature fields; target tags are extracted from the meaning tags, and the fields in the set of feature fields corresponding to the target tags are compared with blacklist fields in the database to determine whether the set of feature fields contains the same fields as the blacklist fields. The callback module is used to send the IM information back to the agent when the IM information contains a blacklist field, and to obtain the blacklist category tag of the blacklist field corresponding to the IM information; and to generate a recall prompt corresponding to the IM information based on the category tag.

6. A bank customer service interactive information control system, characterized in that, The system includes a memory and a processor. The memory stores a bank customer service interaction information control program, and the processor runs the bank customer service interaction information control program to enable the bank customer service interaction information control system to perform the bank customer service interaction information control method as described in any one of claims 1-4.

7. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a bank customer service interaction information control program, which, when executed by a processor, implements the bank customer service interaction information control method as described in any one of claims 1-4.

Images