Distributed customer service team intelligent management system and method based on multi-dimensional data acquisition

Abstract

The invention discloses a distributed customer service team intelligent management system and method based on multi-dimensional data acquisition, and relates to the technical field of data acquisition and storage management.The method comprises the steps that a double-channel screenshot mechanism is established, and screenshot data acquisition is conducted on a customer service terminal; controlling a dual-channel screenshot mechanism to perform collaborative sampling in a manner of avoiding repeated sampling; capturing an input event flow and an interface state flow of the customer service terminal in real time; generating a monitoring evidence packet which cannot be tampered; a multi-dimensional efficiency evaluation model is constructed, and continuous evaluation of multi-dimensional efficiency evaluation indexes is carried out on each customer service terminal based on the monitoring evidence packet; monitoring the change of the efficiency evaluation index value of the customer service terminal, and identifying value data causing the fluctuation of the efficiency evaluation index value; and according to the value data, intelligent distribution of a storage period is carried out on the data in the monitoring evidence packet, and storage maintenance or destruction operation of the data is automatically executed according to a mapped storage duration strategy.

Description

Technical Field

[0001] The present invention relates to the technical field of data acquisition, storage and management, and in particular to a distributed customer service team intelligent management system and method based on multi-dimensional data acquisition. Background Art

[0002] Current mainstream customer service team management technologies primarily utilize the following two models: timekeeping systems based on manual reporting (e.g., Excel / paper records) and centralized management platforms based on remote desktop monitoring. The former suffers from data lag (typically an average delay > 24 hours) and subjective bias (error rate > 35%), while the latter, due to continuous video streaming, leads to excessive network bandwidth usage (> 2Mbps per terminal) and skyrocketing storage costs (average daily storage volume > 50GB / terminal). Each model has the following disadvantages: The lack of structured collection of behavioral data such as software usage and website visits results in a single data collection dimension, and the use of traditional screen recording solutions generates a large amount of redundant data (effective data percentage). Summary of the Invention

[0003] The purpose of the present invention is to provide a distributed customer service team intelligent management system and method based on multi-dimensional data collection to solve the problems raised in the prior art.

[0004] To achieve the above objectives, the present invention provides the following technical solution: a distributed customer service team intelligent management method based on multi-dimensional data collection, the method comprising: Establish a dual-channel screenshot mechanism to collect screenshot data from customer service terminals; sort out the different distribution characteristics of screenshot data collected by different screenshot mechanisms, and control the dual-channel screenshot mechanism to avoid duplicate sampling and perform collaborative sampling; The Hook technology layer is used to capture the input event stream and interface status stream of the customer service terminal in real time. The screenshot data is combined with the Hook operation log and network message with the corresponding timestamp to generate a blockchain fingerprint to generate an unalterable monitoring evidence package. Construct a multi-dimensional performance evaluation model and conduct continuous evaluation of multi-dimensional performance evaluation indicators for each customer service terminal based on the monitoring evidence package; monitor the changes in the performance evaluation indicator values ​​of the customer service terminal and identify the value data that causes fluctuations in the performance evaluation indicator values; With reference to the value data, the storage period of the data in the monitoring evidence package is intelligently allocated, and the storage maintenance or destruction of the data is automatically performed according to the mapped storage duration strategy.

[0005] Preferably, the dual-channel screenshot mechanism includes executing screenshots at preset time intervals and instant screenshots triggered by preset business events.

[0006] Preferably, the process of sorting out different distribution characteristics of screenshot data collected through different screenshot mechanisms includes: Step S101: extracting the real-time screenshot sequences collected from each service terminal based on the preset business event trigger mechanism, and traversing the event tags corresponding to each real-time screenshot in each real-time screenshot sequence; Step S102: If M consecutive instant screenshots with the same event tag are found in the instant screenshot sequence of a certain service terminal, all corresponding to a certain preset business event, where M ≥ Q, Q is the number threshold, and the interval between the timestamps corresponding to each two adjacent instant screenshots is less than the interval threshold, the screenshot sequence consisting of the M consecutive instant screenshots is used as a characteristic screenshot sequence of the certain preset business event; and all characteristic screenshot sequences corresponding to each preset business event are collected separately; The screenshot trigger mechanism based on preset business events determines whether to take an immediate screenshot of the currently displayed screen image based on whether feature information that will appear when the corresponding preset business event occurs can be extracted from the currently displayed screen image. That is, the screenshot trigger mechanism based on preset business events can only judge whether a preset business event has occurred and the category of the currently occurring preset business event (which can generate a corresponding event tag), but cannot judge whether the collected immediate screenshots belong to the same preset business event. For example, due to the recognition of text features of emotional expressions, it can only be judged that a customer complaint event has occurred. However, the complaint behavior of the customer due to a reason often presents a certain duration. That is, each preset business event has a corresponding duration. The above-mentioned feature screenshot sequence is more likely to be an immediate screenshot collected for a certain type of preset business event that occurred at the customer service terminal because the instant screenshots are arranged continuously and the corresponding timestamps are close. Step S103: From each characteristic screenshot sequence of any preset business event, extract the timestamp T1 corresponding to the first instant screenshot and the timestamp T2 corresponding to the last instant screenshot, extract a duration Te=T2-T1 of any preset business event, and calculate a trigger density value β=M / (T2-T1) of any preset business event to the instant screenshot; the above timestamp T1 is the time when a preset business event first occurs, and the timestamp T2 is the time when the preset business event ends.

[0007] Preferably, the process of controlling the collaborative sampling of the dual-channel screenshot mechanism includes: Step S111: Calculate the average value of the trigger density values ​​calculated based on all characteristic screenshot sequences for any preset business event to obtain the average trigger density value β' of the instant screenshots for any preset business event. When β'>η, where η is the density threshold, assume that any preset business event is a characteristic business event. Calculate the average value of the durations extracted based on all characteristic screenshot sequences for any preset business event to obtain the characteristic duration of any preset business event. According to the above method, the characteristic business events identified are all preset business events that trigger the screenshot mechanism at a high frequency once they begin to appear, and collect more instant screenshots. Step S112: Every time an instant screenshot is obtained due to the triggering of any characteristic business event, the current timestamp t1 and the timestamp t2 of the next scheduled screenshot executed at the preset time interval are extracted, and the characteristic duration Ty of any characteristic business event is extracted at the same time. When t2-t1≦Ty, the timestamp of the next scheduled screenshot is temporarily adjusted to t1+Ty+t0, where t0 is the preset compensation duration.

[0008] If the next scheduled screenshot is executed at timestamp t2, the screenshot data collected and the screenshot data collected based on the preset business event trigger mechanism will most likely be repeated. That is, there is a high probability that a screenshot will be collected immediately at timestamp t2 due to the triggering of any characteristic business event, resulting in redundant storage data.

[0009] Preferably, the preset business events include customer complaints, service timeouts, and sensitive operations; the performance evaluation indicators include activity, resolution rate, error rate, and professional index.

[0010] Preferably, whenever the value of a certain performance evaluation indicator is monitored to rise or fall at a certain customer service terminal, a monitoring evidence package U1 corresponding to the value of the certain performance evaluation indicator before the rise or fall is extracted and evaluated, and a monitoring evidence package U2 corresponding to the value of the certain performance evaluation indicator after the rise or fall is extracted and evaluated. If U2=U1+Uy is satisfied, Uy is a newly added data set, and the data in Uy is judged to be valuable data that causes fluctuations in the value of the performance evaluation indicator, and the storage period of the data in Uy is set to be greater than the data contained in the monitoring evidence package U1.

[0011] In order to better implement the above method, an intelligent management system is also proposed, which includes a data acquisition management module, a multi-dimensional data packaging management module, a value data extraction management module, and a data storage cycle intelligent allocation module; The data collection management module is used to establish a dual-channel screenshot mechanism to collect screenshot data from customer service terminals; sort out the different distribution characteristics of screenshot data collected by different screenshot mechanisms, and control the dual-channel screenshot mechanism to avoid duplicate sampling and perform collaborative sampling; The multi-dimensional data encapsulation management module is used to capture the input event stream and interface status stream of the customer service terminal in real time through the Hook technology layer; the screenshot data is fingerprinted with the Hook operation log and network message with the corresponding timestamp to generate an unalterable monitoring evidence package; The value data extraction and management module is used to build a multi-dimensional performance evaluation model and conduct continuous evaluation of multi-dimensional performance evaluation indicators for each customer service terminal based on the monitoring evidence package; monitor the changes in the performance evaluation indicator values ​​of the customer service terminal and identify the value data that causes fluctuations in the performance evaluation indicator values; The data storage cycle intelligent allocation module is used to intelligently allocate storage cycles for data in the monitoring evidence package with reference to value data, and automatically perform data storage maintenance or destruction operations according to the mapped storage duration strategy.

[0012] Preferably, the data collection management module includes a timing trigger and a business event trigger to achieve dual-mode screenshot collection.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention constructs a dual-channel screenshot mechanism, which, on the one hand, provides continuous time series data through timed screenshots, which is suitable for recording gradual changes in system / application status, such as memory leaks and CPU load growth; on the other hand, instant screenshots capture key business moments, avoiding the "time blind spot" of timed screenshots, ensuring that key nodes are not missed, and ensuring the comprehensiveness of the entire supervision process, thereby reflecting the advantages of dual-channel screenshots. In the present invention, by designing collaborative control of the dual-channel screenshot mechanism in some scenarios, the present invention can not only achieve efficient collection of dynamic screen information, but also achieve data deduplication optimization to a certain extent, and reduce the storage of redundant data; the present invention also monitors the changes in the performance evaluation index values ​​of the customer service terminal, identifies and extracts valuable data that can cause fluctuations in the performance evaluation index values, so as to have a clear and relatively scientific reference basis for intelligent allocation of data storage cycles; In addition, the present invention adopts a dual-channel screenshot mechanism, which makes it possible to plan the transmission of data, ensure the accuracy and non-tamperability of data transmission, and ensure the security of data. At the same time, while reducing redundant data storage, the data is accurately screened, effectively improving the efficiency of subsequent data retrieval and extraction, and realizing intelligent, comprehensive and diversified management of distributed customer service teams. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Figure 1 This is a flow chart of the distributed customer service team intelligent management system based on multi-dimensional data collection of the present invention; Figure 2 This is a schematic diagram of a first embodiment of a distributed customer service team intelligent management method based on multi-dimensional data collection according to the present invention; Figure 3 This is a schematic diagram of a second embodiment of the distributed customer service team intelligent management method based on multi-dimensional data collection of the present invention. DETAILED DESCRIPTION

[0015] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0016] Example: Figure 1-Figure 3 As shown, the present invention provides a distributed customer service team intelligent management method and system based on multi-dimensional data collection. One implementation of the method is as follows: Step S1: Establish a dual-channel screenshot mechanism to collect screenshot data from the customer service terminal; sort out the different distribution characteristics of the screenshot data collected by different screenshot mechanisms, and control the dual-channel screenshot mechanism to perform collaborative sampling in a manner to avoid repeated sampling; The dual-channel screenshot mechanism includes screenshots taken at preset time intervals and instant screenshots triggered by preset business events; preset business events include customer complaints, service timeouts, and sensitive operations; Timed screenshots provide continuous time series data, suitable for recording gradual changes in system / application status, such as memory leaks and CPU load increases. Instant screenshots capture key business moments, avoiding the "time blind spots" of timed screenshots, ensuring that key nodes are not missed and ensuring the comprehensiveness of the entire supervision process. The process of sorting out the different distribution characteristics of screenshot data collected through different screenshot mechanisms includes: Step S101: extracting the real-time screenshot sequences collected from each service terminal based on the preset business event trigger mechanism, and traversing the event tags corresponding to each real-time screenshot in each real-time screenshot sequence; Step S102: If M consecutive instant screenshots with the same event tag are found in the instant screenshot sequence of a certain service terminal, all corresponding to a certain preset business event, where M ≥ Q, Q is the number threshold, and the interval between the timestamps corresponding to each two adjacent instant screenshots is less than the interval threshold, the screenshot sequence consisting of the M consecutive instant screenshots is used as a characteristic screenshot sequence of the certain preset business event; and all characteristic screenshot sequences corresponding to each preset business event are collected separately; Step S103: From each characteristic screenshot sequence of any preset business event, extract the timestamp T1 corresponding to the first instant screenshot and the timestamp T2 corresponding to the last instant screenshot, extract a duration Te = T2 - T1 of the preset business event, and calculate a trigger density value β = M / (T2 - T1) of the preset business event on the instant screenshots; The present invention provides an embodiment: a quantity threshold Q=10, and 14 consecutive instant screenshots corresponding to the same event mark are extracted from the instant screenshot sequence of the service terminal A, all corresponding to customer complaint events, and the interval length of the timestamps corresponding to each two adjacent instant screenshots is less than the interval threshold. Then, the screenshot sequence composed of the above 14 consecutive instant screenshots is used as a characteristic screenshot sequence of the customer complaint event; if the timestamp T1 corresponding to the first instant screenshot extracted from the above characteristic screenshot sequence of the customer complaint event is 14:20, and the timestamp T2 corresponding to the last instant screenshot is 14:24; then a duration Te=4min of the customer complaint event is extracted; in summary, a trigger density value of the customer complaint event on the instant screenshot is calculated: β=M / (T2-T1)=14 / 4=7 / 2(1 / min).

[0017] The process of controlling the collaborative sampling of the dual-channel screenshot mechanism includes: Step S111: Calculate the average trigger density value β' of any preset business event for instant screenshots based on all characteristic screenshot sequences, and when β'>η, where η is the density threshold, assume that any preset business event is a characteristic business event, calculate the average duration of any preset business event extracted from all characteristic screenshot sequences, and obtain the characteristic duration of any preset business event; Step S112: Every time a real-time screenshot is obtained by triggering any characteristic business event, the current timestamp t1 and the timestamp t2 of the next scheduled screenshot executed at the preset time interval are extracted, and the characteristic duration Ty of any characteristic business event is extracted. When t2-t1≦Ty, the timestamp of the next scheduled screenshot is temporarily adjusted to t1+Ty+t0, where t0 is the preset compensation duration; wherein, if t2 and t1 are to satisfy the time distribution relationship of t2-t1≦Ty, Figure 2 and Figure 3 As shown; Step S2: The input event stream and interface status stream of the customer service terminal are captured in real time through the Hook technology layer; the screenshot data is fingerprinted with the Hook operation log and network message with the corresponding timestamp to generate an unalterable monitoring evidence package; Step S3: Construct a multi-dimensional performance evaluation model and conduct continuous evaluation of multi-dimensional performance evaluation indicators for each customer service terminal based on the monitoring evidence package; monitor changes in the performance evaluation indicator values ​​of the customer service terminal and identify the value data that causes fluctuations in the performance evaluation indicator values; wherein the performance evaluation indicators include activity, resolution rate, error rate, and professional index; Among them, whenever the value of a certain performance evaluation indicator monitored on a certain customer service terminal increases or decreases, the monitoring evidence package U1 corresponding to the value of the performance evaluation indicator before the increase or decrease is extracted and evaluated, and the monitoring evidence package U2 corresponding to the value of the performance evaluation indicator after the increase or decrease is extracted and evaluated. If U2=U1+Uy is satisfied, Uy is a newly added data set, and the data in Uy is judged to be valuable data that causes the performance evaluation indicator value fluctuation. The storage period of the data in Uy is set to be longer than the data contained in the monitoring evidence package U1; By monitoring the changes in the performance evaluation index values ​​of the customer service terminal, the valuable data that can cause fluctuations in the performance evaluation index values ​​are identified and extracted, so as to have a clear and relatively scientific reference basis when intelligently allocating the storage period of the data.

[0018] Step S4: With reference to the value data, intelligently allocate storage periods for the data in the monitoring evidence package, and automatically perform data storage maintenance or destruction operations according to the mapped storage duration strategy.

[0019] In order to better implement the above method, an intelligent management system is also proposed, which includes a data acquisition management module, a multi-dimensional data packaging management module, a value data extraction management module, and a data storage cycle intelligent allocation module; The data collection management module is used to establish a dual-channel screenshot mechanism to collect screenshot data from customer service terminals; sort out the different distribution characteristics of screenshot data collected by different screenshot mechanisms, and control the dual-channel screenshot mechanism to avoid duplicate sampling and perform collaborative sampling; The multi-dimensional data encapsulation management module is used to capture the input event stream and interface status stream of the customer service terminal in real time through the Hook technology layer; the screenshot data is fingerprinted with the Hook operation log and network message with the corresponding timestamp to generate an unalterable monitoring evidence package; The value data extraction and management module is used to build a multi-dimensional performance evaluation model and conduct continuous evaluation of multi-dimensional performance evaluation indicators for each customer service terminal based on the monitoring evidence package; monitor the changes in the performance evaluation indicator values ​​of the customer service terminal and identify the value data that causes fluctuations in the performance evaluation indicator values; The data storage cycle intelligent allocation module is used to intelligently allocate storage cycles for data in the monitoring evidence package with reference to value data, and automatically perform data storage maintenance or destruction operations according to the mapped storage duration strategy.

[0020] Among them, the data collection management module includes timing triggers and business event triggers to realize dual-mode screen snapshot collection.

[0021] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above and that the invention can be embodied in other specific forms without departing from the spirit or essential characteristics of the invention. Therefore, the embodiments should be considered in all respects as illustrative and non-restrictive, and the scope of the invention is defined by the appended claims, not the foregoing description, and all variations within the meaning and range of equivalents of the claims are intended to be included therein. Any reference sign in a claim should not be construed as limiting the claim to which it relates.

Claims

1. A distributed customer service team intelligent management method based on multi-dimensional data collection, characterized by: The method comprises: Establish a dual-channel screenshot mechanism to collect screenshot data from customer service terminals; sort out the different distribution characteristics of screenshot data collected by different screenshot mechanisms, and control the dual-channel screenshot mechanism to avoid duplicate sampling and perform collaborative sampling; The Hook technology layer is used to capture the input event stream and interface status stream of the customer service terminal in real time. The screenshot data is combined with the Hook operation log and network message with the corresponding timestamp to generate a blockchain fingerprint to generate an unalterable monitoring evidence package. Construct a multi-dimensional performance evaluation model to continuously evaluate the multi-dimensional performance evaluation indicators of each customer service terminal based on the monitoring evidence package; monitor the changes in the performance evaluation indicator values ​​of the customer service terminal and identify the value data that causes fluctuations in the performance evaluation indicator values; With reference to the value data, the storage period of the data in the monitoring evidence package is intelligently allocated, and the storage maintenance or destruction of the data is automatically performed according to the mapped storage duration strategy.

2. The distributed customer service team intelligent management method based on multi-dimensional data collection according to claim 1 is characterized by: The dual-channel screenshot mechanism includes executing screenshots at preset time intervals and instant screenshots triggered by preset business events.

3. The distributed customer service team intelligent management method based on multi-dimensional data collection according to claim 2 is characterized by: The process of sorting out the different distribution characteristics of screenshot data collected through different screenshot mechanisms includes: Step S101: extracting the real-time screenshot sequences collected from each service terminal based on the preset business event trigger mechanism, and traversing the event tags corresponding to each real-time screenshot in each real-time screenshot sequence; Step S102: If M consecutive instant screenshots with the same event tag are extracted from the instant screenshot sequence of a certain service terminal, all corresponding to a certain preset business event, where M ≥ Q, Q is the number threshold, and the interval between the timestamps corresponding to each two adjacent instant screenshots is less than the interval threshold, the screenshot sequence consisting of the M consecutive instant screenshots is used as a characteristic screenshot sequence of the certain preset business event; all characteristic screenshot sequences corresponding to each preset business event are collected separately; Step S103: From each characteristic screenshot sequence of any preset business event, extract the timestamp T1 corresponding to the first instant screenshot and the timestamp T2 corresponding to the last instant screenshot, extract a duration Te=T2-T1 of the arbitrary preset business event, and calculate a trigger density value β=M / (T2-T1) of the arbitrary preset business event on the instant screenshot.

4. The distributed customer service team intelligent management method based on multi-dimensional data collection according to claim 2 is characterized in that: The process of controlling the collaborative sampling of the dual-channel screenshot mechanism includes: Step S111: Calculate the average value of the trigger density values ​​calculated based on all characteristic screenshot sequences for the arbitrary preset business event to obtain the average trigger density value β' of the arbitrary preset business event for the instant screenshots. When β'>η, where η is the density threshold, assume that the arbitrary preset business event is a characteristic business event. Calculate the average value of the durations extracted based on all characteristic screenshot sequences for the arbitrary preset business event to obtain the characteristic duration of the arbitrary preset business event. Step S112: Every time an instant screenshot is obtained due to the triggering of any characteristic business event, the current timestamp t1 and the timestamp t2 of the next scheduled screenshot executed at the preset time interval are extracted, and the characteristic duration Ty of the said arbitrary characteristic business event is extracted. When t2-t1≦Ty, the timestamp of the next scheduled screenshot is temporarily adjusted to t1+Ty+t0, where t0 is the preset compensation duration.

5. The distributed customer service team intelligent management method based on multi-dimensional data collection according to claim 2 is characterized in that: The preset business events include customer complaints, service timeouts, and sensitive operations; the performance evaluation indicators include activity, resolution rate, error rate, and professional index.

6. The distributed customer service team intelligent management method based on multi-dimensional data collection according to claim 1 is characterized by: Whenever the value of a certain performance evaluation indicator is monitored to rise or fall on a certain customer service terminal, the monitoring evidence package U1 corresponding to the value of the certain performance evaluation indicator before the rise or fall is extracted and evaluated, and the monitoring evidence package U2 corresponding to the value of the certain performance evaluation indicator after the rise or fall is extracted and evaluated. If U2=U1+Uy is satisfied, Uy is a newly added data set, and the data in Uy is judged to be valuable data that causes fluctuations in the value of the performance evaluation indicator, and the storage period of the data in Uy is set to be greater than the data contained in the monitoring evidence package U1.

7. An intelligent management system for executing the distributed customer service team intelligent management method based on multi-dimensional data collection according to any one of claims 1 to 6, characterized in that: The system includes a data acquisition management module, a value data extraction management module, and a data storage cycle intelligent allocation module; The data collection and management module is used to establish a dual-channel screenshot mechanism to collect screenshot data from the customer service terminal; sort out the different distribution characteristics of the screenshot data collected by different screenshot mechanisms, and control the dual-channel screenshot mechanism to perform collaborative sampling in a manner to avoid repeated sampling; The value data extraction and management module is used to build a multi-dimensional performance evaluation model and conduct continuous evaluation of the multi-dimensional performance evaluation indicators of each customer service terminal based on the monitoring evidence package; monitor the changes in the performance evaluation indicator values ​​of the customer service terminals and identify the value data that causes fluctuations in the performance evaluation indicator values; The data storage period intelligent allocation module is used to intelligently allocate storage periods for data in the monitoring evidence package with reference to the value data, and automatically perform storage maintenance or destruction operations on the data according to the mapped storage duration strategy.

8. The intelligent management system according to claim 7, characterized in that: The intelligent management system also includes a multi-dimensional data packaging management module; The multi-dimensional data encapsulation management module is used to capture the input event stream and interface status stream of the customer service terminal in real time through the Hook technology layer; the screenshot data is fingerprinted with the Hook operation log and network message with the corresponding timestamp to generate an unalterable monitoring evidence package.

9. The intelligent management system according to claim 7, characterized in that: The data acquisition management module includes a timing trigger and a business event trigger to realize dual-mode screen snapshot acquisition.

Images