Method for evaluating working behavior of customer service staff

Abstract

The invention relates to the technical field of big data processing, in particular to a working behavior evaluation method for customer service staff. The method for evaluating the working behavior of the customer service staff comprises the following steps: evaluating the working normalization of the customer service staff; evaluating the emergency guarantee capability of the customer service agent personnel; the method has the beneficial effects that rapid integration and analysis of related information are realized through a technical means, a non-standard behavior personnel list is timely and accurately formed every day, and effective support is provided for operation management; emergency energy efficiency data of customer service personnel are subjected to normalized statistics, effective support is provided for emergency management, and the overall support capacity is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of big data processing, and in particular to a method for evaluating the work behavior of customer service agents. Background Art

[0002] At present, the main work method for the standardization analysis of the work of 95598 customer service agents is to manually export relevant reports from the system and manually analyze and judge about 250,000 records in the data tables of agent absence, shift scheduling, etc. every day. Each analysis takes a long time and is prone to errors. In order to improve the quality and efficiency of work, it is necessary to use technical forces to carry out data analysis of related work, support operation monitoring, and save human resources. Summary of the invention

[0003] The present invention proposes a customer service agent work behavior evaluation method based on 95598 customer service big data, which can achieve an objective evaluation of the behavior of each staff member. The specific technical solution is as follows:

[0004] The customer service agent work behavior evaluation method includes the following process:

[0005] S100: Standardized evaluation of customer service agent work; the specific process is as follows:

[0006] Establish "absentee list for the day", "abnormal list of meal leave selection for the day", "abnormal list of training duration less than 15 minutes for the day", "wrong list of training leave selection for the day", "wrong list of shift leave selection for the day", "list of absence ratio greater than 8% for the day", "list of single break leave greater than 10 minutes for the day", "list of absences greater than 30 times for the day", "list of single absences less than 1 minute for the day", "list of affairs duration greater than 0 for the day", "list of sign-outs greater than 12 times for the day", "list of delayed readiness for the day", "list of extra-long blank for the day", and the initial list is "empty";

[0007] Set up "Number of absences in the month", "Number of abnormal selections for meal leave in the month", "Number of abnormal daily training duration less than 15 minutes in the month", "Number of incorrect selections for training leave in the month", "Number of incorrect selections for shift leave in the month", "Number of times the ratio of leave for rest is greater than 8% in the month", "Number of single leave for rest is greater than 10 minutes in the month", "Number of days with more than 30 absences in a single day in the month", "Number of single absences less than 1 minute in the month", "Number of days with more than 0 daily affairs duration in the month", "Number of days with more than 12 sign-outs in a single day in the month", "Number of days with delayed readiness in a single day in the month", "Number of extra-long blank days in a single day in the month", "Number of emergency times in the month", "Emergency duration in the month", and the initial data is 0;

[0008] The initial value of the customer service agent's daily work standardization deduction value is set to 0;

[0009] S101: Determine whether it is absence. Specific process: If both the first sign-in time and the first ready time are empty, the agent is included in the "absentee list of the day", the "absent times of the month" of the agent increases by 1, and the customer service agent's work standardization deduction value for the day is increased by 10 and then go to S115, otherwise go to S102;

[0010] S102: Determine whether it is absence. Specific process: If the last sign-out time, the last customer active hang-up time, the last agent active hang-up time, and the last transfer satisfaction time are all empty, the agent will be included in the "absentee list of the day", the agent's "absent times of the month" will increase by 1, and the customer service agent's work standardization deduction value for the day will be increased by 10, and then go to S115, otherwise go to S103;

[0011] S103: Determine whether the dining leave selection is abnormal. If the dining time of the customer service agent is greater than 0, the agent is included in the "List of Abnormal Dining Leave Selections of the Day", the agent's "Number of Abnormal Dining Leave Selections of the Month" is increased by 1, and the customer service agent's work standardization deduction value for the day is increased by 1, and then the process goes to S104; otherwise, the process goes directly to S104;

[0012] S104: Determine whether the training time is less than 15 minutes. If the training time of the customer service agent is less than 15 minutes, the agent is included in the "abnormal list of training time less than 15 minutes on the same day", and the "abnormal number of daily training time less than 15 minutes in the same month" of the agent is increased by 1, and the customer service agent's work standardization deduction value on the same day is increased by 1, and then go to S105; otherwise, go directly to S105;

[0013] S105: If the training time of the customer service agent is not from the department and is greater than 0, the agent is included in the "List of Wrong Training Leave Clicks on the Day", the agent's "Number of Wrong Training Leave Clicks on the Month" is increased by 1, and the customer service agent's work standardization deduction value on the day is increased by 1, and then the process goes to S106; otherwise, the process goes directly to S106;

[0014] S106: Determine whether the shift leave selection is wrong; first determine whether the agent is a shift leader, if yes, go directly to S107; otherwise, determine whether the shift duration is greater than 0, if yes, put the agent in the "list of wrong shift leave selections for the day", the agent's "number of wrong shift leave selections for the month" is increased by 1, and the customer service agent's work standardization deduction value for the day is +1 and then go to S107, otherwise go directly to S107;

[0015] S107: Determine whether the daily rest absence ratio is greater than 8%; determine whether the rest time / sign-in time is greater than 8%. If so, the agent is included in the "list of agents with a daily rest absence ratio greater than 8%", the agent's "monthly rest absence ratio greater than 8% times" is increased by 1, and the customer service agent's daily work standardization deduction value is increased by 1 and then go to S108, otherwise go directly to S108;

[0016] S108: Determine whether the single break absence on the day is greater than 10 minutes; if yes, the agent is included in the "list of single break absences greater than 10 minutes on the day", the agent's "number of single break absences greater than 10 minutes in the month" is increased by 1, and the customer service agent's work standardization deduction value on the day is increased by 1 and then go to S109; otherwise, go directly to S109;

[0017] S109: Determine whether the number of absences on the day is greater than 30 times. If yes, the agent is included in the "list of absences greater than 30 times on the day", the number of days with more than 30 absences on a single day in the month for the agent is increased by 1, and the standardization deduction value of the customer service agent's work on the day is increased by 1 and then the process goes to S110. Otherwise, the process goes directly to S110;

[0018] S110: Determine whether there is a single absence of less than 1 minute, then add the agent to the "list of single absences of less than 1 minute on the day", increase the "number of single absences of less than 1 minute on the month" of the agent by 1, and increase the deduction score of the customer service agent's work standardization on the day by 1, then go to S111, otherwise go directly to S111;

[0019] S111: Determine whether there is a transaction duration greater than 0. If yes, add the agent to the "list of transactions of the day greater than 0", increase the number of days of the agent's "single-day transaction duration greater than 0 times in the month" by 1, and increase the deduction score of the customer service agent's work standardization on that day by 1, and then go to S112. Otherwise, go directly to S112;

[0020] S112: Determine whether there is a person who has signed out more than 12 times. If yes, add the agent to the "list of people who have signed out more than 12 times on the same day", increase the number of days for the agent to sign out more than 12 times on the same day by 1 day, and increase the deduction score for the standardization of the customer service agent's work on that day by 1, and then go to S113. Otherwise, go directly to S113.

[0021] S113: When the staff is not on a rest shift, determine whether there is a delay in readiness; specific process: determine whether the first readiness time is more than 2 minutes later than the working time, if yes, put the staff in the "list of delayed readiness for the day", increase the "number of delayed readiness days for a single day in the month" of the staff by 1, and increase the deduction point value of the customer service staff's work standardization on that day by 1 and then go to S114, otherwise go directly to S114;

[0022] S114: When the staff is not on rest shift, determine whether there is an extra-long blank; specific process: determine whether the latest time among the last sign-out time, the last customer active hang-up time, the last agent active hang-up time, and the last transfer satisfaction time is more than 10 minutes earlier than the end of get off work time. If so, the agent will be included in the "Extra-long Blank List of the Day", the "Extra-long Blank Days per Day of the Month" of the agent will be increased by 1 day, and the customer service agent's work standardization deduction value for the day will be increased by 1 and then go to S115, otherwise go directly to S115; return to "Extra-long Blank"

[0023] S115: End the standardized evaluation of customer service agents' work;

[0024] S200: Evaluation of customer service agent emergency support capabilities; the specific process is as follows:

[0025] The initial value of the number of emergency calls for customer service agents in the month is set to 0, and the initial value of the emergency duration is set to 0;

[0026] S201: Determine whether the customer service agent is on break shift, if yes, go to S202, otherwise go to;

[0027] S202: Determine the emergency duration of the day by checking the agent calls, agent softphone applications, and agent softphone detailed online data, and add it to the number of emergency times and emergency duration of the month;

[0028] S203: End the evaluation of the customer service agent's emergency support capability for the day.

[0029] Preferably, the following subsequent process is also included:

[0030] A detailed evaluation list for each customer service agent is formed by day, month, and year. The smaller the deduction score for the customer service agent's work standardization on that day, the better the agent is; the greater the number of emergency responses and the longer the emergency duration, the better the agent is.

[0031] Preferably, the following subsequent process is also included:

[0032] Grey clustering can be divided into grey association clustering and grey possibility function clustering. The former can be used to check whether there are several factors that roughly belong to the same category among many factors, while the latter is mainly used to observe whether the objects belong to different categories set in advance. Grey possibility function clustering can be divided into grey variable weight clustering and grey fixed weight clustering. When the meaning and dimension of clustering indicators are different and the difference in quantity is large, grey fixed weight clustering is a better clustering method.

[0033] This patent adopts the grey possibility function, combined with the indicator weight determined by the extreme entropy method, to divide all agents into three grey categories: excellent agents, general agents, and agents to be improved;

[0034] Suppose there are n clustering objects, m clustering indices and s gray classes; x ij (i=1, 2, ..., n; j=1, 2, ..., m) is the observed value of cluster object i with respect to cluster index j; f j k (·)(j=1, 2, …, m; k=1, 2, …, s) is the possible degree function of index j with respect to gray class k;

[0035]

[0036] w j is the fixed-weight clustering coefficient of object i belonging to k gray classes, where w j is the clustering weight of the indicator and like Then object i is considered to belong to the gray class k*;

[0037] All customer service agents are taken as evaluation objects. The evaluation index system is shown in Table 1. The observed value of object i with respect to index j is recorded as x ij , i = 1, 2, ...; j = 1, 2, ..., 15, Table 1 Evaluation index

[0038] Serial number Evaluation indicators Unit of measurement 1 Number of absences this month Second-rate 2 The number of abnormal number of meal leave orders in the month Second-rate 3 The number of abnormal times of daily training time less than 15 minutes in the current month Second-rate 4 Number of incorrect clicks when leaving training in the current month Second-rate 5 Number of incorrect clicks on the class leave in the current month Second-rate 6 Number of times the absence ratio during the month was greater than 8% Second-rate 7 Number of times a person was absent from his seat for a break longer than 10 minutes in the month Second-rate 8 Number of days with more than 30 absences per day in the month sky 9 Number of absences less than 1 minute in the month Second-rate 10 Number of days in the month with a single-day transaction duration greater than 0 sky 11 Number of days with more than 12 check-outs per day in the current month sky 12 Number of days of delayed readiness per day in the current month sky 13 Number of extra long blank days in a single day of the month sky 14 Number of emergencies this month Second-rate 15 Emergency time this month Hour

[0039] S300: Set the turning point and establish the gray possibility function;

[0040] The possible degree function f of the subclass with index j (k=1, 2, 3) j k (·) can be specifically expressed as f j k (x j k (1), x j k (2), x j k (3), x j k (4)), where x j k (1), x j k (2), x j k (3), x j k (4) is f j k (·) is the turning point; the possibility functions of the three gray classes divided by index j are set as

[0041]

[0042] in,

[0043] Then the possible degree function f of index j about gray class k is j k (·), j = 1, 2, 3, ..., 15; k = 1, 2, 3, determined by the following calculation formula, the specific function formula is shown in Table 3,

[0044]

[0045] S400: Calculate clustering weights;

[0046] S500: Calculate the grey clustering coefficient to determine the grey class to which the evaluation object belongs;

[0047] According to the following formula

[0048] (i=1, 2, ...; j=1, 2, ..., 15; k=1, 2, 3) Calculate the grey clustering coefficient of each customer service agent, obtain the grey fixed weight clustering coefficient, and determine the grey class to which each customer service agent belongs according to the following formula;

[0049] i=1,2,…;k=1,2,3.

[0050] Preferably, in S400, the clustering weight is determined according to the following method:

[0051] S410: Determine the initial weight of each indicator using the analytic hierarchy process; Invite relevant experts in the industry to score the three first-level evaluation indicators and the three second-level evaluation indicators under each indicator according to their importance, and construct a judgment matrix. The integrated expert opinions are used to construct four judgment matrices, and the weight vectors of each indicator are obtained after calculation and consistency test;

[0052] S420: using the entropy weight calculated by the entropy value method to adjust the initial weight of the indicator;

[0053] S430: Integrate the two groups of indicator weight vectors according to the following formula to calculate the weight vectors of 15 evaluation indicators:

[0054]

[0055] Beneficial effects of the present invention:

[0056] By using online data such as absence duration, number of absences and specific time, number of sign-ins and sign-outs, as well as offline personnel information such as shifts, leave, and training, we can quickly integrate and analyze relevant information through technical means, accurately locate customer service personnel who have irregular behaviors before, during, and after work, and form a timely and accurate list of personnel with irregular behaviors every day, providing effective support for operations management.

[0057] In terms of emergency support analysis, by cleaning, converting, processing and analyzing online data such as agent calls, agent softphone applications, agent softphone details and offline data such as scheduling and personnel information, we mine the emergency data of customer service specialists during non-scheduled hours, and regularly count the emergency efficiency data of customer service personnel, to provide effective support for emergency management and improve the overall bottom-line support capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a schematic diagram of the evaluation method flow in an embodiment of the present invention. DETAILED DESCRIPTION

[0059] The customer service agent work behavior evaluation method includes the following process:

[0060] S100: Standardized evaluation of customer service agent work; the specific process is as follows:

[0061] Establish "absentee list for the day", "abnormal list of meal leave selection for the day", "abnormal list of training duration less than 15 minutes for the day", "wrong list of training leave selection for the day", "wrong list of shift leave selection for the day", "list of absence ratio greater than 8% for the day", "list of single break leave greater than 10 minutes for the day", "list of absences greater than 30 times for the day", "list of single absences less than 1 minute for the day", "list of affairs duration greater than 0 for the day", "list of sign-outs greater than 12 times for the day", "list of delayed readiness for the day", "list of extra-long blank for the day", and the initial list is "empty";

[0062] Set up "Number of absences in the month", "Number of abnormal selections for meal leave in the month", "Number of abnormal daily training duration less than 15 minutes in the month", "Number of incorrect selections for training leave in the month", "Number of incorrect selections for shift leave in the month", "Number of times the ratio of leave for rest is greater than 8% in the month", "Number of single leave for rest is greater than 10 minutes in the month", "Number of days with more than 30 absences in a single day in the month", "Number of single absences less than 1 minute in the month", "Number of days with more than 0 daily affairs duration in the month", "Number of days with more than 12 sign-outs in a single day in the month", "Number of days with delayed readiness in a single day in the month", "Number of extra-long blank days in a single day in the month", "Number of emergency times in the month", "Emergency duration in the month", and the initial data is 0;

[0063] The initial value of the customer service agent's daily work standardization deduction value is set to 0;

[0064] S101: Determine whether it is absence. Specific process: If both the first sign-in time and the first ready time are empty, the agent is included in the "absentee list of the day", the "absent times of the month" of the agent increases by 1, and the customer service agent's work standardization deduction value for the day is increased by 10 and then go to S115, otherwise go to S102;

[0065] S102: Determine whether it is absence. Specific process: If the last sign-out time, the last customer active hang-up time, the last agent active hang-up time, and the last transfer satisfaction time are all empty, the agent will be included in the "absentee list of the day", the agent's "absent times of the month" will increase by 1, and the customer service agent's work standardization deduction value for the day will be increased by 10, and then go to S115, otherwise go to S103;

[0066] S103: Determine whether the dining leave selection is abnormal. If the dining time of the customer service agent is greater than 0, the agent is included in the "List of Abnormal Dining Leave Selections of the Day", the agent's "Number of Abnormal Dining Leave Selections of the Month" is increased by 1, and the customer service agent's work standardization deduction value for the day is increased by 1, and then the process goes to S104; otherwise, the process goes directly to S104;

[0067] S104: Determine whether the training time is less than 15 minutes. If the training time of the customer service agent is less than 15 minutes, the agent is included in the "abnormal list of training time less than 15 minutes on the same day", and the "abnormal number of daily training time less than 15 minutes in the same month" of the agent is increased by 1, and the customer service agent's work standardization deduction value on the same day is increased by 1, and then go to S105; otherwise, go directly to S105;

[0068] S105: If the training time of the customer service agent is not from the department and is greater than 0, the agent is included in the "List of Wrong Training Leave Clicks on the Day", the agent's "Number of Wrong Training Leave Clicks on the Month" is increased by 1, and the customer service agent's work standardization deduction value on the day is increased by 1, and then the process goes to S106; otherwise, the process goes directly to S106;

[0069] S106: Determine whether the shift leave selection is wrong; first determine whether the agent is a shift leader, if yes, go directly to S107; otherwise, determine whether the shift duration is greater than 0, if yes, put the agent in the "list of wrong shift leave selections for the day", the agent's "number of wrong shift leave selections for the month" is increased by 1, and the customer service agent's work standardization deduction value for the day is +1 and then go to S107, otherwise go directly to S107;

[0070] S107: Determine whether the daily rest absence ratio is greater than 8%; determine whether the rest time / sign-in time is greater than 8%. If so, the agent is included in the "list of agents with a daily rest absence ratio greater than 8%", the agent's "monthly rest absence ratio greater than 8% times" is increased by 1, and the customer service agent's daily work standardization deduction value is increased by 1 and then go to S108, otherwise go directly to S108;

[0071] S108: Determine whether the single break absence on the day is greater than 10 minutes; if yes, the agent is included in the "list of single break absences greater than 10 minutes on the day", the agent's "number of single break absences greater than 10 minutes in the month" is increased by 1, and the customer service agent's work standardization deduction value on the day is increased by 1 and then go to S109; otherwise, go directly to S109;

[0072] S109: Determine whether the number of absences on the day is greater than 30 times. If yes, the agent is included in the "list of absences greater than 30 times on the day", the number of days with more than 30 absences on a single day in the month for the agent is increased by 1, and the standardization deduction value of the customer service agent's work on the day is increased by 1 and then the process goes to S110. Otherwise, the process goes directly to S110;

[0073] S110: Determine whether there is a single absence of less than 1 minute, then add the agent to the "list of single absences of less than 1 minute on the day", increase the "number of single absences of less than 1 minute on the month" of the agent by 1, and increase the deduction score of the customer service agent's work standardization on the day by 1, then go to S111, otherwise go directly to S111;

[0074] S111: Determine whether there is a transaction duration greater than 0. If yes, add the agent to the "list of transactions of the day greater than 0", increase the number of days of the agent's "single-day transaction duration greater than 0 times in the month" by 1, and increase the deduction score of the customer service agent's work standardization on that day by 1, and then go to S112. Otherwise, go directly to S112;

[0075] S112: Determine whether there is a person who has signed out more than 12 times. If yes, add the agent to the "list of people who have signed out more than 12 times on the same day", increase the number of days for the agent to sign out more than 12 times on the same day by 1 day, and increase the deduction score for the standardization of the customer service agent's work on that day by 1, and then go to S113. Otherwise, go directly to S113.

[0076] S113: When the staff is not on a rest shift, determine whether there is a delay in readiness; specific process: determine whether the first readiness time is more than 2 minutes later than the working time, if yes, put the staff in the "list of delayed readiness for the day", increase the "number of delayed readiness days for a single day in the month" of the staff by 1, and increase the deduction point value of the customer service staff's work standardization on that day by 1 and then go to S114, otherwise go directly to S114;

[0077] S114: When the staff is not on rest shift, determine whether there is an extra-long blank; specific process: determine whether the latest time among the last sign-out time, the last customer active hang-up time, the last agent active hang-up time, and the last transfer satisfaction time is more than 10 minutes earlier than the end of get off work time. If so, the agent will be included in the "Extra-long Blank List of the Day", the "Extra-long Blank Days per Day of the Month" of the agent will be increased by 1 day, and the customer service agent's work standardization deduction value for the day will be increased by 1 and then go to S115, otherwise go directly to S115; return to "Extra-long Blank"

[0078] S115: End the standardized evaluation of customer service agents' work;

[0079] S200: Evaluation of customer service agent emergency support capabilities; the specific process is as follows:

[0080] The initial value of the number of emergency calls for customer service agents in the month is set to 0, and the initial value of the emergency duration is set to 0;

[0081] S201: Determine whether the customer service agent is on break shift, if yes, go to S202, otherwise go to;

[0082] S202: Determine the emergency duration of the day by checking the agent calls, agent softphone applications, and agent softphone detailed online data, and add it to the number of emergency times and emergency duration of the month;

[0083] S203: End the evaluation of the customer service agent's emergency support capability for the day.

[0084] This patent adopts the grey possibility function, combined with the indicator weight determined by the extreme entropy method, to divide all agents into three grey categories: excellent agents, general agents, and agents to be improved;

[0085] There are n clustering objects, m clustering indicators and s gray classes; x ij (i=1, 2, ..., n; j=1, 2, ..., m) is the observed value of cluster object i with respect to cluster index j; f j k (·)(j=1, 2, …, m; k=1, 2, …, s) is the possible degree function of index j with respect to gray class k;

[0086]

[0087] w j is the fixed-weight clustering coefficient of object i belonging to k gray classes, where w j is the clustering weight of the indicator and

[0088] like Then object i is considered to belong to the gray class k*;

[0089] All customer service agents are evaluated. The evaluation index system is shown in Table 1. The observed value of i with respect to index j is denoted as x. ij ,i=1,2,…;j=1,2,…,15,

[0090] Table 1 Evaluation indicators

[0091] Serial number Evaluation indicators Unit of measurement 1 Number of absences this month Second-rate 2 The number of abnormal number of meal leave orders in the month Second-rate 3 The number of abnormal times of daily training time less than 15 minutes in the current month Second-rate 4 Number of incorrect clicks when leaving training in the current month Second-rate 5 Number of incorrect clicks on the class leave in the current month Second-rate 6 Number of times the absence ratio during the month was greater than 8% Second-rate 7 Number of times a person was absent from his seat for a break longer than 10 minutes in the month Second-rate 8 Number of days with more than 30 absences per day in the month sky 9 Number of absences less than 1 minute in the month Second-rate 10 Number of days in the month with a single-day transaction duration greater than 0 sky 11 Number of days with more than 12 check-outs per day in the current month sky 12 Number of days of delayed readiness per day in the current month sky 13 Number of extra long blank days in a single day of the month sky 14 Number of emergencies this month Second-rate 15 Emergency time this month Hour

[0092] S300: Set the turning point and establish the gray possibility function;

[0093] The possible degree function f of the subclass with index j (k=1, 2, 3) j k (·) can be specifically expressed as f j k (x j k (1), x j k (2), x j k (3), x j k (4)), where x j k (1), x j k (2), x j k (3), x j k (4) is f j k (·) is the turning point; the possibility functions of the three gray classes divided by index j are set as

[0094]

[0095] in,

[0096] Then the possible degree function f of index j about gray class k is j k (·), j = 1, 2, 3, ..., 15; k = 1, 2, 3, determined by the following calculation formula, the specific function formula is shown in Table 3,

[0097]

[0098] S400: Calculate clustering weights; the clustering weights are determined according to the following method:

[0099] S410: Determine the initial weight of each indicator using the analytic hierarchy process; Invite relevant experts in the industry to score the three first-level evaluation indicators and the three second-level evaluation indicators under each indicator according to their importance, and construct a judgment matrix. The integrated expert opinions are used to construct four judgment matrices, and the weight vectors of each indicator are obtained after calculation and consistency test;

[0100] S420: using the entropy weight calculated by the entropy value method to adjust the initial weight of the indicator;

[0101] S430: Integrate the two groups of indicator weight vectors according to the following formula to calculate the weight vectors of 15 evaluation indicators:

[0102]

[0103] S500: Calculate the grey clustering coefficient to determine the grey class to which the evaluation object belongs;

[0104] According to the following formula:

[0105] (i=1, 2, ...; j=1, 2, ..., 15; k=1, 2, 3) Calculate the grey clustering coefficient of each customer service agent, obtain the grey fixed weight clustering coefficient, and determine the grey class to which each customer service agent belongs according to the following formula;

[0106] i=1,2,…;k=1,2,3.

Claims

1. A method for evaluating the work behavior of customer service agents, characterized in that: The process includes the following: S100: Standardized evaluation of customer service agent work; the specific process is as follows: Establish "absentee list for the day", "abnormal list of meal leave selection for the day", "abnormal list of training duration less than 15 minutes for the day", "wrong list of training leave selection for the day", "wrong list of shift leave selection for the day", "list of absence ratio for rest greater than 8% for the day", "list of single break absence for more than 10 minutes for the day", "list of absences for more than 30 times for the day", "list of single absence for less than 1 minute for the day", "list of affairs duration greater than 0 for the day", "list of sign-outs for more than 12 times for the day", "list of delayed readiness for the day", "list of extra-long blank for the day", and the initial list is "empty"; Set up "Number of absences in the month", "Number of abnormal selection of meal leave in the month", "Number of abnormal daily training duration less than 15 minutes in the month", "Number of incorrect selection of training leave in the month", "Number of incorrect selection of class leave in the month", "Number of rest leave ratio greater than 8% in the month", "Number of single rest leave greater than 10 minutes in the month", "Number of days with more than 30 absences in a single day in the month", "Number of single absences less than 1 minute in the month", "Number of days with more than 0 daily affairs duration in the month", "Number of days with more than 12 daily sign-outs in the month", "Number of days with delayed readiness in a single day in the month", "Number of extra-long blank days in a single day in the month", "Number of emergency times in the month", "Emergency duration in the month", and the initial data is 0; The initial value of the customer service agent's daily work standardization deduction value is set to 0; S101: Determine whether it is absence. Specific process: If both the first sign-in time and the first ready time are empty, the agent will be included in the "absentee list of the day", the "absent times of the month" of the agent will increase by 1, and the customer service agent's work standardization deduction value for the day will be increased by 10, then go to S115, otherwise go to S102; S102: Determine whether it is absence. Specific process: If the last sign-out time, the last customer active hang-up time, the last agent active hang-up time, and the last satisfaction transfer time are all empty, the agent will be included in the "absentee list of the day", the agent's "absent times of the month" will increase by 1, and the customer service agent's work standardization deduction value for the day will be increased by 10, and then go to S115, otherwise go to S103; S103: Determine whether the dining leave selection is abnormal. If the dining time of the customer service agent is greater than 0, the agent is included in the "List of Abnormal Dining Leave Selections of the Day", the agent's "Number of Abnormal Dining Leave Selections of the Month" is increased by 1, and the customer service agent's work standardization deduction value for the day is increased by 1, and then the process goes to S104; otherwise, the process goes directly to S104; S104: Determine whether the training time is less than 15 minutes. If the training time of the customer service agent is less than 15 minutes, the agent is included in the "abnormal list of training time less than 15 minutes on the same day", and the "abnormal number of daily training time less than 15 minutes in the same month" of the agent is increased by 1, and the customer service agent's work standardization deduction value on the same day is increased by 1, and then go to S105; otherwise, go directly to S105; S105: If the training time of the customer service agent is not from the department and is greater than 0, the agent is included in the "List of Wrong Training Leave Clicks on the Day", the agent's "Number of Wrong Training Leave Clicks on the Month" is increased by 1, and the customer service agent's work standardization deduction value on the day is increased by 1, and then the process goes to S106; otherwise, the process goes directly to S106; S106: Determine whether the shift leave selection is wrong; first determine whether the agent is a shift leader, if yes, go directly to S107; otherwise, determine whether the shift duration is greater than 0, if yes, put the agent in the "list of wrong shift leave selections for the day", the agent's "number of wrong shift leave selections for the month" increases by 1, and the customer service agent's work standardization deduction value for the day is +1 and then go to S107, otherwise go directly to S107; S107: Determine whether the daily rest absence ratio is greater than 8%; determine whether the rest time / sign-in time is greater than 8%. If so, the agent is included in the "list of agents with a daily rest absence ratio greater than 8%", and the agent's "number of times the monthly rest absence ratio is greater than 8%" is increased by 1, and the customer service agent's work standardization deduction value for the day is increased by 1 and then go to S108. Otherwise, go directly to S108; S108: Determine whether the single absence from the seat for a break on the day is greater than 10 minutes; if so, the agent is included in the "list of absences from the seat for a single break greater than 10 minutes on the day", the agent's "number of absences from the seat for a single break greater than 10 minutes in the month" is increased by 1, and the customer service agent's work standardization deduction value for the day is increased by 1 and then the process goes to S109, otherwise, the process goes directly to S109; S109: Determine whether the number of absences on the day is greater than 30 times. If yes, the agent is included in the "list of absences greater than 30 times on the day", and the "number of days with more than 30 absences on a single day in the month" of the agent is increased by 1 day, and the customer service agent's work standardization deduction value on the day is increased by 1 and then go to S110. Otherwise, go directly to S110; S110: Determine whether there is a single absence of less than 1 minute, then add the agent to the "list of single absences of less than 1 minute on the day", increase the "number of single absences of less than 1 minute on the month" of the agent by 1, and increase the deduction score of the customer service agent's work standardization on the day by 1, then go to S111, otherwise go directly to S111; S111: Determine whether there is a transaction duration greater than 0. If yes, add the agent to the "list of transactions of the day greater than 0", increase the number of days with transaction duration greater than 0 for the agent by 1, and increase the deduction score of the agent's work standardization on the day by 1, then go to S112. Otherwise, go directly to S112; S112: Determine whether there is a person who has signed out more than 12 times. If yes, add the agent to the "list of people who have signed out more than 12 times on the same day", increase the number of days for the agent to sign out more than 12 times on the same day by 1 day, and increase the deduction score for the standardization of the customer service agent's work on that day by 1, then go to S113. Otherwise, go directly to S113. S113: When the staff is not on a rest shift, determine whether there is a delay in readiness; specific process: determine whether the first readiness time is more than 2 minutes later than the working time. If so, the seat staff is included in the "list of delayed readiness for the day", the "number of delayed readiness days for a single day in the month" of the seat staff is increased by 1, and the customer service seat staff's work standardization deduction value for the day is increased by 1 and then go to S114, otherwise directly go to S114; S114: When the staff is not on rest shift, determine whether there is an extra-long blank; specific process: determine whether the latest time among the last sign-out time, the last customer active hang-up time, the last agent active hang-up time, and the last transfer satisfaction time is more than 10 minutes earlier than the end of get off work time. If so, the agent is included in the "Extra-long blank list of the day", the "Extra-long blank days per day of the month" of the agent is increased by 1 day, and the customer service agent's work standardization deduction value for the day is increased by 1 and then go to S115, otherwise go directly to S115; return to "Extra-long blank" S115: End the standardized evaluation of customer service agents' work; S200: Evaluation of customer service agent emergency support capabilities; the specific process is as follows: The initial value of the number of emergency calls for customer service agents in the month is set to 0, and the initial value of the emergency duration is set to 0; S201: Determine whether the customer service agent is on break shift, if yes, go to S202, otherwise go to; S202: Determine the emergency duration of the day by checking the agent calls, agent softphone applications, and agent softphone detailed online data, and add it to the number of emergency times and emergency duration of the month; S203: End the evaluation of the customer service agent's emergency support capability for the day.

2. The method for evaluating the work behavior of customer service agents according to claim 1, characterized in that: The following follow-up processes are also included: A detailed evaluation list for each customer service agent is formed by day, month, and year. The smaller the deduction score for the customer service agent's work standardization on that day, the better the agent is; the greater the number of emergency responses and the longer the emergency duration, the better the agent is.

3. The method for evaluating the work behavior of customer service agents according to claim 1, characterized in that: The following follow-up processes are also included: This patent adopts the grey possibility function, combined with the indicator weight determined by the extreme entropy method, to divide all agents into three grey categories: excellent agents, general agents, and agents to be improved; There are n clustering objects, m clustering indicators and s gray classes; x ij (i=1, 2, ..., n; j=1, 2, ..., m) is the observed value of cluster object i with respect to cluster index j; f j k (·)(j=1, 2, …, m; k=1, 2, …, s) is the possible degree function of index j with respect to gray class k; w j is the fixed-weight clustering coefficient of object i belonging to k gray classes, where w j is the clustering weight of the indicator and like Then object i is considered to belong to the gray class k*; All customer service agents are taken as evaluation objects. The evaluation index system is shown in Table 1. The observed value of object i with respect to index j is recorded as x ij ,i=1,2,…;j=1,2,…,15, Table 1 Evaluation indicators S300: Set the turning point and establish the gray possibility function; The possible degree function f of the subclass with index j (k=1, 2, 3) j k (·) can be specifically expressed as f j k (x j k (1), x j k (2), x j k (3), x j k (4)), where x j k (1), x j k (2), x j k (3), x j k (4) is f j k (·) is the turning point; the possibility functions of the three gray classes divided by index j are set as in, Then the possible degree function f of index j about gray class k is j k (·), j = 1, 2, 3, ..., 15; k = 1, 2, 3, determined by the following calculation formula, the specific function formula is shown in Table 3, S400: Calculate clustering weights; S500: Calculate the grey clustering coefficient to determine the grey class to which the evaluation object belongs; According to the following formula Calculate the grey clustering coefficient of each customer service agent, obtain the grey fixed-weight clustering coefficient, and determine the grey class to which each customer service agent belongs according to the following formula; 4. The method for evaluating the work behavior of customer service agents according to claim 1, characterized in that: In S400, the clustering weight is determined according to the following method: S410: Determine the initial weight of each indicator using the analytic hierarchy process; Invite relevant experts in the industry to score the three first-level evaluation indicators and the three second-level evaluation indicators under each indicator according to their importance, and construct a judgment matrix. The integrated expert opinions are used to construct four judgment matrices, and the weight vectors of each indicator are obtained after calculation and consistency test; S420: using the entropy weight calculated by the entropy value method to adjust the initial weight of the indicator; S430: Integrate the two groups of indicator weight vectors according to the following formula to calculate the weight vectors of 15 evaluation indicators:

Images