A system for scoring the performance of call center employees

Abstract

The present invention relates to a system (1) which enables calls with customers (U) calling a call center (CM) affiliated to at least one company to be recorded and performance evaluations to be performed regarding the call center (CM) representatives (A) dealing with the relevant customers (U).

Description

[0001] DESCRIPTION

[0002] A SYSTEM FOR SCORING THE PERFORMANCE OF CALL CENTER EMPLOYEES

[0003] Technical Field

[0004] The present invention relates to a system which enables calls with customers calling a call center affiliated to at least one company to be recorded and performance evaluations to be performed regarding the call center representatives dealing with the relevant customers.

[0005] Background of the Invention

[0006] Today, customers commonly contact the call center when they encounter any problem. With the established connection, the customer representative in the call center initiates a voice conversation with the customer on the other side. After the contact with the representative, the relevant customer is asked to participate in a survey, and the performance of the customer representative is requested to be evaluated by the relevant customer with the said survey. However, there may be a difference between the expectations of the customer from the solution and the desired action with the solution. At this point, the score given to the customer representative by the customer can be misleading at the point of evaluation. In the state of the art, various solutions have been found that enable the weight averaging of multiple metrics in the form of how many questions the call center employee asks during the conversation and how many keywords s / he uses in the scoring of the representative working in the call center, and responses in the related conversations to be processed by analyzing the conversation data performed in the call center.However, in the state of the art, no solution has been found that enables to determine whether the response given by the representative to the customer is correct or not, to include the mood information as a multiplier in the calculation of the performance of the customer representative, to take the difficulty levels in the customer and representative scenarios as a basis, and to match the difficulty levels with the correct answers transmitted to the customer.

[0007] The United States patent document no. US2024267466, an application included in the state of the art, discloses a new approach for representative performance evaluation that enables a call or chat coming from a customer to be converted into text in real time and a real-time scorecard analysis conducted on the device of the representative to be performed while the representative and the customer are communicating with each other. The real-time scorecard analysis may be conducted based on a set of live scoring rules from the server side. Additional analyses such as sentiment analysis and summary analysis can be performed on the server side and displayed to the representative through the user interface on the device of the representative. The automatically generated performance score of the representative is presented to the manager on demand or in real time, allowing the manager to monitor the performance of the representative while the representative is talking to the customer. The voice-to-text and sentiment service enables voice packets to be received and sent to a service that converts the voice data into text and to a service that analyzes the text for sentiment. Sentiment analysis can be performed on the text converted from voice data into text regarding the customer, the representative or both.

[0008] Summary of the Invention

[0009] An object of the present invention is to realize a system for scoring the performance of call center employees, which determines whether the response given by the representative to the customer is correct or not; enables the mood information to be included as a multiplier in the calculation of the performance ofthe customer representative; and the difficulty levels in the customer and representative scenarios to be taken as a basis.

[0010] Detailed Description of the Invention

[0011] “A System for Scoring the Performance of Call Center Employees” realized to fulfil the objective of the present invention is shown in the figure attached, in which:

[0012] Figure 1 is a schematic view of the inventive system for scoring the performance of call center employees.

[0013] The components illustrated in the figure are individually numbered, where the numbers refer to the following:

[0014] 1. System

[0015] 2. Electronic device

[0016] 3. Client device

[0017] 4. Analysis server

[0018] CM. Call center

[0019] A. Representative

[0020] U. Customer

[0021] An inventive system (1) for scoring the performance of call center employees, which enables the call center (CM) affiliated to at least one company to communicate with callers, and performance evaluation to be performed regarding the call center (CM) representatives (A) dealing with the relevant individuals; comprises

[0022] at least one electronic device (2) which is configured to enable communication to be initiated by at least one customer (U) with the call center (CM) through the hardware located thereon;at least one client device (3) which is used by the representative (A) in the call center (CM), and is configured to enable the communication initiated through the electronic device (2) to be responded to by the representative (A); and

[0023] at least one analysis server (4) which is configured to access various data through the call center (CM) in the form of products and customers (U) recorded in the call center (CM); to establish connection with the client device (3) and the electronic device (2); to convert the conversations of the representative (A) with the customer (U) into text; to identify the named entity recognition (NER) expectations in the conversation of the customer (U); to determine whether the identified sentences of the customer (U) are complaints and / or requests through the classification algorithm thereon; to enable the transactions made by the representative (A) during the call between the representative (A) and the customer (U) to be extracted from the logs stored in the call center (CM); to access the documents of the product and campaign information asked by the customer (U) by converting the information on the screen opened by the representative (A) on the client device (3) into text with OCR; to each request of the customer (U) to be determined by running BERT-based algorithms on the documents it accesses; to enable word embedding to be applied by converting the text expressions of the responses given by the representative (A) and the questions of the customer (U) during the related conversation into vector; responses provided to the said customer (U) to be determined by analyzing the similarities between the responses corresponding to each request of the customer (U) obtained with BERT-based algorithms and the responses given by the representative (A) whose texts have been converted into vector; and to enable the performance of the call center (CM) representative (A) to be determined by taking the average or median of the similarities in all responses given to the customer (U).The electronic device (1) included in the inventive system (2) is a smart device in the form of a mobile phone, tablet computer or computer configured to enable the customer (U) to establish audio and video communication with the call center (CM) through the hardware located thereon.

[0024] The client device (3) included in the inventive system (1) is a smart device in the form of a mobile phone, tablet computer or computer that establishes connection with the call center (CM), and is configured to enable the electronic device (2) calls received by the call center (CM) to be responded to by the representative (A); and to transmit the conversations of the customer (U) transmitted through the electronic device (2) and the responses of the representative (A) given to the relevant customer (U) to the analysis server (4).

[0025] The analysis server (4) included in the inventive system (1) is configured to receive documents related to the product and campaign that the customer (U) asks through the electronic device (2) from at least one database located in the call center (CM), and to enable, in the event that a response that does not contain general information about the product specific to the customer (U) is transmitted to the customer (U) by the representative (A), the relevant representative (A) to be given a score of “0” from the said customer (U) request by enabling the responses that should be given to each request of the customer (U) to be determined by processing the documents it receives with BERT-based algorithms. The analysis server (4) is configured to convert the texts in the question of the customer (U) into a vector with the responses given by the representative (A) during the conversation between the representative (A) and the customer (U) taking place between the client device (3) and the electronic device (2), and to determine which response the representative (A) gave to which question by looking at the similarities between the responses by applying the word embedding method on the data it has converted into a vector. The analysis server (4) is configured to subtract the probability of adverse and negative marked moods being negative from one and multiply it by the customer (U) score, by enabling the moodanalysis of the representative (A) during the conversation taking place between the client device (3) and the electronic device (2) to be performed. The analysis server (4) is then configured to enable the mood of the conversation that the customer (U) had with the call center representative (A) to be analyzed, and to perform clustering by evaluating the data in the form of the number of requests in the said customer (U) conversation, the mood, the number of times the customer (U) called the call center (CM) in the last one and the last week, the mood of the customer (U) in the last one week, and the difficulties experienced by the customer (U). The analysis server (4) is configured to determine the difficulty level of calls by processing new incoming calls to the call center (CM) with the same key performance indicators and classification algorithm, to enable the representative (A) to be re-scored by normalizing the responses given by the representative (A) according to the difficulty level of the incoming call, and thus, to enable the relevant representative (A) to be scored by evaluating the responses given by the representative (A) within the determined difficulty level.

[0026] Industrial Application of the Invention

[0027] In the inventive system, (1) customer (U) calls the call center (CM) with his / her electronic device (2). The representative (A) at the call center (CM) starts a conversation with the customer (A) by answering the call of the customer (U) coming from the electronic device (2). The speech of the customer (U) is converted into text by the analysis server (4) by using OCR. The information requested in the transcribed speech is retrieved from the stored documents and given to the customer (U) by the representative (A) as a response. Meanwhile, by running BERT -based algorithms on the speech text, responses and controls are found for the information requested by the customer. By looking at the similarities between the responses given by the representative (A) to the customer (U), the accuracy of the response given to the customer (U) is determined. In determining the performance of the representative (A), the evaluation of the data in the form of similarity average given for the responses given by the representative (A), themood of the representative (A), the number of calls of the customer (A) in the last month, and the mood of the customer (A) during the conversation is carried out together by means of the analysis server (4).

[0028] With the inventive system (1), it is enabled to determine whether the response of the representative (A) to the customer (U) is correct or not, to include the mood information as a multiplier in the calculation of the performance of the representative (A), and to take the difficulty levels in the customer (U) and representative (A) scenarios as a basis.

[0029] The inventive system (1) operates within the scope of the Personal Data Protection Law (KVKK). In this system (1), after obtaining approval from the customer (U), the information received from the relevant customer (U) is used in the performance evaluation of the representative (A).

[0030] Within these basic concepts; it is possible to develop various embodiments of the inventive “A System (1) for Scoring the Performance of Call Center Employees”; the invention cannot be limited to examples disclosed herein and it is essentially according to claims.

Claims

CLAIMS1. A system (1) for scoring the performance of call center employees, which enables the call center (CM) affiliated to at least one company to communicate with callers, and performance evaluation to be performed regarding the call center (CM) representatives (A) dealing with the relevant individuals; comprising;at least one electronic device (2) which is configured to enable communication to be initiated by at least one customer (U) with the call center (CM) through the hardware located thereon;at least one client device (3) which is used by the representative (A) in the call center (CM), and is configured to enable the communication initiated through the electronic device (2) to be responded to by the representative (A); and characterized byat least one analysis server (4) which is configured to access various data through the call center (CM) in the form of products and customers (U) recorded in the call center (CM); to establish connection with the client device (3) and the electronic device (2); to convert the conversations of the representative (A) with the customer (U) into text; to identify the named entity recognition (NER) expectations in the conversation of the customer (U); to determine whether the identified sentences of the customer (U) are complaints and / or requests through the classification algorithm thereon; to enable the transactions made by the representative (A) during the call between the representative (A) and the customer (U) to be extracted from the logs stored in the call center (CM); to access the documents of the product and campaign information asked by the customer (U) by converting the information on the screen opened by the representative (A) on the client device (3) into text with OCR; to each request of the customer (U) to be determined by running BERT-based algorithms on the documents it accesses; to enable word embedding to be applied by converting the text expressions of the responses given by the representative (A) and the questions of the customer (U) during the relatedconversation into vector; responses provided to the said customer (U) to be determined by analyzing the similarities between the responses corresponding to each request of the customer (U) obtained with BERT- based algorithms and the responses given by the representative (A) whose texts have been converted into vector; and to enable the performance of the call center (CM) representative (A) to be determined by taking the average or median of the similarities in all responses given to the customer (U).

2. A system (1) for scoring the performance of call center employees, according to Claim 1; characterized by the electronic device (2) which is a smart device in the form of a mobile phone, tablet computer or computer configured to enable the customer (U) to establish audio and video communication with the call center (CM) through the hardware located thereon.

3. A system (1) for scoring the performance of call center employees, according to Claim 1 or 2; characterized by the client device (3) which is a smart device in the form of a mobile phone, tablet computer or computer that establishes connection with the call center (CM), and is configured to enable the electronic device (2) calls received by the call center (CM) to be responded to by the representative (A); and to transmit the conversations of the customer (U) transmitted through the electronic device (2) and the responses of the representative (A) given to the relevant customer (U) to the analysis server (4).

4. A system (1) for scoring the performance of call center employees, according to any one of the preceding claims; characterized by the analysis server (4) which is configured to receive documents related to the product and campaign that the customer (U) asks through the electronic device (2) from at least one database located in the call center (CM), and to enable, in the event that a response that does not contain general information about the product specific to the customer (U) is transmitted to the customer (U) by the representative (A), the relevant representative (A) to be given a score of “0” from the said customer (U) requestby enabling the responses that should be given to each request of the customer (U) to be determined by processing the documents it receives with BERT-based algorithms.

5. A system (1) for scoring the performance of call center employees, according to any one of the preceding claims; characterized by the analysis server (4) which is configured to convert the texts in the question of the customer (U) into a vector with the responses given by the representative (A) during the conversation between the representative (A) and the customer (U) taking place between the client device (3) and the electronic device (2), and to determine which response the representative (A) gave to which question by looking at the similarities between the responses by applying the word embedding method on the data it has converted into a vector.

6. A system (1) for scoring the performance of call center employees, according to any one of the preceding claims; characterized by the analysis server (4) which is configured to subtract the probability of adverse and negative marked moods being negative from one and multiply it by the customer (U) score, by enabling the mood analysis of the representative (A) during the conversation taking place between the client device (3) and the electronic device (2) to be performed.

7. A system (1) for scoring the performance of call center employees, according to any one of the preceding claims; characterized by the analysis server (4) which is then configured to enable the mood of the conversation that the customer (U) had with the call center representative (A) to be analyzed, and to perform clustering by evaluating the data in the form of the number of requests in the said customer (U) conversation, the mood, the number of times the customer (U) called the call center (CM) in the last one and the last week, the mood of the customer (U) in the last one week, and the difficulties experienced by the customer (U).

8. A system (1) for scoring the performance of call center employees, according to any one of the preceding claims; characterized by the analysis server (4) which is configured to determine the difficulty level of calls by processing new incoming calls to the call center (CM) with the same key performance indicators and classification algorithm, to enable the representative (A) to be re-scored by normalizing the responses given by the representative (A) according to the difficulty level of the incoming call, and thus, to enable the relevant representative (A) to be scored by evaluating the responses given by the representative (A) within the determined difficulty level.

Images