A SYSTEM THAT ENABLES CUSTOMER SATISFACTION TO BE PREDICTED.

Abstract

This invention relates to a system (1) that enables the conversion of customer conversations recorded during call centre operations into text using speech-to-text technology and the analysis of these texts using data science and artificial intelligence methods to automatically predict customer satisfaction survey results for all calls.

Description

1 TARIFF A SYSTEM THAT ENABLES CUSTOMER SATISFACTION TO BE PREDICTED. SYSTEM Technical Area 5 This invention records customer data during call center operations. their speeches are converted into text using speech-to-text technology and this All calls are analyzed using data science and artificial intelligence methods. 10. Automatically predicting customer satisfaction survey results It is related to a system that provides. Previous Technique Today, due to the rules established for measuring customer satisfaction, at least 15 Surveys are sent to a large number of customers, but very few customers complete these surveys. prefers to answer. This applies to each customer interview. This prevents a satisfaction assessment from being conducted. Furthermore, the current situation... Because assessment methods vary from person to person, The evaluation scores do not provide consistent and measurable results. 20 Therefore, considering the studies and shortcomings in the current technique... when considered, customer service handled by call center operations labeling the data from the interviews using data programming methods and BERT (Bidirectional Encoder Representations from Transformers- 25 Classification model of bidirectional encoder representations from transformers. This allows predicting customer satisfaction survey scores. It is clear that a system is needed. Chinese patent no. CN112801676A, which is included in the prior art, 30 In the document, call center customers were analyzed using a satisfaction prediction model. 2 predicting satisfaction, improving customer satisfaction, and a system that improves the service quality of the call center It is mentioned that the invention in question concerns a session of a call center system. Voice call data is obtained in real time, automated speech Voice search data is converted to text data using the recognition function, 5 based on customer text data and previously acquired historical logistics information customer using a pre-trained satisfaction prediction model Satisfaction is predicted and customer satisfaction is a predetermined factor. When it reaches the threshold, a guidance instruction for customer service guidance. is being sent. 10 Another document included in the known state of the art is US2022383329A1. In the United States patent document number [number], call transcripts and call quality a call center using an artificial intelligence model trained on its data a system that enables the prediction of customer satisfaction scores for 15 It is mentioned that customer satisfaction is predicted in this invention. A root cause analysis of factors related to customer satisfaction is being carried out, and The anticipated customer satisfaction is used for various purposes. The issue in question... The invention takes sentiment analysis into account in training the artificial intelligence model. Other optimizations include optimizing hyperparameters, over 20 Choosing between sampling and avoiding oversampling is crucial for education. Segmentation, development and testing according to call definition parameters. This involves determining the token size and choosing between BERT and XLNet. This includes the use of various artificial intelligence tools. Brief Description of the Invention The purpose of this invention is to record customer data during call center operations. their speeches are converted into text using speech-to-text technology and this All calls are analyzed using data science and artificial intelligence methods. 3 automatic prediction of customer satisfaction survey results It is about implementing a system developed to provide this. Another purpose of this invention is to improve the performance of call center operations. 5. Labeling customer interview data using data programming methods and customer satisfaction survey using the BERT classification model. a system developed to enable the prediction of scores to accomplish. Another aim of this invention is to increase the satisfaction level of each customer call by 10. determining this level and monitoring it as a quality performance indicator, When data is labeled with the help of an artificial intelligence model, each customer is identified for each call. so that satisfaction can be observed objectively, thus, Satisfaction survey scores for each call are faster and require more human intervention. the ability to predict without needing to use, call center operations 15 increasing its effectiveness and efficiency, eliminating the cost required for this process removal, continuous improvements aimed at customer satisfaction and to significantly improve the overall customer experience It is about implementing a developed system. Detailed Description of the Invention The "Customer Satisfaction" process was carried out to achieve the purpose of this invention. A system that enables prediction is shown in the attached figure; Figure 1 shows a schematic view of the system that is the subject of the invention. The parts shown in the figure are individually numbered, and these numbers correspond to... The corresponding answers are given below. 1. System 4 2. Database 3. Server Customer conversations recorded during call center operations speech-to-text technology that converts speech into text and these texts into data 5 Customer data for all calls is analyzed using scientific and artificial intelligence methods. To automatically predict satisfaction survey results. The system in question, developed for the purpose of invention (1); - data converted into text using speech-to-text technology. at least one database structured to keep records (2), 10 − connect to the database (2) using any communication protocol to set up, access data on the database (2) and the database (2) recording data, customer representatives within the company access to the conversations held, customer representatives The interviews he conducted were transcribed into text using speech-to-text technology. convert and convert the converted text into a database in the form of big data (2) recording conversations conducted by customer representatives by analyzing and monitoring customer satisfaction and employee performance. to analyze the transcripts of conversations with customer representatives by matching the transcripts with customer satisfaction survey results 20 to create a unique dataset and use this dataset to help the customer an artificial intelligence model that can automatically predict satisfaction to develop, various word phrases used in conversations, manually labeled BERT models created from a small amount of data versus big data. Prediction results, customer service representative and customer conversations, emotion 25 analysis, customer satisfaction survey scores and threshold values ​​of various KPIs data using values ​​in the form of rules based on those values to carry out the programming process, customer with the BERT model to enable prediction of satisfaction survey scores It includes at least one configured server (3). 30 The invention involves establishing a connection between the database (2) and the server (3) in the system (1). It is structured in such a way. The database (2) contains speech to text. to keep records of data converted into text using technology It is being structured. In the system in question (1), the server (3) is located in the known state of the art. the area connects to the database (2) using any communication protocol to set up, access data on the database (2) and data on the database (2) It is configured to save. The server (3) is configured to save the customer within the company. It is structured to provide access to the meetings held by their representatives. 10 The server (3) presents the conversations conducted by customer representatives, without speaking. converting text to text using writing technology and converting the converted text into big data. It is configured to save to the database (2) as follows. Server (3), By analyzing the conversations conducted by customer representatives, the customer 15 to monitor and analyze employee satisfaction and performance is configured. The server (3) handles the conversations with customer representatives. by matching text transcripts with customer satisfaction survey results to create a unique dataset and use this dataset to help the customer an artificial intelligence model that can automatically predict satisfaction It is structured to develop. Server (3) 20 data programming method and automatically predicting survey data using converter models It is configured to enable data programming. The server (3) Relabeling data using this technique and creating a reliable customer satisfaction survey. It is structured to generate values. The server (3) is used to generate values ​​in conversations. BERT 25, which consists of various word phrases created from a small amount of manually labeled data. (Bidirectional Encoder Representations from Transformers Prediction in big data using the Bidirectional Encoder Representations (BIR) model. The results include sentiment analysis of customer service and customer conversations, and customer satisfaction survey scores and various KPIs (Key Performance Indicators) Rules based on threshold values ​​of key performance indicators (KPIs) 30 to perform the data programming process using values ​​in this form 6 It is structured. Server (3), customer satisfaction survey with BERT model. It is configured to enable the prediction of the score. Server (3), Customer satisfaction labeled with data programming while training the BERT model. using the survey values ​​and one, two, three, four and five as model outputs It is structured to generate satisfaction survey scores in this way. Server 5 (3), by making sense of big data, customer satisfaction survey scores, all To enable automatic predictive behavior for conversations. It is structured accordingly. Industrial application of the invention 10 In the system that is the subject of the invention, (1) recorded during call center operations converting customer conversations from speech to text using speech-to-text technology and these texts are analyzed using data science and artificial intelligence methods to determine all Automatically predict customer satisfaction survey results for calls 15 This is ensured. The system in question (1) provides the user with data privacy principles in accordance with the principles of data privacy. This provides information and consent, in accordance with the Law on the Protection of Personal Data. It operates within the scope of (KVKK - Turkish Personal Data Protection Law). 20 Around these fundamental concepts, the subject of the invention is "Customer Satisfaction Prediction". A wide variety of applications related to "A System that Enables (1)" It is possible to develop further, and the invention cannot be limited to the examples described here. It is essentially as stated in the requests. 25

Claims

7 REQUESTS 1. Customer conversations recorded during call center operations the conversion of speech to text using speech-to-text technology and this All 5 texts are analyzed using data science and artificial intelligence methods. Customer satisfaction survey results for calls are automatically generated. developed to enable prediction; - data converted into text using speech-to-text technology. containing at least one database (2) structured to keep records; − Connect to the database (2) using any communication protocol 10 to set up, access data on the database (2) and the database (2) recording data, customer representatives within the company access to the conversations held, customer representatives The interviews he conducted were transcribed into text using speech-to-text technology. convert and convert the text into a database in the form of big data (2) 15 recording conversations conducted by customer representatives by analyzing and monitoring customer satisfaction and employee performance. to analyze the transcripts of conversations with customer representatives by matching the transcripts with the customer satisfaction survey results create a unique dataset and use this dataset to analyze customer 20 an artificial intelligence model that can automatically predict satisfaction to develop, various word phrases used in conversations, manually labeled BERT models created from a small amount of data versus big data. forecast results, customer service representative and customer conversation emotions analysis, customer satisfaction survey scores and various KPI values ​​threshold 25 data using values ​​in the form of rules based on those values to carry out the programming process, customer with the BERT model to enable prediction of satisfaction survey scores a system characterized by having at least one server (3) configured (1). 8 2. Database (2) configured to connect with Server (3) A system like the one in Claim 1, characterized (1).

3. Data that is converted into text using speech-to-text technology. 5 characterized by a database structured to keep records (2) A system like the one in Request 1 or 2 (1).

4. Connect to the database (2) using any communication protocol. to set up, access data on the database (2) and the database (2) 10 characterized by the server (3) configured to record data on it. a system like any of the above-mentioned requests (1).

5. Meetings conducted by customer representatives within the company. the above characterized by the server (3) configured to access a system like any of the requests (1). 15 6. Transcribing conversations conducted by customer representatives from spoken to written form. converting the text to a digital format using technology and transforming the converted text into big data. Server (3) configured to save to database (2) in the following way a 20 as in any of the above claims characterized by system (1).

7. By analyzing the conversations conducted by customer representatives, customers to monitor and analyze employee satisfaction and performance 25 of the above requests characterized by the configured server (3). a system like any other (1).

8. Transcripts of conversations with customer representatives, customer a unique data set by matching it with satisfaction survey results to create and use this dataset to automatically improve customer satisfaction 30 9 to develop an artificial intelligence model that can predict from the above requests characterized by the configured server (3) a system like any other (1).

9. Survey 5 using data programming methods and converter models to enable automatic prediction of data from the above requests characterized by the configured server (3) a system like any other (1).

10. Relabeling data and making it reliable using data programming techniques. customer satisfaction survey structured to create values in any of the above requests characterized by the server (3) such a system (1).

11. Various word phrases used in the conversations were labeled manually in a small number of 15 Predictions made on big data using the BERT model generated from the data. The results include sentiment analysis of customer service and customer conversations. customer satisfaction survey scores and thresholds for various KPI values data using values ​​in the form of rules based on those values 20 with server (3) configured to carry out the programming process as in any of the above characterized claims system (1).

12. Predicting customer satisfaction survey scores using the BERT model. The above 25 is characterized by the server (3) configured to provide a system like any of the requests (1).

13. While training the BERT model, the customer was labeled with data programming. using satisfaction survey values ​​and one, two, as model outputs, 30 to generate satisfaction survey scores in the form of three, four and five. from the above requests characterized by the configured server (3) a system like any other (1).

14. Making sense of big data by interpreting customer satisfaction survey scores, all 5. Making conversations automatically predictable the above characterized by the server (3) configured to provide a system like any of the requests (1). 15 25

Images