Deep fusion method of artificial intelligence and customer service scene

Abstract

The invention relates to the technical field of artificial customer service, and discloses a method for deep fusion of artificial intelligence and customer service scenes. The artificial intelligence and customer service scene deep fusion method comprises the following steps: establishing an artificial intelligence customer service learning system; establishing a classification database; establishing an entry analysis system; according to the invention, the established classification database contains most customer service scene fields in the market, and the artificial intelligence customer service can automatically answer the questions in different fields proposed by the customer, so that the adaptability and flexibility of the artificial intelligence customer service are greatly improved; when an artificial intelligence customer service cannot respond to a customer according to the content in the classification database, the artificial intelligence customer service can be automatically switched to ensure the flexibility of the artificial intelligence customer service, and entries are established and stored in the industry sub-database according to the content required by the customer and responded by the artificial customer service in combination with product words and customer demand words, so that the user experience is improved. And automatic expansion of classified database contents is realized.

Description

Technical Field

[0001] The present invention belongs to the field of artificial customer service technology, and specifically is a method for deep integration of artificial intelligence and customer service scenarios. Background Art

[0002] Artificial intelligence customer service is an intelligent customer service system that uses artificial intelligence technology to provide interactive services to customers. Through natural language processing technology, speech recognition technology, machine learning technology, etc., it can understand customer needs, answer customer questions, and provide solutions. Artificial intelligence customer service is more efficient than manual customer service when handling simple and repetitive questions, and can be online 24 hours a day, greatly saving labor costs. It can automatically identify and answer customer questions, handle customer complaints and suggestions, provide fast and accurate service responses, and improve customer satisfaction.

[0003] Traditional artificial intelligence customer service is mostly targeted at a specific service field during training and learning, which results in a single artificial intelligence customer service program being unable to adapt to service industries in different fields. In addition, traditional artificial intelligence customer service is mostly passive artificial intelligence customer service. When communicating with customers, it gives customers choices based on pre-designed question options and provides corresponding answers based on the questions selected by the customers. This results in artificial intelligence customer service being not flexible enough in practical work. Therefore, in view of the current situation, it needs to be improved. Summary of the invention

[0004] In view of the above situation, in order to overcome the defects of the prior art, the present invention provides a method for the deep integration of artificial intelligence and customer service scenarios, which effectively solves the problem that a single artificial intelligence customer service program cannot adapt to service industries in different fields, and that most traditional artificial intelligence customer services are passive artificial intelligence customer services. When communicating with customers, they give customers choices based on pre-designed question options and make corresponding answers based on the questions selected by the customers, which leads to the problem that artificial intelligence customer service is not flexible enough in practical work.

[0005] To achieve the above purpose, the present invention provides the following technical solution: a method for deep integration of artificial intelligence and customer service scenarios:

[0006] The method is based on:

[0007] An artificial intelligence customer service learning system is used to split the conversation between the human customer service and the customer into entries, including product words, customer demand words, and customer service response words;

[0008] Classification databases, including retail industry sub-database, financial industry sub-database, telecommunications industry sub-database, gaming industry sub-database, catering industry sub-database and other industry sub-databases;

[0009] a term analysis system for searching the term in the classification database;

[0010] The method comprises the following steps:

[0011] Step 1: The AI ​​customer service learning system collects the conversation content between the human customer service and the customer in real time, splits the content, and splits the content related to the product in the split content into entries;

[0012] Step 2: The classification database analyzes the product words in the entry obtained in step 1 through big data to determine the industry field to which it belongs, and then stores the content of the entry in the corresponding sub-database of the classification database;

[0013] If the industry field to which the product word belongs does not match the specific industry sub-database in the classification database, it will be stored in other industry sub-databases, and the entry content in other industry sub-databases will be manually classified by humans, and the number of specific industry sub-databases in the classification database will be expanded according to the classification content;

[0014] If there are two groups of entries about the same product word in the same sub-database, the two groups of entries are compared in content. When the repetition rate of the compared content reaches 85%, the entry with fewer words in the two groups is deleted;

[0015] Step 3: The term analysis system analyzes the product terms proposed by the customer through big data to determine the industry field to which they belong. If there is a sub-database of the industry in the classification database, the product terms and customer demand terms proposed by the customer are searched in the sub-database of the industry. When the corresponding content is retrieved, the customer service response terms corresponding to the product terms and customer demand terms stored in the sub-database are automatically replied;

[0016] If there is a specific industry sub-database corresponding to the product word proposed by the customer in the classification database, but the customer demand word stored in the industry sub-database is not retrieved, the call will be automatically transferred to the manual customer service, and the manual customer service will respond to the customer's needs, combine the product word and the customer demand word, create an entry and directly store it in the industry sub-database;

[0017] If there is no specific industry sub-database corresponding to the product word proposed by the customer in the classification database, the call will be automatically transferred to manual customer service, and the manual customer service will respond to the customer's needs, combine the product words and customer demand words, and create an entry that is directly stored in other industry sub-databases.

[0018] Preferably, the formula for the content repetition rate proposed in step 2 is repetition rate = number of repeated words / total number of words×100%.

[0019] Preferably, the artificial intelligence customer service proposed in step one consists of a data layer, an algorithm layer and a technical support layer, the algorithm layer includes machine learning algorithms, deep learning algorithms and NLP technology, and the technical support layer includes speech recognition technology, intent recognition technology and speech synthesis technology.

[0020] Preferably, the machine learning algorithm consists of a linear regression algorithm, a logistic regression algorithm, a decision tree algorithm and a K-means algorithm.

[0021] Preferably, the NLP technology integrates knowledge in the fields of linguistics, computer science, machine learning, mathematics, and cognitive psychology, and is composed of word vector technology and neural network technology.

[0022] Preferably, the deep learning algorithm consists of a convolutional neural network, a recurrent neural network, an optimization algorithm and an activation function.

[0023] Preferably, the classification database proposed in step 1 is a tree database, which adopts a tree data structure, including a group of root nodes and multiple groups of element nodes, and each group of element nodes includes a parent node and zero to two child nodes.

[0024] Compared with the prior art, the beneficial effects of the present invention are: 1. The present invention, through the established classification database, includes most of the customer service scenario fields on the market, so that the artificial intelligence customer service can automatically answer the questions in different fields raised by the customers, thereby greatly improving the adaptability of the artificial intelligence customer service, and when the artificial intelligence customer service cannot respond to the customer's question according to the content in the classification database, it can automatically transfer to the manual customer service, thereby ensuring the flexibility of the artificial intelligence customer service, and according to the content of the manual customer service's response to the customer's needs, combined with product words and customer demand words, establish entries and store them in the industry sub-database, thereby realizing the automatic expansion of the content of the classification database and improving the intelligence of the work of the artificial intelligence customer service.

[0025] 2. The present invention can automatically transfer the call to manual customer service when there is no specific industry sub-database corresponding to the product word proposed by the customer in the classification database, and establish entries for the content of the manual customer service's response to the customer's needs, combined with the product words and customer demand words, and store them in other industry sub-databases. In this way, the classification database can automatically store customer questions in unknown fields and the answers to the questions by manual customer service. Engineers can manually classify the content of the entries in other industry sub-databases and expand the number of specific industry sub-databases in the classification database according to the classification content, so as to increase the number of artificial intelligence customer services that adapt to different fields and provide a basis for the application of artificial intelligence customer service in all fields. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The accompanying drawings are used to provide further understanding of the present invention and constitute a part of the specification. They are used to explain the present invention together with the embodiments of the present invention and do not constitute a limitation of the present invention.

[0027] In the attached picture:

[0028] Figure 1 This is a structural framework diagram of a method for deep integration of artificial intelligence and customer service scenarios in the present invention;

[0029] Figure 2 It is a structural framework diagram of the classification database in the present invention;

[0030] Figure 3 It is a structural framework diagram of the term analysis system in the present invention. DETAILED DESCRIPTION

[0031] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments; based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0032] See also Figure 1-Figure 3 , the present invention provides a method for deep integration of artificial intelligence and customer service scenarios:

[0033] The method is based on:

[0034] An artificial intelligence customer service learning system is used to split the conversation between the human customer service and the customer into entries, including product words, customer demand words, and customer service response words;

[0035] Classification databases, including retail industry sub-database, financial industry sub-database, telecommunications industry sub-database, gaming industry sub-database, catering industry sub-database and other industry sub-databases;

[0036] a term analysis system for searching the term in the classification database;

[0037] The method comprises the following steps:

[0038] Step 1: The AI ​​customer service learning system collects the conversation content between the human customer service and the customer in real time, splits the content, and splits the content related to the product in the split content into entries;

[0039] Step 2: The classification database analyzes the product words in the entry obtained in step 1 through big data to determine the industry field to which it belongs, and then stores the content of the entry in the corresponding sub-database of the classification database;

[0040] If the industry field to which the product word belongs does not match the specific industry sub-database in the classification database, it will be stored in other industry sub-databases, and the entry content in other industry sub-databases will be manually classified by humans, and the number of specific industry sub-databases in the classification database will be expanded according to the classification content;

[0041] If there are two groups of entries about the same product word in the same sub-database, the two groups of entries are compared in content. When the repetition rate of the compared content reaches 85%, the two entries are judged to be the same entry, and the entry with fewer words in the two groups is deleted, so as to reduce the repetitive content in the sub-database and improve the operation efficiency. On the contrary, when the repetition rate is less than 85%, the two groups of entries are stored;

[0042] Step 3: The term analysis system analyzes the product terms proposed by the customer through big data to determine the industry field to which they belong. If there is a sub-database of the industry in the classification database, the product terms and customer demand terms proposed by the customer are searched in the sub-database of the industry. When the corresponding content is retrieved, the customer service response terms corresponding to the product terms and customer demand terms stored in the sub-database are automatically replied;

[0043] If there is a specific industry sub-database corresponding to the product word proposed by the customer in the classification database, but the customer demand word stored in the industry sub-database is not retrieved, the call will be automatically transferred to the manual customer service, and the manual customer service will respond to the customer's needs, combine the product word and the customer demand word, create an entry and directly store it in the industry sub-database;

[0044] If there is no specific industry sub-database corresponding to the product word proposed by the customer in the classification database, the call will be automatically transferred to manual customer service, and the manual customer service will respond to the customer's needs, combine the product words and customer demand words, and create an entry that is directly stored in other industry sub-databases.

[0045] Among them, the formula for the content repetition rate proposed in step 2 is repetition rate = number of repeated words / total number of words × 100%. The artificial intelligence customer service proposed in step 1 consists of a data layer, an algorithm layer, and a technical support layer. The data layer is responsible for the collection, processing, and storage of data, including the user's historical conversation records and behavior data. These data provide the basis for subsequent natural language processing, speech recognition, and large-scale data analysis. The algorithm layer includes machine learning, deep learning, and NLP technologies. The algorithm layer can understand the user's natural language expression, identify the user's intention, and generate corresponding replies. The technical support layer includes speech recognition technology, intent recognition technology, and speech synthesis technology. The artificial intelligence customer service proposed in step 1 consists of a data layer, an algorithm layer, and a technical support layer. The layer includes machine learning algorithms and deep learning algorithms. Through NLP technology, the technical support layer includes speech recognition technology, intent recognition technology and speech synthesis technology. The machine learning algorithm consists of linear regression algorithm, logistic regression algorithm, decision tree algorithm and K-means algorithm. NLP technology integrates knowledge in the fields of linguistics, computer science, machine learning, mathematics, and cognitive psychology, and consists of word vector technology and neural network technology. The deep learning algorithm consists of convolutional neural network, recurrent neural network, optimization algorithm and activation function. The classification database proposed in step 1 is a tree database. The tree database adopts a tree data structure, which includes a set of root nodes and multiple groups of element nodes. Each group of element nodes includes a parent node and zero to two child nodes.

[0046] Embodiment 1

[0047] A method for deep integration of artificial intelligence and customer service scenarios:

[0048] The method is based on:

[0049] An artificial intelligence customer service learning system is used to split the conversation between the human customer service and the customer into entries, including product words, customer demand words, and customer service response words;

[0050] Classification databases, including retail industry sub-database, financial industry sub-database, telecommunications industry sub-database, gaming industry sub-database, catering industry sub-database and other industry sub-databases;

[0051] a term analysis system for searching the term in the classification database;

[0052] Step 3: The term analysis system uses big data to analyze the product terms proposed by the customer and determine the industry field to which it belongs. When a sub-database of the industry exists in the classified database, the product terms and customer demand terms proposed by the customer are searched in the sub-database of the industry. When the corresponding content is retrieved, the customer service response words corresponding to the product terms and customer demand words stored in the sub-database are automatically replied. If a specific industry sub-database corresponding to the product terms proposed by the customer exists in the classified database, but the customer demand words stored in the industry sub-database are not retrieved, the system automatically transfers to manual customer service, and the content of the manual customer service's response to the customer's needs is combined with the product terms and customer demand words to create an entry and directly store it in the industry sub-database.

[0053] The classification database established by the above method covers most customer service scenarios on the market, and automatically answers questions in different fields raised by customers for artificial intelligence customer service, thereby greatly improving the adaptability of artificial intelligence customer service. When artificial intelligence customer service cannot respond to customer questions based on the content in the classification database, it can automatically transfer to manual customer service to ensure the flexibility of artificial intelligence customer service. Based on the content of manual customer service's response to customer needs, combined with product words and customer demand words, entries are established and stored in the industry sub-database, thereby realizing automatic expansion of the content of the classification database and improving the intelligence of the work of artificial intelligence customer service.

[0054] Embodiment 2

[0055] A method for deep integration of artificial intelligence and customer service scenarios:

[0056] The method is based on:

[0057] An artificial intelligence customer service learning system is used to split the conversation between the human customer service and the customer into entries, including product words, customer demand words, and customer service response words;

[0058] Classification databases, including retail industry sub-database, financial industry sub-database, telecommunications industry sub-database, gaming industry sub-database, catering industry sub-database and other industry sub-databases;

[0059] a term analysis system for searching the term in the classification database;

[0060] Step 3: The term analysis system uses big data to analyze the product terms proposed by the customer to determine the industry field to which they belong. If the classification database does not contain a specific industry sub-database corresponding to the product terms proposed by the customer, it will automatically transfer to manual customer service and respond to the customer's needs. The manual customer service will combine the product terms and customer demand terms to create terms and store them directly in other industry sub-databases.

[0061] The artificial intelligence customer service established by the above method can automatically transfer to manual customer service when there is no specific industry sub-database corresponding to the product word proposed by the customer in the classification database, and establish entries for the content of the manual customer service's response to customer needs, combined with product words and customer demand words, and stored in other industry sub-databases. In this way, the classification database can automatically store customer questions in unknown fields and the answers to the questions by manual customer service. Engineers manually classify the content of the entries in other industry sub-databases and expand the number of specific industry sub-databases in the classification database according to the classification content, so as to increase the number of artificial intelligence customer services that adapt to different fields and provide a basis for the application of artificial intelligence customer service in all fields.

[0062] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0063] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for deep integration of artificial intelligence and customer service scenarios, characterized by: The method is based on: An artificial intelligence customer service learning system is used to split the conversation between the human customer service and the customer into entries, including product words, customer demand words, and customer service response words; Classification databases, including retail industry sub-database, financial industry sub-database, telecommunications industry sub-database, gaming industry sub-database, catering industry sub-database and other industry sub-databases; a term analysis system for searching the term in the classification database; The method comprises the following steps: Step 1: The AI ​​customer service learning system collects the conversation content between the human customer service and the customer in real time, splits the content, and splits the content related to the product in the split content into entries; Step 2: The classification database analyzes the product words in the entry obtained in step 1 through big data to determine the industry field to which it belongs, and then stores the content of the entry in the corresponding sub-database of the classification database; If the industry field to which the product word belongs does not match the specific industry sub-database in the classification database, it will be stored in other industry sub-databases, and the entry content in other industry sub-databases will be manually classified by humans, and the number of specific industry sub-databases in the classification database will be expanded according to the classification content; If there are two groups of entries about the same product word in the same sub-database, the two groups of entries are compared in content. When the repetition rate of the compared content reaches 85%, the entry with fewer words in the two groups is deleted; Step 3: The term analysis system analyzes the product terms proposed by the customer through big data to determine the industry field to which they belong. If there is a sub-database of the industry in the classification database, the product terms and customer demand terms proposed by the customer are searched in the sub-database of the industry. When the corresponding content is retrieved, the customer service response terms corresponding to the product terms and customer demand terms stored in the sub-database are automatically replied; If there is a specific industry sub-database corresponding to the product word proposed by the customer in the classification database, but the customer demand word stored in the industry sub-database is not retrieved, the call will be automatically transferred to the manual customer service, and the manual customer service will respond to the customer's needs, combine the product word and the customer demand word, create an entry and directly store it in the industry sub-database; If there is no specific industry sub-database corresponding to the product word proposed by the customer in the classification database, the call will be automatically transferred to manual customer service, and the manual customer service will respond to the customer's needs, combine the product words and customer demand words, and create an entry that is directly stored in other industry sub-databases.

2. The method for deep integration of artificial intelligence and customer service scenarios according to claim 1, characterized in that: The formula for the content repetition rate proposed in step 2 is repetition rate = number of repeated words / total number of words×100%.

3. The method for deep integration of artificial intelligence and customer service scenarios according to claim 1, characterized in that: The artificial intelligence customer service proposed in step 1 is composed of a data layer, an algorithm layer, and a technical support layer. The algorithm layer includes machine learning algorithms, deep learning algorithms, and NLP technologies. The technical support layer includes speech recognition technology, intent recognition technology, and speech synthesis technology.

4. The method for deep integration of artificial intelligence and customer service scenarios according to claim 3 is characterized by: The machine learning algorithm consists of a linear regression algorithm, a logistic regression algorithm, a decision tree algorithm and a K-means algorithm.

5. The method for deep integration of artificial intelligence and customer service scenarios according to claim 3 is characterized by: The NLP technology integrates knowledge in the fields of linguistics, computer science, machine learning, mathematics, and cognitive psychology, and is composed of word vector technology and neural network technology.

6. The method for deep integration of artificial intelligence and customer service scenarios according to claim 3 is characterized by: The deep learning algorithm consists of a convolutional neural network, a recurrent neural network, an optimization algorithm and an activation function.

7. The method for deep integration of artificial intelligence and customer service scenarios according to claim 1, characterized in that: The classification database proposed in step 1 is a tree database, which adopts a tree data structure, including a group of root nodes and multiple groups of element nodes, and each group of element nodes includes a parent node and zero to two child nodes.

Images