Operation management platform based on AI voice communication

Abstract

The present invention discloses an operation management platform based on AI voice communication, including a voice response analysis unit, a voice feature analysis unit and an information storage management unit. The present invention relates to the field of voice communication management technology, and solves the problem that when data is stored, the characteristics of the data itself cannot be comprehensively analyzed and processed, which further leads to the technical problem of data storage security. The present invention classifies and analyzes the data, stores them according to different data types, and when classifying, performs comprehensive classification processing according to the data acquisition method, and performs storage analysis on the classified data. When performing storage analysis, repeated data in the data is analyzed and processed, and corresponding data is stored according to the characteristics of the repeated data, so as to further realize data compression processing, and at the same time, encrypted storage of data can be realized, thereby improving the security of data storage.

Description

Technical Field

[0001] The present invention relates to the field of voice communication management technology, and specifically to an operation management platform based on AI voice communication. Background Art

[0002] The operation and management platform of AI voice communication usually provides a series of intelligent functions to optimize the customer service experience and improve operational efficiency. Specific functions include intelligent customer service system, intelligent outbound call system, call center voice solution and intelligent contact center.

[0003] According to the patent application number CN202210578883.9, the system includes a test and training service module, a first AI service module, a second AI service module and a database module, which are respectively connected to the switch for communication. The test and training service module is used to provide test and training papers to users, manage the test and training process, manage user information, evaluate test results, and interact with the first AI service module for data; the first AI service module is used to control the call process, monitor, manage, record data, and provide gateway services; the second AI service module is used to provide intelligent voice interaction services to users; the database module is used to store test questions, test and training record data, and update the collected corpus knowledge.

[0004] However, AI voice communication platforms involve the collection and processing of a large amount of personal and sensitive data, which requires the platform to have strict data protection measures. Any data leakage or abuse may lead to serious legal and reputation issues. Some existing data storage methods cannot comprehensively analyze and protect data, which will further lead to security issues when storing data. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides an operation and management platform based on AI voice communication, which solves the problem that the characteristics of the data themselves cannot be comprehensively analyzed and processed when the data is stored, which further leads to the problem of data storage security.

[0006] To achieve the above objectives, the present invention is implemented through the following technical solutions: an operation management platform based on AI voice communication, comprising:

[0007] A voice information acquisition unit, used to acquire the customer's voice and transmit the acquired customer's voice to the voice response analysis unit;

[0008] The voice response analysis unit is used to analyze the response of the acquired customer voice, generate response information or secondary interaction information by matching the customer voice recognition, re-match the secondary interaction information, and transmit the generated response information to the voice information feature analysis unit;

[0009] A voice information feature analysis unit is used to comprehensively analyze the acquired response information and customer voice, and classify the response information according to the number of interactions to obtain single interaction information and multiple interaction information, and generate feature information according to the corresponding response time, the feature information includes single information features and multiple information features, and transmit the feature information to the information storage management unit;

[0010] An information storage management unit is used to store and analyze the response information in combination with the acquired feature information. For single interaction information, duplicate data is first extracted, and different storage analyses are performed according to the types of duplicate data, and storage information is generated at the same time. For multiple interaction information, duplicate data is first extracted, and a matching template is generated according to the duplicate data, and multiple interaction information is further stored according to the matching template to generate storage information, and then the generated storage information is transmitted to the operation management output unit;

[0011] The operation management output unit is used to store the response information according to the acquired storage information.

[0012] As a further solution of the present invention: the specific manner in which the voice response analysis unit performs response analysis on the customer voice is:

[0013] S1: Recognize the acquired customer voice to obtain recognition content, and determine whether the recognition content has corresponding response content. If there is response content, obtain the corresponding response content and record it as response information. Otherwise, if there is no response content, generate secondary interaction information.

[0014] S2: For the generated secondary interaction information, the system reminds the customer to conduct a secondary interaction and obtain the secondary interaction voice, and at the same time recognizes and analyzes the secondary interaction voice, and repeats S1 to generate corresponding response information, and then transmits the generated response information to the voice information feature analysis unit.

[0015] As a further solution of the present invention: the specific manner in which the voice feature analysis unit classifies the response information is as follows:

[0016] All customer voices within a time period t are obtained, and all customer voices are marked as i, where i=1, 2, ..., j. Then, the response information corresponding to the customer voice i is obtained, and the customer voices are classified according to the number of interactions. The response information with only one interaction is classified as single interaction information, and vice versa, the response information with two or more interactions is classified as multiple interaction information.

[0017] As a further solution of the present invention: the specific manner in which the voice feature analysis unit extracts the features of the response information is:

[0018] The method of extracting and analyzing the features of the single interaction information is as follows: obtaining the response time corresponding to the single interaction information and recording it as T1, and taking the response time as the information feature of the single interaction information, and generating the single information feature;

[0019] The method of extracting and analyzing the features of the multiple interactive information is as follows: obtaining the response time corresponding to the multiple interactive information and recording it as T2, and taking the response time as the information feature of the multiple interactive information, and generating multiple information features;

[0020] Then, the single feature information and the multiple feature information are integrated to generate feature information, and the generated feature information is transmitted to the information storage management unit.

[0021] As a further solution of the present invention: the specific manner in which the information storage management unit stores and analyzes the single interaction information in combination with the feature information is:

[0022] Obtain all single interaction information and label them as a, where a=1, 2, ..., n. Then, select a set of single interaction information as the analysis object k, and obtain the data capacity of the analysis object as Lk. Then, obtain the repeated data in the analysis object, extract the repeated data, and record the remaining single interaction information as single difference and similarity data.

[0023] Then the type of repeated data is judged. When there is only one type of repeated data, the data capacity of the repeated data is obtained, and the single different and similar data is segmented based on the data capacity of the repeated data. After segmenting the single different and similar data, multiple groups of single different and similar data segmentation packages are obtained.

[0024] As a further solution of the present invention, the specific method of storing and analyzing multiple groups of single-time different and similar data segmentation packages when the type of repeated data is one is:

[0025] Obtain any group of single-time data segmentation packets of similarities and differences as analysis data packets, then obtain duplicate data corresponding to the analysis data packets, perform binary conversion on the duplicate data to obtain binary conversion information, and simultaneously obtain the data capacity of the analysis data packets and judge the data capacity thereof;

[0026] When the single feature information value is an odd number, all odd-bit binary numbers in the binary conversion information are obtained and combined to generate a replacement template. Conversely, when the single feature information value is an even number, all even-bit binary numbers in the binary conversion information are obtained and combined to generate a matching template.

[0027] Then, the matching template is matched and analyzed with the single difference and similarity data segmentation package, and the binary number in the single difference and similarity data segmentation package that is the same as the matching template is obtained, and the obtained binary number is converted. Finally, the single difference and similarity data segmentation package is stored, storage information is generated, and the storage information is transmitted to the operation management output unit.

[0028] As a further solution of the present invention: the specific method of storing and analyzing multiple groups of single-time different and similar data segmentation packages when there are multiple types of repeated data is:

[0029] All duplicate data and their corresponding data capacity are obtained, and then the capacity mean corresponding to all duplicate data is calculated, and the calculated capacity mean is used as the segmentation standard. At the same time, the different and similar data are segmented according to the segmentation standard to obtain multiple groups of different and similar data packets. Then, the corresponding storage information is generated in a similar way for the type of duplicate data, and the storage information is transmitted to the operation management output unit.

[0030] As a further solution of the present invention: the specific manner in which the information storage management unit stores and analyzes multiple interactive information is:

[0031] All the multiple interaction information are obtained, and a group of multiple interaction information is selected as the target object. Then, the identification content and the response content in the target object are obtained, and the two are compared and screened to obtain the content with the highest matching value as the high matching content, and the high matching content is binary converted to obtain binary matching information;

[0032] Extract the duplicate data in the target object, record the remaining data as multiple similarity and difference data, then obtain multiple feature information corresponding to the target object, and segment the target object using the sum of the multiple feature information values ​​as the segmentation standard to obtain multiple groups of multiple similarity and difference data segmentation packages;

[0033] Then, any group of multiple different and similar data segmentation packages are obtained, and the repeated data is converted into binary form. At the same time, the data capacity of the repeated data is analyzed. When the data capacity of the repeated data is an odd number, the odd-bit binary number corresponding to the binary conversion is obtained and selected. Then, a matching template is generated with the selected odd-bit binary number. At the same time, the matching template is matched with the multiple different and similar data segmentation packages, and the matched binary numbers are converted. Finally, the multiple different and similar data segmentation packages are stored to generate storage information.

[0034] The present invention provides an operation management platform based on AI voice communication. Compared with the prior art, it has the following beneficial effects:

[0035] The present invention classifies and analyzes data and stores them according to different data types. When classifying, it performs comprehensive classification processing based on the data acquisition method, and performs storage analysis on the classified data. When performing storage analysis, it analyzes and processes duplicate data in the data, and stores the corresponding data according to the characteristics of the duplicate data, further realizes data compression processing, and can also realize encrypted storage of data, thereby improving the security of data storage. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 It is the principle diagram of the system of the present invention;

[0037] Figure 2 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0038] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0039] See also Figure 1 and Figure 2 , this application provides an operation management platform based on AI voice communication, including: a voice information acquisition unit, a voice response analysis unit, a voice feature analysis unit, an information storage management unit and an operation management output unit.

[0040] As embodiment 1 of the present invention

[0041] The voice information acquisition unit is used to acquire the customer's voice and transmit the acquired customer's voice to the voice response analysis unit.

[0042] The voice response analysis unit is used to analyze the response of the acquired customer voice, and the specific response analysis method is as follows:

[0043] Recognize the acquired customer voice, and the voice recognition method here is to convert the acquired voice, usually into a format that can be understood by the computer, such as text or command, to obtain the recognition content, and then match and interact according to the recognition content to determine whether there is corresponding response content for the recognition content. When there is response content, the corresponding response content is obtained and recorded as response information. On the contrary, when there is no response content, secondary interaction information is generated;

[0044] For the generated secondary interaction information, the system will remind the corresponding customer to conduct a secondary interaction, further obtain the customer's secondary interaction voice, and recognize and analyze the secondary interaction voice, and repeat the above steps to generate corresponding response information, and then transmit the generated response information to the voice information feature analysis unit.

[0045] Specifically, when customers are interacting, the system may be unable to recognize the acquired voice due to various factors. Specific influencing factors include: unclear voice, vague speech, or excessive external noise resulting in failure to acquire the voice. In the above situations, it is necessary to re-acquire and further recognize the response.

[0046] The voice information feature analysis unit is used to comprehensively analyze the acquired response information and customer voice, and the specific comprehensive analysis method is as follows:

[0047] All customer voices within a time period t are obtained, and the specific value of the time period t is set by the operator. At the same time, all customer voices are marked as i, and i=1, 2, ..., j. Then, the response information corresponding to the customer voice i is obtained, and the customer voice is classified according to the number of interactions. The response information with only one interaction is classified as single interaction information, and vice versa, when there are two or more response information, it is classified as multiple interaction information;

[0048] Then, feature extraction and analysis are performed on the single interaction information and the multiple interaction information respectively, and the method of feature extraction and analysis on the single interaction information is as follows:

[0049] The response time corresponding to the single interaction information is obtained and recorded as T1, and the response time is used as the information feature of the single interaction information, and a single information feature is generated;

[0050] The method of feature extraction and analysis of multiple interactive information is as follows:

[0051] The response time corresponding to the multiple interactive information is obtained and recorded as T2. At the same time, the response time is used as the information feature of the multiple interactive information, and the multiple information features are generated. Then, the single feature information and the multiple feature information are integrated to generate feature information, and the generated feature information is transmitted to the information storage management unit. Similarly, the same analysis is performed on all multiple interactive information and single interactive information.

[0052] The information storage management unit is used to store and analyze the response information in combination with the acquired feature information. Specifically, the response information acquired here includes the customer voice recognition content and the response content, and the specific storage and analysis method is as follows:

[0053] The specific method of storing and analyzing single interaction information is as follows:

[0054] Obtain all single interaction information and label them as a, where a=1, 2, ..., n. Then, select a set of single interaction information as the analysis object k, and obtain the data capacity of the analysis object as Lk. Then, obtain the repeated data in the analysis object, extract the repeated data, and record the remaining single interaction information as single difference and similarity data.

[0055] Then, the type of the repeated data is judged. When there is only one type of repeated data, the data capacity of the repeated data is obtained, and the single different and similar data is segmented based on the data capacity of the repeated data. After segmenting the single different and similar data, multiple groups of single different and similar data segmentation packages are obtained.

[0056] Obtain any group of single-time difference and similarity data segmentation packages as analysis data packages, then obtain duplicate data corresponding to the single-time difference and similarity data, perform binary conversion on the duplicate data to obtain binary conversion information, and simultaneously obtain the data capacity of the analysis data package and determine its data capacity;

[0057] The specific judgment method is: when the value of the single feature information is an odd number, all the odd-numbered binary numbers in the binary conversion information are obtained, and they are combined to generate a replacement template; conversely, when the value of the single feature information is an even number, all the even-numbered binary numbers in the binary conversion information are obtained, and they are combined to generate a matching template;

[0058] Combined with actual analysis, for example, if the value of a single feature information is an odd number, it means that the overall numerical property is an odd number, that is, the value of the last digit is a series of odd numbers such as 1, 3, 5, etc. When it is an odd number, select the binary numbers of all odd bits in the binary information, and then combine the selected binary numbers, and match them according to the combined bit standard, replace the binary numbers of the same form in the analysis object, and finally store the whole.

[0059] Then, the matching template is matched and analyzed with the single difference and similarity data segmentation package, and the binary number in the single difference and similarity data segmentation package that is the same as the matching template is obtained, and the obtained binary number is converted. Finally, the single difference and similarity data segmentation package is stored, storage information is generated, and the storage information is transmitted to the operation management output unit.

[0060] The operation management output unit is used to store data according to the acquired storage information.

[0061] As the second embodiment of the present invention, this embodiment is implemented on the basis of the first embodiment, and is different from the first embodiment in that.

[0062] The information storage management unit analyzes when there are multiple types of duplicate data, and the specific analysis method is as follows:

[0063] When there are multiple types of duplicate data, all duplicate data and their corresponding data capacities are obtained, and then the capacity mean corresponding to all duplicate data is calculated, and the calculated capacity mean is used as the segmentation criterion. At the same time, the different and similar data are segmented according to the segmentation criterion to obtain multiple groups of different and similar data packets, and then the corresponding storage information is generated in the same way as in Example 1, and the storage information is transmitted to the operation management output unit.

[0064] As the third embodiment of the present invention, it is implemented on the basis of the first and second embodiments, and is different from the first and second embodiments in that.

[0065] The information storage management unit is used to store multiple interactive information and generate storage information, and the specific analysis method is as follows:

[0066] All the multiple interaction information are obtained and labeled as b, where b=1, 2, ..., m. At the same time, a group of multiple interaction information is selected as the target object, and then the recognition content and response content in the target object are obtained, and the two are compared and screened to obtain the content with the highest matching value as the high matching content, and the high matching content is binary converted to obtain binary matching information;

[0067] Specific high-matching content refers to the information that exists most in the identification content and the response content. For example, the identification content is "Query today's weather" and the response content is "Today's highest temperature is 30° and the lowest temperature is 20°". Therefore, after comparing and screening the two, "Today's weather" is the content with the highest matching value, so it is screened and binary converted at the same time.

[0068] Obtain repeated data in the target object, extract the repeated data, and record the remaining data as multiple similarity and difference data, then obtain multiple feature information corresponding to the target object, and the multiple feature information obtained here is represented as multiple feature information corresponding to the target object one by one, and the specific different multiple interaction information will correspond to one multiple feature information, and the target object is segmented by taking the sum of the multiple feature information values ​​as the segmentation standard to obtain multiple groups of multiple similarity and difference data segmentation packages;

[0069] Then, any group of multiple different and similar data segmentation packages are obtained, and the repeated data is converted into binary form. At the same time, the data capacity of the repeated data is analyzed. When the data capacity of the repeated data is an odd number, the odd-bit binary number corresponding to the binary conversion is obtained and selected. Then, a matching template is generated with the selected odd-bit binary number. At the same time, the matching template is matched with the multiple different and similar data segmentation packages, and the matched binary numbers are converted. Finally, the multiple different and similar data segmentation packages are stored to generate storage information, and the generated storage information is transmitted to the operation management output unit.

[0070] The operation management output unit is used to store data according to the acquired storage information.

[0071] Some of the data in the above formulas are dimensionless and numerically calculated. Meanwhile, the contents not described in detail in this specification belong to the prior art known to those skilled in the art.

[0072] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical method of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present invention.

Claims

1. The operation management platform based on AI voice communication is characterized by: include: A voice information acquisition unit, used to acquire the customer's voice and transmit the acquired customer's voice to the voice response analysis unit; The voice response analysis unit is used to analyze the response of the acquired customer voice, generate response information or secondary interaction information by matching the customer voice recognition, re-match the secondary interaction information, and transmit the generated response information to the voice information feature analysis unit; The voice information feature analysis unit is used to comprehensively analyze the acquired response information and customer voice, and classify the response information according to the number of interactions to obtain single interaction information and multiple interaction information, and generate feature information according to the corresponding response time, the feature information includes single information features and multiple information features, and transmit the feature information to the information storage management unit. The specific method of classifying the response information is: All customer voices within the time period t are obtained, and all customer voices are marked as i, where i=1, 2, ..., j. Then, the response information corresponding to the customer voice i is obtained, and the customer voices are classified according to the number of interactions. The response information with only one interaction is classified as single interaction information, and the response information with two or more interactions is classified as multiple interaction information. The information storage management unit is used to store and analyze the response information in combination with the acquired feature information. For single interaction information, the repeated data is first extracted, and different storage analyses are performed according to the types of repeated data, and storage information is generated at the same time. For multiple interaction information, the repeated data is first extracted, and a matching template is generated according to the repeated data. The multiple interaction information is further stored according to the matching template to generate storage information, and then the generated storage information is transmitted to the operation management output unit. The specific method of storing and analyzing the single interaction information is as follows: Obtain all single interaction information and label them as a, where a=1, 2, ..., n. Then, select a group of single interaction information as the analysis object k, and obtain the data capacity of the analysis object as Lk. Then, obtain the repeated data in the analysis object, extract the repeated data, and record the remaining single interaction information as single difference and similarity data. Then, the type of the repeated data is judged. When there is only one type of repeated data, the data capacity of the repeated data is obtained, and the single different and similar data is segmented based on the data capacity of the repeated data. After segmenting the single different and similar data, multiple groups of single different and similar data segmentation packages are obtained. The specific method for storing and analyzing multiple groups of single-time data segmentation packages when the type of repeated data is one is as follows: Obtain any group of single-time data segmentation packets of similarities and differences as analysis data packets, then obtain duplicate data corresponding to the analysis data packets, perform binary conversion on the duplicate data to obtain binary conversion information, and simultaneously obtain the data capacity of the analysis data packets and judge the data capacity thereof; When the single feature information value is an odd number, all odd-bit binary numbers in the binary conversion information are obtained and combined to generate a replacement template. Conversely, when the single feature information value is an even number, all even-bit binary numbers in the binary conversion information are obtained and combined to generate a matching template. Then, the matching template is matched and analyzed with the single-time difference and similarity data segmentation package, and the binary number in the single-time difference and similarity data segmentation package that is the same as the matching template is obtained, and the obtained binary number is converted at the same time. Finally, the single-time difference and similarity data segmentation package is stored, storage information is generated, and the storage information is transmitted to the operation management output unit; The specific method of storing and analyzing multiple interactive information is as follows: All the multiple interaction information are obtained, and a group of multiple interaction information is selected as the target object. Then, the identification content and the response content in the target object are obtained, and the two are compared and screened to obtain the content with the highest matching value as the high matching content, and the high matching content is binary converted to obtain binary matching information; Extract the duplicate data in the target object, record the remaining data as multiple similarity and difference data, then obtain multiple feature information corresponding to the target object, and segment the target object using the sum of the multiple feature information values ​​as the segmentation standard to obtain multiple groups of multiple similarity and difference data segmentation packages; Then, any group of multiple different and similar data segmentation packages are obtained, and the repeated data is converted into binary, and the data capacity of the repeated data is analyzed. When the data capacity of the repeated data is an odd number, the odd-bit binary number corresponding to the binary conversion is obtained and selected, and then a matching template is generated with the selected odd-bit binary number, and the matching template is matched with the multiple different and similar data segmentation packages, and the binary numbers obtained by matching are converted, and finally the multiple different and similar data segmentation packages are stored to generate storage information; The operation management output unit is used to store the response information according to the acquired storage information.

2. The operation management platform based on AI voice communication according to claim 1, characterized in that: The specific method of the voice response analysis unit performing response analysis on the customer voice is as follows: S1: Recognize the acquired customer voice to obtain recognition content, and determine whether the recognition content has corresponding response content. If there is response content, obtain the corresponding response content and record it as response information. Otherwise, if there is no response content, generate secondary interaction information. S2: For the generated secondary interaction information, the system reminds the customer to conduct a secondary interaction and obtain the secondary interaction voice, and at the same time recognizes and analyzes the secondary interaction voice, and repeats S1 to generate corresponding response information, and then transmits the generated response information to the voice information feature analysis unit.

3. The operation management platform based on AI voice communication according to claim 1, characterized in that: The specific method in which the voice feature analysis unit extracts the features of the response information is as follows: The method of extracting and analyzing the features of the single interaction information is as follows: obtaining the response time corresponding to the single interaction information and recording it as T1, and taking the response time as the information feature of the single interaction information, and generating the single information feature; The method of extracting and analyzing the features of the multiple interactive information is as follows: obtaining the response time corresponding to the multiple interactive information and recording it as T2, and taking the response time as the information feature of the multiple interactive information, and generating multiple information features; Then, the single feature information and the multiple feature information are integrated to generate feature information, and the generated feature information is transmitted to the information storage management unit.

4. The operation management platform based on AI voice communication according to claim 1, characterized in that: The specific method of storing and analyzing multiple groups of single-time data segmentation packages with different types of repeated data is as follows: All duplicate data and their corresponding data capacity are obtained, and then the capacity mean corresponding to all duplicate data is calculated, and the calculated capacity mean is used as the segmentation standard. At the same time, the different and similar data are segmented according to the segmentation standard to obtain multiple groups of different and similar data packets. Then, the corresponding storage information is generated in a similar way for the type of duplicate data, and the storage information is transmitted to the operation management output unit.

Images